1
Face Processing Impairments in Schizophrenia and Other Psychiatric Disorders
Hayley Darke (519351)
ORCID ID: 0000-0003-0262-5330
Submitted in partial fulfilment of the requirements of the combined degree of Master of Psychology
(Clinical Neuropsychology) and Doctor of Philosophy.
May 2018
Melbourne School of Psychological Sciences
Faculty of Medicine, Dentistry and Health Sciences
University of Melbourne
2
Abstract
The ability to recognise and interpret the facial expressions of others is shown to be impaired in
individuals with schizophrenia and may contribute to poor social functioning. In contrast, it remains
unclear whether other aspects of face processing (such as identity recognition) are also impaired, and
whether such deficits correlate with specific symptoms. This thesis explores face processing in
schizophrenia and other psychiatric disorders. To test this, a novel set of video-based tasks were
designed to improve upon traditional image-based tasks, as these have been criticised for lacking
ecological validity. We show that these video-based tasks are recognised more easily than image-
based versions and are sensitive to schizophrenia-like experiences in a study of healthy individuals
(n=82). Eighty-six psychiatric inpatients and an additional twenty healthy controls completed a series
of five video-based tasks: Emotion Discrimination (same or different), Emotion Labelling (fear or
disgust), Identity Discrimination (same or different), Sex Labelling (male or female), and a non-face
control task – Car Discrimination (same or different). Schizophrenia patients (n=36) were impaired
compared to healthy controls across all tasks except for Identity Discrimination (which showed
marginal impairment), and Sex Labelling. When all patient groups were compared, it was revealed
that the schizophrenia (n=36) and bipolar disorder groups (n=15) were significantly impaired on the
emotion-processing tasks and Car Discrimination compared to both healthy controls (n=20) and
patients with non-psychotic diagnoses (n=18). There were no significant differences between patients
and healthy controls on the general face tasks (Identity Discrimination and Sex Labelling). These
findings hint that the “emotion-specific” processing deficits reported in previous studies may
represent a more general cognitive or perceptual impairment. Correlational analyses revealed that
non-cognitive positive symptoms – such as delusions and suspiciousness – were negatively correlated
with emotion-processing ability and car discrimination, but not identity-processing ability. In contrast,
more severe cognitive symptoms were associated with generally reduced performance across tasks.
Overall, results suggest that deficits in emotion processing reflect symptom pathology independent of
diagnosis and support the idea of a generalised cognitive deficit that is particularly prominent in
patients showing positive symptoms of psychosis. These findings are discussed in terms of
dimensional models of psychosis. Furthermore, they highlight the importance of including appropriate
control tasks when assessing allegedly specific deficits.
3
Declaration
This is to certify that:
i. the thesis comprises only my original work towards the PhD except where indicated in the
Preface,
ii. due acknowledgement has been made in the text to all other material used,
iii. the thesis is fewer than 100 000 words in length, exclusive of tables, maps, bibliographies and
appendices,
iv. the research reported in this thesis was conducted in accordance with the principles of the
ethical treatment of human subjects as approved for research by Human Research Ethics
Committee at the University of Melbourne and Northern Health.
Hayley Darke 20th May 2018
4
Preface
Acknowledgement of contributors and funding
Data for the experiment described in Chapters 5 and 6 were collected as part of a larger
battery which was administered by a visiting master’s student, Lola Kourouma. Data for the
additional experiment described in Chapter 9 were collected with the assistance of an honours student,
Letian Wang. However, all task development, programming, data cleaning and analyses, and
interpretation of results were conducted by the PhD candidate alone, and therefore have been included
as a part of this thesis.
Parts of this thesis have appeared in peer-reviewed manuscripts with multiple authors (listed
below). However, all chapters of this thesis were drafted by the PhD candidate alone. This research
was funded by an Australian Research Council Grant to A/Prof Olivia Carter: FT140100807.
Publications arising from this thesis
Darke, H., Sundram, S., Cropper, S. J., Carter, O. (In preparation). Deficits in facial emotion
processing are associated with psychotic symptoms and are not limited to schizophrenia.
Darke, H., Cropper, S. & Carter, O. (Manuscript under review). Using morphing to vary person
and emotion attributes in face video stimuli. Behavior Research Methods.
Peterman, J. S., Darke, H., Carter, O, Sundram, S. & Park, S. (2017). M124. Perception of socio-
emotional information in dynamic gait in inpatients with schizophrenia spectrum disorders.
Schizophrenia Bulletin, 43(suppl_1), S255-S256.
Darke, H., Peterman, J. S., Park, S., Sundram, S. & Carter, O. (2013). Are patients with
schizophrenia impaired in processing non-emotional features of human faces? Frontiers in
Psychology, 4, 529.
Darke, H. Dynamic stimulus database. The emotional faces, non-emotional faces, and car stimuli
developed for this thesis can be downloaded from http://go.unimelb.edu.au/e3t6.
Posters and presentations
Darke, H., Sundram, S. & Carter, O. (2014). Face processing in schizophrenia and other
psychiatric disorders. Talk presented at the Department of Psychiatry Research Symposium,
University of Melbourne, Australia. Winner of the Best Presenter Award for Early Career
Researcher.
Darke, H., Carter, O. & Sundram, S. (2014). Deficits in facial expression and identity recognition
in schizophrenia. Poster presented at the Biological Psychiatry Australia Conference, Melbourne,
Australia.
Darke, H., Sundram, S. & Carter, O. (2014). Facial expression and identity recognition
impairments in schizophrenia and other disorders. Poster presented at the Integrative Brain
Function Workshop, MBI, Monash University, Melbourne, Australia.
5
Acknowledgements
Holy moly, I can’t believe I’m writing this. We did it! And I must say “we”, because this
thesis could never have existed without the many generous, patient, and ridiculously clever people
who have supported me through this journey. Foremost, I want to thank my primary supervisor A/Prof
Olivia Carter for her tireless guidance and encouragement throughout my candidature. I also want to
thank my co-supervisors Professor Suresh Sundram and Dr Simon Cropper for their excellent input at
all stages of the project. I am indebted to my partner, friends, colleagues, and family for their
constant encouragement, assistance and, of course, gentle prodding to stop procrastinating and finish
this thesis. I am also extremely grateful to the staff at the School of Psychological Sciences and the
Northern Hospital for supporting my data collection efforts.
Finally, I want to thank the participants who took part in this study. I know it was boring
(sorry!), but it is thanks to your generosity that research like this is made possible.
6
Table of Contents
Chapter 1: Introduction ......................................................................................................................... 10
Chapter 2: Face Processing in Healthy versus Schizophrenia Populations .......................................... 14
Chapter 3: Is it really all about emotion? Non-emotional Face Processing is (Maybe) Impaired in
Schizophrenia ........................................................................................................................................ 29
Chapter 4: What do we really want to know about face processing in schizophrenia? Rationale for the
Inpatient Study ...................................................................................................................................... 40
Chapter 5: Is our data only as good as our tools? Issues of Static versus Dynamic Faces, and
Development of a Novel Stimuli Set ...................................................................................................... 44
Chapter 6: Are these tasks sensitive to schizophrenia-like traits? Schizotypy in Healthy Controls and
the Relation to Emotion-processing Performance ................................................................................ 70
Chapter 7: Method for Inpatient Study ................................................................................................. 78
Chapter 8: So what kind of deficit is it, really? Characterising Face Processing Deficits in Inpatients
with Schizophrenia ................................................................................................................................ 87
Chapter 9: Could face-processing impairments in schizophrenia be mediated by deficits in allocating
visuospatial attention? ........................................................................................................................... 98
Chapter 10: Are these deficits specific to schizophrenia? Overlap with Face Processing Deficits in
Other Psychiatric Disorders ............................................................................................................... 115
Chapter 11: What else accompanies these deficits? Symptom Correlates of Face-processing Deficits
in Schizophrenia and Other Disorders ............................................................................................... 131
Chapter 12: General Discussion .......................................................................................................... 162
References ........................................................................................................................................... 178
Appendices .......................................................................................................................................... 208
7
List of Tables
Table 3. Identity processing tasks used in schizophrenia research from the last 25 years. .................. 32
Table 5.1. Behavioural studies comparing emotional processing of dynamic and static face stimuli in
healthy controls. .................................................................................................................................... 49
Table 5.2. Behavioural studies using dynamic stimuli to investigate emotion processing in
schizophrenia. ....................................................................................................................................... 50
Table 5.3. Participants’ (n=82) demographic information. ................................................................. 57
Table 5.4. Pearson correlations between task performance and demographic factors. ...................... 67
Table 6.1. Healthy controls’ (n=82) demographics and mean raw scores for the O-LIFE scales. ...... 74
Table 6.2. Spearman rank correlations between face-processing performance and O-LIFE scores for
healthy controls (n=82). ....................................................................................................................... 75
Table 7.1. Breakdown of patient diagnoses by group. ......................................................................... 79
Table 7.2. Mean participant demographics and questionnaire scores by group (SD in parentheses). 84
Table 7.3. Medication status for inpatient groups. ............................................................................... 85
Table 8. Mean participant demographics by group (SD in parentheses). ............................................ 88
Table 9.1. Mean participant demographics and questionnaire scores by group (SD in parentheses).
............................................................................................................................................................ 104
Table 9.2. Uncorrected Pearson correlations between demographic factors and task performance. 105
Table 9.3. Mean c scores for patients with schizophrenia and healthy controls on the five tasks. ..... 105
Table 9.4. Bonferroni-corrected Pearson correlations for performance across different tasks. ........ 110
Table 9.5. Uncorrected partial correlations for performance across different tasks, controlling for
Age, years of Education, and IQ. ........................................................................................................ 111
Table 10. Mean d’ scores for each group on the five tasks. ................................................................ 121
Table 11.1. Studies reporting symptom correlates with reduced performance on emotion tasks in
schizophrenia spectrum disorders. ..................................................................................................... 138
Table 11.2. Studies reporting symptom correlates with reduced performance on non-emotion face
tasks in schizophrenia spectrum disorders. ........................................................................................ 146
Table 11.3. Pearson correlations between PANSS scores and performance (d’) on the five dynamic
tasks. ................................................................................................................................................... 154
Table 11.4. Spearman rank correlations between individual PANSS items and performance (d’) on the
five dynamic tasks ............................................................................................................................... 154
Table 11.5. Characteristics of four clusters based on patient performance. ...................................... 156
8
List of Figures
Figure 2.1. A visual representation showing the separation of emotion and identity processing during
face perception. ..................................................................................................................................... 15
Figure 2.2. Bruce and Young’s theory of face perception. ................................................................... 16
Figure 2.3. The distributed neural system for face perception. ............................................................ 17
Figure 2.4. Examples of labelling and discrimination tasks. ................................................................ 21
Figure 3. Examples of tasks used to assess the perception of identity. ................................................ 31
Figure 5.1. Example of morphed stimuli created by Norton and colleagues (2009, p.1095) used in an
emotion discrimination task. ................................................................................................................. 46
Figure 5.2. Partial image sequences (every other frame) from five of the video stimuli ranging in
intensity of emotion. ............................................................................................................................. 54
Figure 5.3. Partial image sequences from six different non-emotional face video stimuli. ................. 55
Figure 5.4. Partial image sequences from two different 3D video stimuli used in the Car
Discrimination task. .............................................................................................................................. 56
Figure 5.5. Trial sequences for Emotion Discrimination (A) and Emotion Labelling (B .................... 59
Figure 5.6. Comparison of percent accuracy by emotional intensity for dynamic and static conditions
on the Discrimination task (A) and the Labelling task (B). Error bars indicate 95% confidence
intervals around means. ........................................................................................................................ 61
Figure 5.7. Raw accuracy rates (A) and d prime scores (B) for Dynamic and Static stimuli for the two
tasks: Emotion Discrimination and Emotion Labelling. ....................................................................... 62
Figure 5.8. Mean response bias as evidenced by c rates across the four task conditions. Error bars
indicate 95% confidence intervals. ....................................................................................................... 63
Figure 7. Example trials for each of the five tasks: Emotion Discrimination, Emotion Labelling,
Identity Discrimination, Sex Labelling, and Car Discrimination. ........................................................ 81
Figure 8.1. Performance (d’) of healthy controls and patients with schizophrenia across the five
dynamic tasks. ....................................................................................................................................... 88
Figure 8.2. Performance (d’) of healthy controls across the five dynamic tasks. ................................. 90
Figure 8.3. Performance (d’) of patient with schizophrenia across the five dynamic tasks. ................ 90
Figure 8.4. Mean response bias as evidenced by c rates across the five tasks for schizophrenia patients
versus controls. ..................................................................................................................................... 91
Figure 8.5. Mean accuracy performance for the schizophrenia spectrum (SZ) and control groups for
(A) emotion discrimination and (B) emotion labelling. ........................................................................ 92
9
Figure 8.6. Mean accuracy performance for the schizophrenia spectrum (SZ) and healthy control
groups for (A) identity discrimination and (B) sex labelling. ............................................................... 94
Figure 9.1. Example of hierarchical stimuli used by Johnson and colleagues (2005, p.938). .............. 99
Figure 9.2. The Navon task ................................................................................................................ 102
Figure 9.3. Performance (d’) of healthy controls, psychiatric controls and schizophrenia spectrum
patients for the identity discrimination and emotion discrimination tasks .......................................... 106
Figure 9.4. Performance (d’) of healthy controls, psychiatric controls and schizophrenia spectrum
patients for the global and local attentional tasks. .............................................................................. 107
Figure 9.5. Mean local processing bias for healthy controls, psychiatric controls and schizophrenia
spectrum patients. ............................................................................................................................... 108
Figure 9.6. Reaction times (ms) of healthy controls, psychiatric controls and schizophrenia spectrum
patients for the global and local attentional tasks. .............................................................................. 109
Figure 10.1. Mean d’ performance for the schizophrenia spectrum (SZ), bipolar disorder (BPAD),
other psychosis (Other), non-psychosis (NP) and healthy control groups across the five tasks ......... 122
Figure 10.2. Performance (d’) of patients with bipolar disorder across the five dynamic tasks. ....... 123
Figure 10.3. Performance (d’) of patients with non-schizophrenia psychosis across the five dynamic
tasks. ................................................................................................................................................... 124
Figure 10.4. Performance (d’) of patients with non-psychotic disorders across the five dynamic tasks.
............................................................................................................................................................ 125
Figure 10.5. Mean accuracy performance for the schizophrenia spectrum, bipolar disorder, non-
schizophrenia psychosis (Other psychosis), non-psychotic disorders and control groups across the four
morphed face tasks: Emotion Discrimination (A), Emotion Labelling for disgust faces (B), Emotion
Labelling for fear faces (C), Identity Discrimination (D), and Sex Labelling (E). ............................. 127
Figure 11. Mean performance across five dynamic tasks for the four patient clusters. ..................... 157
Figure 12.1. Pentagonal bifactor model of psychosis. ........................................................................ 171
Figure 12.2. A stress-diathesis model demonstrating the interaction between exposure and genetic
vulnerability to psychosis. .................................................................................................................. 173
10
Chapter 1: Introduction
Schizophrenia is a debilitating psychiatric disorder affecting an estimated 0.3 -.66% of
individuals globally (McGrath, Saha, Chant, & Welham, 2008). It is classically characterised by a
triad of symptoms: positive symptoms, including hallucinations and delusions; negative symptoms,
such as diminished emotional expression; and disorganised symptoms, which describe disordered
thoughts and behaviour (Tandon, Nasrallah, & Keshavan, 2009). With onset of the disorder usually
occurring in early adulthood, the course of schizophrenia typically involves fluctuating or episodic
positive and disorganised symptoms accompanied by persisting negative symptoms (Keshavan,
Nasrallah, & Tandon, 2011). In addition, schizophrenia is associated with persisting cognitive
impairment in areas such as basic attention, processing speed, working memory, verbal learning and
memory, and reasoning (Keefe & Harvey, 2012). Although cognitive impairment is not included in
the DSM-V description of schizophrenia, such impairments are increasingly recognised as a core
feature of this disorder, with the average patient performing two standard deviations below the mean
for healthy individuals (Keefe et al., 2011).
Despite their relatively low prevalence, schizophrenia and other psychotic disorders account
for the majority of spending on specialist mental health services in Australia (Australian Institute of
Health and Welfare, 2015). It was estimated that in 2014, psychotic disorders such as schizophrenia
cost the Australia population approximately $10 billion in medical costs, disability support, and lost
productivity (Victoria Institute of Strategic Economic Studies, 2016). A national psychosis survey
conducted in 2010 highlights the extent of disability and disadvantage associated with psychotic
disorders in Australia (Morgan et al., 2012). It was found that almost 20% of patients accessing
mental health services had difficulties with reading and writing, two-thirds had not completed high
school, and only one-third had held some form of employment in the past 12 months. Half of
respondents reported attempting suicide in the past, and 13% had been homeless in the past 12
months. Moreover, the life expectancy of individuals with schizophrenia is 14.5 years below the
average, and has not improved over recent years (Hjorthoj, Sturup, McGrath, & Nordentoft, 2017).
Suggested reasons for premature death include an increased risk of cardiovascular and metabolic
disease, substance use, and reduced health-seeking behaviour. However, risk of death from accidents
and homicide is also increased, and the risk of suicide is 22 times higher in schizophrenia compared to
the general population (McGrath et al., 2008).
The etiology of schizophrenia remains unknown, and is likely complex and multifactorial. It
is known that the disease is highly heritable (approximately 80%; Van Os, 2009) and has also been
linked to a variety of environmental factors such as maternal malnutrition, prenatal infection, urban
birth, obstetric complications, trauma and cannabis exposure (Keshavan et al., 2011). While a range of
11
different models have been proposed to explain the pathophysiology of schizophrenia (e.g.: aberrant
salience, oxidative stress, dopamine imbalance) there is not yet a unified explanation that can account
for all aspects of the disorder. One major obstacle to clinical research is the high level of
heterogeneity within schizophrenia. For example, it is unclear why one individual presents with
disorganised behaviour and significant cognitive impairment, while another presents with complex
persecutory delusion. Historically, this heterogeneity was explained in terms of four clinical subtypes:
disorganised or hebephrenic, catatonic, paranoid, and undifferentiated (i.e.: none of the above)
schizophrenia. However, studies have consistently failed to show support for these subtypes, instead
finding that the vast majority of patients fall into the undifferentiated category (Tandon et al., 2013).
For this reason, clinical subtypes have since been removed from the DSM-V.
A second obstacle to clinical research is the substantial overlap between schizophrenia and
other psychiatric disorders. Clinically, schizophrenia can be difficult to differentiate from other
diagnoses with psychotic features, such as bipolar disorder and psychotic depression (Dacquino, De
Rossi, & Spalletta, 2015). This is partly because individuals with psychosis show high rates of mood
symptoms which may or may not overlap with non-affective psychosis at different times (Malaspina
et al., 2013). To address this overlap, schizoaffective disorder was first introduced to the DSM in
1980 to serve as an intermediary diagnosis between schizophrenia and bipolar disorder. However, the
reliability and stability of schizoaffective disorder over time is poor, meaning that diagnostic
boundaries between these disorders are still often blurred in clinical practice. In fact, there is
increasing evidence that these disorders are not only similar in symptoms, course, and prognosis, but
are also underlain by a shared polygenic vulnerability (Purcell et al., 2009). An alternative
conceptualisation of these disorders is that they represent a spectrum of psychosis, rather than discrete
nosological categories. This approach shifts the focus of study from diagnostic categories to specific
symptoms, such as depression or hallucinations, and how these relate to neurobiological and
behavioural markers. Although some authors have questioned the utility of applying a dimensional
scale to clinical practice (e.g.: Lawrie et al., 2010) , this approach allows researchers to identify
specific patterns of symptomatology and deficit from an otherwise highly heterogeneous population.
A particular area of deficit in psychosis that has gained interest in recent years is social
cognition. Social cognition broadly includes abilities such as reacting emotionally to others,
experiencing empathy, and inferring the emotions of others (Green, Horan, & Lee, 2015). One critical
aspect of social cognition is the ability to extract emotional cues from faces. Interchangeably referred
to as facial affect recognition, expression recognition, and facial emotion processing, this ability has
been studied extensively in schizophrenia, and to a lesser degree in other psychotic disorders.
Research indicates that facial affect recognition is strongly associated with social, occupational, and
community functioning in patients with schizophrenia, and may also mediate the relationship between
social functioning and broader neurocognitive deficits (Irani, Seligman, Kamath, Kohler, & Gur,
12
2012; Meyer & Kurtz, 2009; Tsunoda et al., 2012). Furthermore, deficits in facial affect recognition
may predict the conversion to schizophrenia in clinically high-risk populations, suggesting that these
deficits precede the formal onset of psychosis (Corcoran et al., 2015).
The exact mechanics that may underlie impairments in facial emotion processing are an
ongoing source of debate. However, there is consensus that individuals with schizophrenia are
significantly impaired in their ability to recognise emotions compared to healthy controls and certain
other psychiatric disorders (Kohler, Walker, Martin, Healey, & Moberg, 2010; Ventura, Wood,
Jimenez, & Hellemann, 2013). Meta-analyses estimate these effect sizes to be quite large, e.g.: d =
−.91, d = −.85, and g = .89 respectively (Chan, Li, Cheung, & Gong, 2010; Kohler et al., 2010; Savla,
Vella, Armstrong, Penn, & Twamley, 2013). Deficits in emotion recognition have been demonstrated
reliably across a wide range of behavioural studies. Furthermore, there is a mounting body of
evidence that individuals with schizophrenia show differences from healthy controls using other
methodologies such as analysis of visual scanpaths during face viewing, neuroimaging, and
electrophysiological studies.
It remains unclear whether these emotion processing deficits reported in schizophrenia are
indeed specific to facial expressions, or whether they could be due to more general impairments in
processing non-emotional faces, or processing visual stimuli more generally. For instance, there is
evidence to suggest that patients with schizophrenia may have impairments in recognising the identity
of a face (i.e.: non-affective face processing). Like emotion-processing, the ability to recognise
identity is a crucial aspect of social cognition, and deficits may contribute to poorer functional
outcome in this disorder (Chen, Norton, McBain, Ongur, & Heckers, 2009). As emotion-processing
and identity-processing are believed to be underlain by different neural routes, the comparison of
these abilities allows us to better characterise these impairments and their relevant pathophysiology.
Through improved characterisation of impairment, it is hoped that we can eventually improve
interventions aimed at remediating these deficits in individuals before the formal onset of psychotic
illness.
The goal of this thesis, therefore, is to better understand facial affect processing impairments
in schizophrenia by addressing the following research questions:
I. Can we better characterise face processing deficits in schizophrenia? i.e.: Can impairments in
emotion-processing be explained by more general deficits in non-emotional face processing or
non-face processing?
II. Are face processing impairments specific to schizophrenia, or are they shared by other psychiatric
disorders on the psychosis continuum?
III. Do specific symptoms correlate with face-processing impairments?
13
These questions will be explored through experiments using a set of original dynamic (video-
based) tasks to better overcome some of the methodological issues raised in previous studies.
Outline of chapters
Chapter 2 of this thesis will begin by introducing the dominant theories of face-processing in
healthy individuals. It will then review behavioural, imaging, and ERP research investigating
emotion-processing impairments in schizophrenia. Chapter 3 will review studies examining non-
emotional face processing in schizophrenia and presents the argument that putative ‘emotion-specific’
deficits may not be so specific after all. Chapter 4 will present the rationale and aims of the
experiments that were conducted for this thesis. Chapter 5 will discuss methodological issues raised
in previous research and outline the development of a novel set of video-based tasks for the
assessment of face processing in schizophrenia. It will then present the results of an experiment
evaluating the efficacy of these video-based tasks in healthy controls. Chapter 6 will introduce the
concept of schizotypy (schizophrenia-like traits) and discuss how these experiences related to task
performance in the previous healthy control study. Chapter 7 will present the methods and initial
results for a large inpatient study examining face-processing in psychiatric inpatients with
schizophrenia, bipolar disorder, other forms of psychosis, and non-psychotic disorders. Chapter 8
will discuss selected results with reference to the first aim of the study: to better characterise face-
processing deficits in schizophrenia using a range of dynamic tasks. Chapter 9 will present the results
of an additional inpatient study designed to explore the possible relationship between emotion-
processing deficits and visuospatial attention in schizophrenia. In Chapter 10, results of the larger
inpatient study will be discussed with reference to the second aim: to determine whether face-
processing deficits are specific to schizophrenia or shared across other psychiatric disorders with and
without psychotic features. Chapter 11 will present a review of research examining associations
between symptomatology and face-processing impairments in schizophrenia. It will also discuss the
results of correlational analyses between symptoms and task performance in the current study. Finally,
the general discussion and practical implications of this thesis will be presented in Chapter 12.
14
Chapter 2: Face Processing in Healthy versus Schizophrenia
Populations
The perception of human faces is arguably one of the most complex and specialized visual
processes that we possess. As a critical aspect of social communication, faces are a unique stimulus
class that bridges the gap between visual and social cognitive sciences. Healthy individuals are able to
extract a wealth of information from faces – such as identity, sex, attractiveness, emotion, and focus
of attention – with almost effortless ease. This chapter will outline the dominant theories of face
processing in healthy adults, highlighting the neural and behavioural dissociation between emotional
(expression) and non-emotional (facial identity) face information (See Figure 2.1). It will then present
a broad review of emotion-processing impairments in schizophrenia, drawing on research from
behavioural, visual scanpath, functional imaging, and electrophysiological studies.
Face processing in the healthy brain
Bruce & Young’s theory of face recognition (1986)
An early iteration of modern theories of face processing was produced by Bruce and Young
(1986), who proposed that different types of information are extracted from a face via both
hierarchical and parallel functions (see Figure 2.2). The first step involves structural encoding, which
involves extracting two types of information from a face: view-centred descriptions such as transient
facial movements and expressions, and expression-independent descriptions which are more abstract,
view-independent information, such as the distance between features. After encoding, this information
then feeds into three parallel routes: the face-recognition route (i.e.: processing the identity of a face)
which relies on expression-independent information, and the expression-analysis and facial speech
analysis routes which both rely on view-centred descriptions. A fourth route, directed visual
processing, describes strategic goal-directed visual processing, such as searching for a friend’s face in
a crowd. In addition to making the critical distinction between processing the identity of a face and
processing emotional expressions, Bruce and Young’s (1986) theory also included a framework for
recognising familiar faces. This hierarchical process involved first the activation of Facial
Recognition Units (triggering a sense of familiarity), then Person Identity Nodes (accessing memories
about the individual) and finally Name Generation (accessing the specific name of the individual).
The distinction between Person Identity Nodes and Name Generation aims to explain the common
‘tip-of-the-tongue phenomenon’ where individuals can recall information about a person without
being able to recall their name (e.g. “that’s whats-his-name who taught my psychology course last
year”). Bruce and Young’s functional theory was compatible with a range of psychological research,
including behavioural studies, observation of everyday errors, and studies of cerebral injury (Hanley,
15
2011). However, this theory made no assumptions about the anatomical correlates that may underpin
these functional routes.
Figure 2.1. A visual representation showing the separation of emotion and identity processing during
face perception.
16
Figure 2.2. Bruce and Young’s theory of face perception. Image adapted from Bruce and Young
(1986, p.312).
The distributed neural system for face perception
Bruce and Young’s (1986) theoretical model was later expanded upon by Haxby and colleagues
(2000), who proposed the distributed neural model or dual-route model of face perception. This
model proposed a neuroanatomical framework for face recognition derived from neuroimaging
research. This framework consists of a Core System of extrastriate areas where visual identity and
non-identity information are analysed separately, then fed into a shared Extended System (see Figure
2.2). Analogous to Bruce and Young’s (1986) expression-independent and view-centred descriptions,
Haxby and colleagues distinguish between two types of information: invariant and changeable.
Invariant information refers to properties that are consistent across different views and facial
expressions, and are necessary for recognising the identity of a face. Changeable information includes
eye gaze, expression, and movements of the eyes and mouth, and is necessary for the recognition of
facial affect. At the level of the Core System, the most basic processing of facial features is thought to
be mediated by neurons in the inferior occipital gyri (the Occipital Face Area; Haxby & Gobbini,
2011). From here, the analysis of changeable visual features is processed largely via a route involving
the posterior superior temporal sulcus (pSTS). In contrast, invariant information is processed via a
ventral temporal route which is thought to include the inferior occipital and fusiform gyri, including
the Fusiform Face Area (FFA). These two routes then feed into the Extended System for higher level
interpretation, which is composed of structures outside the visual extrastriate cortex. This includes
functions such as retrieval of personal knowledge (associated with the medial prefrontal cortex,
anterior temporal cortex, and posterior cingulate gyrus, among others), motor simulation of facial
17
expressions (associated with putative mirror neurons in the inferior parietal lobe and frontal
operculum) and emotion recognition (associated with the limbic system).
Figure 2.3. The distributed neural system for face perception. Image adapted from Haxby and
Gobbini (2011, p.105).
Evidence in support of this model, with particular attention to the separation of identity and
expression recognition, is discussed at length elsewhere (see Calder & Young, 2005, for a helpful
review). Briefly, this support comes from a broad range of sources including behavioural research
(Bruce, 1986; Calder, Young, Keane, & Dean, 2000; Campbell, Brooks, de Haan, & Roberts, 1996)
functional imaging studies (George et al., 1993; Sergent, Ohta, Macdonald, & Zuck, 1994; Winston,
Henson, Fine-Goulden, & Dolan, 2004), and in dissociations exhibited by brain injured individuals
(Hornak, Rolls, & Wade, 1996; Tranel, Damasio, & Damasio, 1988; Young, Newcombe, de Haan,
Small, & Hay, 1993). For instance, it has been demonstrated in healthy participants that the familiarity
of a face does not affect reaction times during an expression recognition task, and vice versa (e.g.:
Calder et al., 2000; Campbell, et al., 1996). Accounts of individuals with prosopagnosia – severe
deficits in recognising the identity of familiar faces, despite intact low-level vision and general
cognitive function – have described preserved ability to recognise facial expressions despite impaired
identity recognition (Baudouin & Humphreys, 2006; Riddoch, Johnston, Bracewell, Boutsen, &
Humphreys, 2008).
18
Compelling evidence for this dissociation has also come from studies of nonhuman primates.
In humans, fMRI has revealed several specialised regions in the temporal lobe that respond much
more strongly to faces than to other objects (Kanwisher, McDermott, & Chun, 1997; Puce, Allison,
Asgari, Gore, & McCarthy, 1996). Similar regions have been found in the temporal lobes of macaques
(Pinsk, DeSimone, Moore, Gross, & Kastner, 2005; Rajimehr, Young, & Tootell, 2009; Tsao,
Freiwald, Knutsen, Mandeville, & Tootell, 2003). In one study, Tsao and colleagues (2008) scanned
humans and macaques while showing images of faces and nonface objects. In macaques, six ‘patches’
of face-selective cortex were identified in the temporal lobe. The locations of these patches were
consistent across 9 of the 10 animals scanned. In humans, the number of these patches in the temporal
lobe varied between three and five. These were found primarily in the occipital face area (Kanwisher
& Yovel, 2006), the superior temporal sulcus (Hoffman & Haxby, 2000), the fusiform face area
(Kanwisher et al., 1997) and one or two regions in the anterior collateral sulcus (Tsao, Moeller, &
Freiwald, 2008). In addition to the temporal lobe face patches, Tsao and colleagues (2008) also
identified two face-selective patches in the macaque frontal lobe. Face-responsive regions have also
been found in the human frontal lobe, particularly in the orbitofrontal cortex (Haxby et al., 1996). The
orbitofrontal cortex has been shown to play a role in the regulation of mood, emotion processing, and
social reinforcement (Bechara, Damasio, & Damasio, 2000), raising the possibility that this brain
region may be selective to facial expressions. Tsao and colleagues (2008) found that when neutral and
expressive faces were shown to macaques, the frontal lobe face patches responded much more
strongly to expressive faces than neutral faces. Conversely, the responses of the temporal lobe patches
were only weakly modulated by expression. This result suggests that in macaques, frontal lobe face
patches are involved in facial expression recognition, while the temporal lobe patches are involved in
non-expressive face processing, i.e. identity recognition.
While it has been established that identity and expression recognition involve separate
processes, the level at which this bifurcation occurs, and the degree to which these parallel processes
interact, are yet to be resolved. A convincing double dissociation between identity and expression
recognition is yet to be demonstrated, and while evidence for some dissociation exists, it is not clear
whether this represents a separation at the visuoperceptual level or at a higher stage of processing
(Calder, 2011). Furthermore, it is likely that some degree of crossover exists at the Core level. For
example, both changeable and invariant cues may be used to identify an individual based on
characteristic facial expressions or face movements (Fitousi & Wenger, 2013). Considering the scope
of this PhD, it will adopt the prevailing view that facial identity and facial expression recognition are
largely, but not entirely dissociable processes that nevertheless provide useful information about the
relative health of two different but interrelated neural routes.
19
Summary
The dominant theory in the face literature is that the two major components of face processing
– identity recognition and emotion recognition – represent largely independent processes. This view
reflects the two most prominent and widely-accepted models of face perception. The first is Bruce and
Young's (1986) classic functional model, which asserts that visual information pertaining to identity
and information pertaining to emotion recognition and speech are processed by two separate routes.
The second is Haxby and colleagues' (2000) dual-route model, which provides a neuroanatomical
framework to support the separation between identity recognition, emotion recognition, and other
face-related information, such as visual processing of speech. While the precise level at which
identity- and emotion-processing routes bifurcate remains an issue of contention, there is
overwhelming evidence from behavioural, neuroimaging, nonhuman primate, and brain-injury studies
to support the division of these neural functions.
The universality of facial emotions
The study of facial emotion recognition has for the most part focused on what is referred to as
the six basic emotions: happiness, sadness, fear, disgust, anger, and surprise (Kohler & Martin, 2006).
While there remains some debate as to whether additional emotions (such as contempt) should also be
included in this list, there is strong evidence to suggest that these six emotions are biologically
determined, rather than learned through social interactions.
The idea that emotional expressions may be innate was first attributed to Charles Darwin.
Upon observing similarities in facial expressions shared by phylogenetically similar species (such as
different primates) Darwin suggested that expressions are an evolutionary adaptation that is
biologically determined (Darwin, 1872, reprinted with commentary in Darwin, Ekman, & Prodger,
1998) . A century later, this idea was revisited in a series of studies designed to examine the
‘universality’ of emotional expressions across different cultures. Revived by three independent
researchers, Tomkin, Ekman and Izard, these studies revealed high consistency in both the expression
and recognition of six primary emotions: joy, sadness, fear, disgust, anger and surprise. This was
demonstrated across a range of literate cultures including Japan, Brazil, USA, Chile and Argentina
(Ekman & Friesen, 1971; Ekman et al., 1987; Izard, 1971; Tomkins & McCarter, 1964), as well as in
a socially isolated, pre-literate cultural group in Papua New Guinea (Ekman & Friesen, 1971; Ekman,
Sorenson, & Friesen, 1969). In the decades since, more than 30 studies have replicated the finding
that emotional expressions are universally recognised across cultures (see Matsumoto, 2001) , and
over 75 studies have shown consistency in the way that these basic emotions are expressed in
response to emotion-eliciting stimuli, such as films (Matsumoto, Keltner, Shiota, Frank, & O'Sullivan,
2008).
20
Other support for the idea that certain emotional expressions are biologically determined
comes from studies of congenitally blind individuals, who also produce the same six facial
expressions despite lacking the ability to perceive others’ expressions visually (Galati, Sini, Schmidt,
& Tinti, 2003; Matsumoto & Willingham, 2009). Considered as a whole, the literature strongly
suggests that the six basic emotional expressions are hard-wired in humans, therefore major
impairment in the identification and expression of these emotions is likely to be underpinned by
neural differences, rather than differences in culture or socialisation.
Behavioural studies of emotion processing in schizophrenia
The most widely used method of evaluating facial emotion processing in schizophrenia is
through behavioural tasks that involve viewing static images of faces. There is an extensive body of
research demonstrating that patients with schizophrenia perform more poorly on expression
processing tasks compared to healthy controls (see Kohler et al., 2010 for a meta-analytic review) as
well as other psychiatric populations such as patients with depression (Bediou, Krolak-Salmon, et al.,
2005; Weniger, Lange, Ruther, & Irle, 2004), schizoaffective disorder (Chen, Cataldo, Norton, &
Ongur, 2012) and affective psychoses (Edwards, Pattison, Jackson, & Wales, 2001). This deficit has
been demonstrated across a wide variety of tasks. These tasks can be broadly divided into two
categories: recognition/labelling tasks, and discrimination tasks. Recognition/labelling tasks involve
selecting the most appropriate label for a given stimulus. Figure 2.4A shows an example which
requires the participant to decide whether each face more closely resembles happiness, sadness, anger,
fear or disgust. In contrast, discrimination tasks involve viewing pairs of faces and making some sort
of judgement, such as deciding whether they show the same or different expression, or deciding which
expression has greater intensity (see Figure 2.4B and C).
The benefit of discrimination tasks is that, unlike labelling tasks, they do not require
participants to explicitly identify the expression shown; they are simply identifying whether the
expressions are the same or not. For this reason they are thought to be less affected by difficulties with
language or semantic retrieval (Macmillan & Creelman, 1991). However, because discrimination
tasks require comparison between two faces – either two presented simultaneously, or serially,
meaning that the first face must be held in iconic memory – they arguably place a greater load on
working memory compared to labelling tasks (Macmillan & Creelman, 1991). This is an important
consideration when assessing clinical samples such as schizophrenia, as working memory deficits are
increasingly recognised as a central feature of this disorder (Forbes, Carrick, McIntosh, & Lawrie,
2009). In recognition of the differing demands of these two types of tasks, many researchers include
both labelling and discrimination tasks in their studies to account for the possibility that they may be
sensitive to different impairments (Kohler et al., 2010). For example, one meta-analysis reported that
performance on only labelling tasks, not discrimination tasks, significantly predicted functional
21
outcomes in schizophrenia, although it is possible that this reflected a lack of power in the analyses
(Irani et al., 2012). Nevertheless, significant emotion perception impairments in schizophrenia have
been repeatedly and reliably demonstrated across a broad range of both labelling (Bediou et al., 2007;
Fakra, Salgado-Pineda, Delaveau, Hariri, & Blin, 2008; Kucharska-Pietura, David, Masiak, &
Phillips, 2005; Martin, Slessor, Allen, Phillips, & Darling, 2012; Penn et al., 2000; Whittaker, Deakin,
& Tomenson, 2001) and discrimination tasks (Addington, Saeedi, & Addington, 2006; Martin,
Baudouin, Tiberghien, & Franck, 2005; Weniger et al., 2004). Together, these results suggest that
these deficits exist even when memory and language demands are minimised.
Figure 2.4. Examples of labelling and discrimination tasks. A: A typical labelling task using Ekman
face stimuli (Ekman & Friesen, 1976). B: An example same-or-different task, also using Ekman
stimuli. C: An intensity discrimination task created by Bediou and colleagues (2005, p.528).
Emotional valence in behavioural studies
Overall, there is a consensus that individuals with schizophrenia show impairment in the
perception of affect in static facial expressions. However, a number of issues remain. One is whether
this impairment may be specific to different emotions. In general, studies have found stronger
evidence for impairment in recognising negative emotions such as fear, disgust, anger and sadness
(Bediou, Franck, et al., 2005; Brune, 2005; Comparelli et al., 2014; Edwards et al., 2001; Kohler et
al., 2003; Lahera et al., 2014; Pinkham, Penn, Perkins, & Lieberman, 2003). In contrast, these studies
have reported either unimpaired, or less impaired processing of positive emotions such as happiness.
It is possible that this distinction indicates differential processing of positive and negative emotions in
22
schizophrenia. However, authors have also pointed out that this finding may simply reflect inherent
differences in the difficulty of identifying these emotions, as even healthy controls typically find
happy faces to be the easiest emotion to identify and fearful faces the most difficult (Marwick & Hall,
2008).
Another interesting trend in the behavioural literature is that patients with schizophrenia also
show impairments in correctly identifying neutral facial expressions. For instance, Kohler and
colleagues (2003) noted that patients with schizophrenia were more likely to mislabel neutral faces as
“fearful” or “disgusted” compared to healthy controls. According to another study, this tendency to
miscategorise neutral faces appears to be more prominent in schizophrenia patients with paranoid
symptoms compared to those without (Pinkham, Brensinger, Kohler, Gur, & Gur, 2011). These
authors found that paranoid patients were significantly more likely to attribute neutral faces as
“angry”, despite performing similarly to non-paranoid patients on all other conditions. It was
suggested that this finding may reflect the broader tendency to misattribute salience to ambiguous
stimuli in schizophrenia, with paranoid patients in particular showing heightened sensitivity to
potentially threatening stimuli (Holt, Titone, et al., 2006). Furthermore, this implies that
schizophrenia is associated with difficulty distinguishing between socially-salient emotional
information and non-salient information, coupled with a tendency to assign meaning to ambiguous or
unimportant details (Phillips, 2011).
Visual scanpath studies in schizophrenia
Another approach to studying facial affect processing is through the recording of visual
scanpaths. A visual scanpath refers to the spatial sequence of eye gaze fixations and saccades
(movements) that occur when viewing a scene or object (Noton & Stark, 1971). This is thought to
represent the real-time allocation of visual attention required for encoding visual information.
Typically, visual scanpaths are measured using paradigms that record eye movements while faces are
shown under free-viewing conditions. When viewing faces, healthy individuals show a characteristic
pattern of fixations across the face, with particular focus on salient features such as the eyes, nose, and
mouth (Radua et al., 2010). In contrast, individuals with schizophrenia tend to show a comparatively
restricted scanning pattern characterised by shorter saccades, more time spent on each fixation, and a
tendency to avoid salient facial features (Toh, Rossell, & Castle, 2011). This pattern has been
demonstrated reliably in studies using neutral faces (Gordon et al., 1992; Loughland, Williams, &
Gordon, 2002b; Manor et al., 1999; Phillips & David, 1998; Rosse, Schwartz, Johri, & Deutsch, 1998;
Williams, Loughland, Gordon, & Davidson, 1999) as well as across the six universal emotions
(Bestelmeyer et al., 2006; Delerue, Laprevote, Verfaillie, & Boucart, 2010; Green & Phillips, 2004;
Green, Williams, & Davidson, 2003; Loughland, Williams, & Gordon, 2002a; Shimizu et al., 2000;
Streit, Wolwer, & Gaebel, 1997). This pattern also corresponds to poorer accuracy for recognising
23
emotions in a behavioural task (Williams et al., 1999), and is more pronounced in schizophrenia
compared to affective disorders (Loughland et al., 2002a).
Importantly, this pattern of fewer, longer fixations is also shown when viewing non-face
stimuli such as complex scenes, geometric shapes, and ambiguous stimuli (Benson, Leonards,
Lothian, St Clair, & Merlo, 2007; Bestelmeyer et al., 2006; Green, Waldron, Simpson, & Coltheart,
2008; Hori, Fukuzako, Sugimoto, & Takigawa, 2002; Obayashi et al., 2001), suggesting that this
scanning strategy is not simply in reaction to the class of object being viewed (Beedie, Benson, & St
Clair, 2011). However, the tendency to avoid salient facial features such as the eyes is also shared by
individuals with autism and social phobia (Horley, Williams, Gonsalvez, & Gordon, 2003; Radua et
al., 2010). It has been suggested patients with these disorders may find faces (especially eyes)
particularly anxiety-inducing and simply prefer to avoid attending to these features (Morris, Weickert,
& Loughland, 2009). For instance, it has been shown that people with schizophrenia choose to stand
further away from emotional faces compared to healthy controls (Mandal, Pandey, & Prasad, 1998).
Does this mean that the avoidance of salient facial features is simply an individual preference? The
majority of past research examined scanpaths under passive viewing conditions, meaning that
participants are requested to look at the face without performing any specific judgements. To
determine whether scanpath abnormalities persist in active viewing conditions, Delerue and
colleagues (2010) examined the visual scanpaths of 20 patients with schizophrenia while completing a
variety of active tasks. These included judging the age or sex of the face, identifying the emotion
shown, and deciding if the face is familiar on unfamiliar. Interestingly, they found that although the
patients with schizophrenia showed the typical restricted gaze pattern when viewing faces passively,
there was no significant difference in gaze pattern between healthy controls and patients on the four
active tasks. That is, patients were no longer avoiding salient features when they were explicitly asked
to make judgements about the face they are viewing. These results suggest that scanpath abnormalities
can be eliminated when task demands require active attention, and are unlikely to account for the
facial processing deficits identified on active behavioural tasks (Delerue et al., 2010).
In conclusion, visual scanpath abnormalities under passive viewing conditions are a persistent
feature of schizophrenia. However, patients show scanpaths that are identical to healthy controls when
performing active behavioural tasks, such as judging the gender of a face. Therefore, abnormalities in
visual scanning are unlikely to explain the pronounced facial emotion processing impairments shown
in schizophrenia across different behavioural tasks in the literature. It is possible that anxiety relating
to eye contact with faces may account for the unusual scanpath patterns produced in studies using
passive (undirected) viewing conditions.
24
Neuroimaging studies in schizophrenia
The neural network activated in healthy adults when viewing emotional faces has been well
established in the literature (Kret & Ploeger, 2015). This network has been shown to include the
amygdala (Haxby et al., 2000) fusiform face area (Kanwisher et al., 1997), occipital face area
(Gauthier et al., 2000), basal ganglia, and regions of the medial and inferior frontal cortex, such as the
anterior cingulate cortex (Morris et al., 2009). In particular, the amygdala appears to play a key role in
modulating the activity of the fusiform face area when viewing emotional faces, especially fearful
expressions (Adams, Gordon, Baird, Ambady, & Kleck, 2003; Demos, Kelley, Ryan, Davis, &
Whalen, 2008; Rotshtein, Malach, Hadar, Graif, & Hendler, 2001; Vuilleumier & Pourtois, 2007).
Accordingly, individuals with bilateral amygdala damage show particular impairment in recognising
fear from faces, and show reduced attention to the eye region (Adolphs et al., 2005; Calder, 1996;
Spezio, Huang, Castelli, & Adolphs, 2007).
Schizophrenia is associated with both structural and functional abnormalities of the amygdala.
On average, the amygdala of individuals with schizophrenia are 6% smaller in volume compared to
healthy controls (Wright et al., 2000). Several meta-analyses of functional imaging studies in
schizophrenia have shown reduced activation of the bilateral amygdala when viewing emotional faces
compared to healthy controls (Anticevic et al., 2012; Li, Chan, McAlonan, & Gong, 2010; Taylor et
al., 2012). Studies suggest that this hypoactivation is particularly pronounced when viewing fearful
faces (R. E. Gur et al., 2002; Michalopoulou et al., 2008). However, these studies typically report a
reduction in amygdala activation compared to activation for neutral faces (Morris et al., 2009). Other
studies suggest that patients instead show increased amygdala activation for neutral faces compared to
healthy controls (Habel et al., 2010; Hall et al., 2008; Holt, Kunkel, et al., 2006; Surguladze et al.,
2006). Therefore, it remains unclear whether this difference indicates true hypoactivation during
fearful face processing, or is simply due to hyperactivation to the comparison stimulus. It is possible
that this increased activation to neutral faces reflects the tendency of patients to mistakenly ascribe
emotional significance to non-emotional stimuli (Mothersill et al., 2014). In further support of this
idea, several studies using non-face baseline comparisons have reported hyperactivation of the
amygdala when viewing emotional faces (Holt, Kunkel, et al., 2006; Kosaka et al., 2002; Rauch et al.,
2010), or have failed to find differences in activation altogether (Holt et al., 2005; Sachs et al., 2012).
It is also possible that these discrepancies represent differences in tasks used (e.g.: passive viewing
versus activate emotion discrimination), differences in analytic techniques used, or heterogeneity in
patient clinical symptoms (Pankow et al., 2013).
Schizophrenia has also been associated with abnormalities in other brain regions involved in
emotion processing. For instance, the insular cortex and basal ganglia are shown to play an important
role in recognising expressions of disgust in healthy controls (Phan, Wager, Taylor, & Liberzon,
2002), and studies indicate that both of these regions have reduced volume in patients with
25
schizophrenia (Honea, Crow, Passingham, & Mackay, 2005; Shenton, Dickey, Frumin, & McCarley,
2001). Furthermore, the insula and basal ganglia also show reduced blood flow in patients compared
to healthy controls while viewing expressions of disgust (Phillips et al., 1999). Similarly, structural
abnormalities in the fusiform face area have been reported in schizophrenia (Lawrence, Calder,
McGowan, & Grasby, 2002), which correspond to reduced activation of this area when viewing
emotional faces (Li et al., 2010; Maher, Ekstrom, Holt, Ongur, & Chen, 2016; Taylor et al., 2012).
Still other regions, including the hippocampal and parahippocampal areas, appear to show increased
activity in response to both emotional (Holt, Kunkel, et al., 2006; Holt et al., 2005; Russell et al.,
2007) and neutral faces (Surguladze et al., 2006) in patients with schizophrenia compared to controls.
Another key brain region implicated in emotion processing is the anterior cingulate cortex
(ACC). The ACC is a region of the prefrontal cortex that is shown to be activated during emotion
perception tasks (Kober et al., 2008). It is believed to play an important role in coordinating emotional
responses to salient environmental stimuli, and is involved in emotional control, motivation, arousal,
and performance monitoring (Kennerley, Walton, Behrens, Buckley, & Rushworth, 2006; Patterson,
Ungerleider, & Bandettini, 2002; Ridderinkhof, Ullsperger, Crone, & Nieuwenhuis, 2004). While
some studies report decreased activity of the ACC in schizophrenia when viewing emotional faces
(Habel et al., 2010; Hempel, Hempel, Schonknecht, Stippich, & Schroder, 2003), others report
overactivity, particularly in response to neutral faces (Hall et al., 2008). Like the amygdala, it has
been suggested that activation of this region may reflect inappropriate attribution of salience to
ambiguous or neutral face stimuli (Mothersill et al., 2014). Dysfunction of this region is also
supported by structural (Baiano et al., 2007; Fornito, Yucel, Patti, Wood, & Pantelis, 2009) and post-
mortem (Benes, 1998) findings that indicate abnormal volumes in individuals with schizophrenia.
Exploring these abnormalities from a different angle, several authors have examined patterns
of neural connectivity in schizophrenia while processing emotional faces. Fakra and colleagues (2008)
found that, on an emotion-matching task, patients with schizophrenia did not show the typical pattern
of functional connectivity between the amygdala and prefrontal cortex as seen in healthy controls. The
authors suggest that this may reflect a tendency to rely on more cognitive strategies to perform this
task, rather than engaging the usual limbic structures. Similarly, other studies have reported patterns
of reduced connectivity between the amygdala and the fusiform gyrus (Hall et al., 2008) and between
the amygdala and regions of the parietal lobe and precuneus (Mukherjee et al., 2012). Importantly,
these regions are shown to be involved in emotion regulation in healthy controls (Mak, Hu, Zhang,
Xiao, & Lee, 2009). Furthermore, one study reported increased activation of these areas in
schizophrenia when emotional load interferes with performance on a working memory task
suggesting that patients are disproportionately recruiting these areas to compensate for abnormal
functioning of emotion processing networks (Becerril & Barch, 2011).
Taken together, functional imaging studies indicate that schizophrenia is associated with
abnormal activation of regions of the temperomedial lobe that are typically involved in emotion
26
processing, including the amygdala, insula, hippocampal and parahippocampal gyri. This includes
patterns of hypoactivation to emotional stimuli as well as hyperactivation to ambiguous or
emotionally-neutral stimuli, suggesting that this dysfunctional activation may reflect a tendency to
misattribute salience to emotional stimuli in schizophrenia. In addition, studies have also reported
altered activation and connectivity of prefrontal areas that are involved in cognitive control, such as
the anterior cingulate cortex. It has been suggested that unusual activation of prefrontal regions may
be indicative of compensatory cognitive strategies to actively overcome deficits in emotion
processing.
EEG studies in schizophrenia
Event-related potential (ERP) studies have identified several electrophysiological markers
that may be linked to face processing deficits in schizophrenia. The most widely studied of these is
the N170, an ERP component that peaks negatively at 170 milliseconds after presentation of a
stimulus (Bentin, Allison, Puce, Perez, & McCarthy, 1996). The N170 is found primarily at
occipitotemporal sites in the right hemisphere, and is thought to reflect activity of the fusiform face
area (Deffke et al., 2007), occipital cortex (Herrmann, Ehlis, Muehlberger, & Fallgatter, 2005; Utama,
Takemoto, Koike, & Nakamura, 2009), and superior temporal gyrus (Nguyen & Cunnington, 2014).
In healthy individuals, this waveform is significantly larger in response to faces and face-related
stimuli such as line drawings of faces, individual facial features, inverted faces and distorted faces
compared to other visual stimuli (Bentin et al., 1996; Rossion & Caharel, 2011; Rossion et al., 1999;
Wynn, Lee, Horan, & Green, 2008). For this reason, it has been suggested that the N170 represents
face detection: the early categorisation of faces as a perceptual category, rather than processing of
specific aspects such as facial identity or expression (Bentin et al., 1996; Maher, Mashhoon, Ekstrom,
Lukas, & Chen, 2016). A meta-analysis by McCleerey and colleagues (2015) noted that of 21
identified studies examining N170 amplitudes in schizophrenia, 15 reported significant deficits in
patients. Specifically, this meta-analysis reported a mean effect size of .64 (a medium effect) with
patients showing reduced N170 waveforms in response to face stimuli across a range of tasks
including both passive and active viewing (McCleery et al., 2015).
Another ERP implicated in face processing is the N250, a negative waveform occurring 250
milliseconds after stimulus presentation that is observed at fronto-central sites (Brenner, Rumak, &
Burns, 2016). This waveform is thought to represent the accumulation and integration of face-related
information from different regions, including basic structural information associated with the N170
(Streit, Wolwer, Brinkmeyer, Ihl, & Gaebel, 2000; Wynn, Jahshan, Altshuler, Glahn, & Green, 2013).
In healthy controls, the N250 is shown to be modulated by properties such as familiarity (Pierce et al.,
2011; Tanaka, Curran, Porterfield, & Collins, 2006) and facial affect (Streit, Wolwer, Brinkmeyer,
Ihl, & Gaebel, 2001), suggesting that this ERP may be involved in processing more complex face
27
information such as identity, gender, and emotional expression. Deficits in the N250 are not as well
documented in schizophrenia compared to the N170. While several studies have found reduced
amplitude of the N250 in response to faces (Streit et al., 2001; Wynn et al., 2013; Wynn et al., 2008),
others reported intact N250 waveforms compared to controls (Jung, Kim, Kim, Im, & Lee, 2012; Lee,
Kim, Kim, & Bae, 2010; Turetsky et al., 2007). However, the previously cited meta-analysis by
McCleery and colleagues (2015) indicated significant reductions in the N250 in patients overall, with
a mean effect size of 0.49 (a medium effect size). Taken together, the authors suggest that the
reduction of both N170 and N250 waveforms in schizophrenia signify impairments in the early stages
of visual processing of faces.
In addition to specific time-locked waveforms such as the N170, researchers have also
examined links between facial emotion processing and sustained activity in the theta frequency range
(4-7 Hz). For instance, healthy controls tend to show increased theta oscillations when viewing
emotional faces compared to neutral faces (Aftanas, Varlamov, Pavlov, Makhnev, & Reva, 2001;
Balconi & Pozzoli, 2008; Knyazev, Slobodskoj-Plusnin, & Bocharov, 2009). However, this activity
may reflect greater neural recruitment in response to biologically meaningful stimuli, rather than a
purely emotion-specific process (Krombholz, Schaefer, & Boucsein, 2007; Palermo & Rhodes, 2007).
In addition to emotion processing, theta oscillations have been implicated in other complex cognitive
processes including maintenance of information in memory (Lee, Simpson, Logothetis, & Rainer,
2005; Reinhart et al., 2012; Tsoneva, Baldo, Lema, & Garcia-Molina, 2011), as well as focussed
attention under low working memory load (Chang & Huang, 2012). It has also been suggested that
increases in theta activity may represent different functions at different time points, with oscillations
within 500 milliseconds of stimulus presentation indicating directed attention and early visual
processing, and oscillations after 950 milliseconds indicating maintenance in working memory
(Brenner et al., 2016; Deiber et al., 2007). In schizophrenia, reduced theta oscillations have been
demonstrated across various several face-related tasks (Missonnier et al., 2012; Schmiedt, Brand,
Hildebrandt, & Basar-Eroglu, 2005). One study found that individuals with schizophrenia showed
reduced theta activity despite intact behavioural performance on a facial emotion recognition task
(Ramos-Loyo, Gonzalez-Garrido, Sanchez-Loyo, Medina, & Basar-Eroglu, 2009). Another study
reported reduced theta activity (specifically between 140-200 ms) in schizophrenia, which was
correlated with performance deficits in an emotion recognition task (Csukly, Stefanics, Komlosi,
Czigler, & Czobor, 2014). Together, these studies suggest that early theta activity (overlapping with
the N170 and N250) following stimulus onset appears to be sensitive to emotion deficits in
schizophrenia. To examine the relationship between N170, N250 and P100 (an ERP associated with
early visual processing), and later-stage theta oscillations in schizophrenia, Brenner and colleagues
(2016) recorded electrophysiological activity during a same-or-different facial emotion discrimination
task. They found that, while both patients and healthy controls showed increased N170 peaks in
response to negative facial expressions compared to neutral, patients showed increased theta
28
oscillations in the inter-stimulus interval. Interestingly, increased theta oscillations during this period
predicted poorer behavioural performance for both patients and healthy controls. The authors propose
that this increased activity may reflect prolonged processing of emotional content that impedes task
performance in schizophrenia (Brenner et al., 2016).
In summary, electrophysiological studies have highlighted dysfunction at different stages of
face and emotion processing in schizophrenia. Evidence suggests that patients have reduced N170 and
N250 responses to emotional faces compared to healthy controls, indicating irregularity at the early
stages of perceptual processing. While these waveforms are stronger for emotional faces, this does
not rule out the possibility of a more generalised deficit in non-emotional face processing.
Furthermore, increased later-stage theta oscillations in response to emotional faces may indicate
dysfunction at the later stages of emotion processing as well.
29
Chapter 3: Is it really all about emotion? Non-emotional Face
Processing is (Maybe) Impaired in Schizophrenia
The previous chapter outlined a large body of literature indicating impaired emotion-
processing schizophrenia. However, modern theories of face-processing in healthy individuals tell us
that information relating to facial emotion is processed largely independently from information
relating to facial identity. This raises the question: Is identity-processing also impaired in
schizophrenia? Answering this question may allow us to determine in which stage of perceptual
processing these abnormalities lie. If found, these deficits challenge the idea that the ‘emotion-
processing’ deficit is truly specific to emotion. This chapter will review the extant research
investigating non-emotional face processing in schizophrenia.
Non-emotional face processing in schizophrenia
Deficits in non-emotional face processing have been explored in schizophrenia using a variety
of behavioural tasks. However, heterogeneity across tasks types and participants has produced mixed
results (see Table 3). Tasks that assess true identity processing are primarily matching tasks, where
the participant views static photographs of faces (with non-face identifying features removed, such as
hair and spectacles), either serially or concurrently. The participant must match the identity of the first
face to one of several options (Addington & Addington, 1998; Chen et al., 2012; Kucharska-Pietura et
al., 2005; Penn et al., 2000). The most commonly used face-matching task is the Benton Test of Facial
Recognition (Benton, 1983; see Figure 3A). Individuals with schizophrenia have shown impaired
performance on the Benton test in many (Addington & Addington, 1998; Evangeli & Broks, 2000;
Kucharska-Pietura et al., 2005; Soria Bauser et al., 2012; Whittaker et al., 2001), but not all studies
(Hall et al., 2004; Pomarol-Clotet et al., 2010; Sachse et al., 2014; Scholten, Aleman, Montagne, &
Kahn, 2005; van 't Wout et al., 2007). Four studies using face-matching tasks with similar properties
to the Benton – that is, matching one image to one of two or three options displayed simultaneously –
also reported no impairment compared to healthy controls (Kosmidis et al., 2007; Quintana et al.,
2011; Quintana, Wong, Ortiz-Portillo, Marder, & Mazziotta, 2003; Williams et al., 1999). One study
using an identity matching task with seven options reported impaired performance in patients with
schizophrenia, suggesting that increasing task difficulty is more likely to precipitate group differences
(Williams et al., 1999). However, another study using a similar 7-option matching design showed no
impairments in patients (Hooker & Park, 2002). Inconsistent findings have also been produced by
studies using a morphed identity-matching task, in which participants choose which of a pair of faces
of varying similarity matches a briefly presented target (see Figure 3B). Two studies using this task
reported significant impairments in patients with schizophrenia (Chen & Ekstrom, 2016; Chen,
McBain, & Norton, 2015), while one only approached a significant difference (Chen et al., 2009), and
30
two reported intact functioning (Chen et al., 2012; Norton, McBain, Holt, Ongur, & Chen, 2009). The
reason for these inconsistent findings is unclear. Given the relatively heterogeneous patient samples,
however, the influence of factors such as age, illness duration and gender may be worth exploring
through a formal meta-analysis in the future.
An alternative measure of identity recognition is the two-alternative forced-choice identity
discrimination paradigm, in which the participant judges whether two serially presented faces are the
same or different. Again, some studies reported significant impairment in schizophrenia (Butler et al.,
2008; Kim et al., 2010; Martin et al., 2005; Shin et al., 2008; Soria Bauser et al., 2012) while others
found no impairment relative to controls (Edwards et al., 2001; Fakra, Jouve, Guillaume, Azorin, &
Blin, 2015; Johnston et al., 2010; Soria Bauser et al., 2012). Another study requiring participants to
distinguish between photographs of two learned identities, person A and person B, also reported
significant impairment (Baudouin, Martin, Tiberghien, Verlut, & Franck, 2002). Again, the reason for
these inconsistent findings is unclear, but given that mean age and illness duration of the patients
(Kucharska-Pietura et al., 2005), and the stimuli used in these tasks (e.g.: Soria Bauser et al., 2012)
are known to influence performance, it is possible these factors contributed here. While it is outside
the scope of the current review it would be interesting to consider the relative severity of face
perception deficits in a non-clinical population of individuals high in psychosis-proneness and
whether these deficits predict transition to psychosis. While impairments have been seen in some non-
emotional aspects of face processing in these populations before (Poreh, Whitman, Weber, & Ross,
1994), they are not reported as frequently as deficits in face emotion perception (Germine & Hooker,
2011; Waldeck & Miller, 2000; Williams, Henry, & Green, 2007).
31
Figure 3. Examples of tasks used to assess the perception of identity. A: Plate from the Benton Facial
Recognition Test (Benton et al., 1983). Participants indicate which of the six images match the target.
(Published in Busigny & Rossion, 2010, p 969). B: Identity matching task used by Norton and
colleagues (2009). C: Example of morphed images ranging from ‘no sex’ (50% male, 50% female) to
100% male face (from Bediou et al., 2005, p 528). D: Examples of upright and inverted stimuli used
in Soria Bauser and colleagues (2012). E: Example of semi-successive frames of the point-light
displays (walking) used in Kim and colleagues (2005).
32
Table 3. Identity processing tasks used in schizophrenia research from the last 25 years. Results indicate performance of patients relative to controls.
Task Results References
Tasks Assessing Identity Discrimination
Benton Test of Facial Recognition
[see Figure 3A]
Impaired
Kerr & Neale, 19931; Borod et al., 19931; Bellack et al., 19962; Mueser et al., 19961; Salem et al. 19961;
Addington & Addington, 19981; Evangeli & Broks, 20001; Penn et al., 20001; Whittaker et al., 20012;
Hooker & Park, 20022; Kucharska-Pietura et al, 20052; Soria Bauser et al., 20123
NOT impaired Hall et al., 20041; Scholten, et al., 20053; Van ’t Wout et al., 20071; Pomarol-Clotet et al., 20103; Sachse
et al., 20141
Identity Matching
[Match target to 1 of 2-3 choices] NOT impaired Williams et al., 1999; Quintana et al., 2003; Kosmidis et al., 2007; Quintana 2011
Identity Matching
[Match target to 1 of 7 choices]
Impaired Williams et al., 1999
NOT impaired Hooker & Park, 2002
Morphed-Faces Identity Matching
[Match target to 1 of 2 choices]
Impaired Chen et al., 2015; Chen & Ekstrom, 2016
Approaching significance Chen et al., 2009
NOT impaired Chen et al., 2012; Norton et al., 2009
Identity Discrimination [same/different
judgements of serially presented faces]
Impaired Martin et al., 2005; Butler et al., 2008; Shin et al., 2008; Kim et al., 2010; Soria Bauser et al., 2012
NOT impaired Edwards et al., 2001; Johnston et al., 2010; Soria Bauser et al., 2012; Fakra et al., 2015
Identity Recognition Task
[Is this Person A or B?]
Slower only (accuracy at
ceiling) Baudouin et al., 2002
Tasks Assessing One Aspect of Identity
Sex recognition task [Male or female?] Impaired Hall et al., 2008
NOT impaired Bigelow et al., 2006; Michalopoulou et al., 2008; Delerue 2010
Morphed sex recognition task Impaired Bigelow et al., 2006
NOT impaired Bediou et al., 2005; Bediou et al., 2007; Bediou et al., 2012
33
Age Discrimination task [indicate the age by
decade: 1=teens... 7= seventies]
Impaired Schneider et al., 1995; Kohler et al., 2000; Habel et al., 2000
NOT impaired
(speed/accuracy trade-off) Schneider et al., 1998
Age Discrimination task
[Judge whether face is older or younger than 30]
Impaired Schneider et al., 2006
NOT impaired Gur et al., 2002a, Gur et al., 2002b; Pinkham et al., 2008; Goghari et al., 2011
Age Discrimination [Child, teen, adult or elder?] Impaired Delerue et al., 2010
Tasks Assessing Familiarity/Memory for Faces
Penn Face Memory Test Impaired Sachs et al., 2004; Calkins et al., 2005; Silver et al., 2009
The Warrington Recognition Memory Test–Faces
subtest
Impaired Soria Bauser et al., 2012; Whittaker, et al., 2001
NOT impaired Evangeli & Broks, 2000
Delayed identity matching (3s and 10s) Impaired Chen et al., 2009
Encoding and recognition task Impaired Walther et al., 2009
Famous Faces
Impaired Pomarol-Clotet et al., 2010
NOT impaired Evangeli & Broks, 2000; Whittaker et al., 2001; Joshua & Rossell, 2009; Lee et al., 2007; Walther et al.,
2009; Heinisch et al., 2013; Yun et al., 2014
Familiarity task [is this face familiar or
unknown?]
Impaired Irani et al., 2006; Caharel et al., 2007; Kircher et al., 2007
NOT impaired Delerue et al., 2010
1Short form, 2Long form, 3Not specified.
34
Tasks assessing a single aspect of face identity
Tasks that do not assess true identity discrimination, but just one aspect of facial identity have
produced slightly more consistent results. For instance, three studies reported that patients with
schizophrenia have an intact ability to identify the sex of a face (Bigelow et al., 2006; Delerue et al.,
2010; Michalopoulou et al., 2008), although one study reported reduced performance compared to
controls (Hall et al., 2008). Bediou and colleagues (2007; 2005, 2012) found that patients were able to
accurately identify the sex of a face even when stimuli were digitally morphed to increase ambiguity
(see Figure 3C). However, Bigelow and colleagues (2006) found that patients were significantly
poorer at identifying the sex of digitally-morphed faces but not non-morphed faces.
Three studies (Habel et al., 2000; Kohler, Bilker, Hagendoorn, Gur, & Gur, 2000; Schneider,
Gur, Gur, & Shtasel, 1995) found that patients were significantly impaired in judging the age of a face
in decades (i.e.: teens, twenties, thirties, etc), while one reported poorer accuracy, but faster reaction
times than controls (Schneider et al., 1998). Another study using a similar paradigm (judge each face
as child, teen, adult or elder) also found impaired performance in patients (Delerue et al., 2010). Yet
another study found that patients were impaired in judging if a face is older or younger than 30
(Schneider et al., 2006), while four studies found no difference using the same paradigm (Goghari,
Macdonald, & Sponheim, 2011; R. C. Gur et al., 2002; R. E. Gur et al., 2002; Pinkham, Hopfinger,
Pelphrey, Piven, & Penn, 2008). Taken together, these findings suggest that patients with
schizophrenia are less likely to show impairments when asked to make gross judgements about the
sex or age of a face, but have greater difficulty on tasks that require more fine-grained judgements,
such as defining the age of a face by decade. This result highlights the importance of selecting the
right measure when assessing face recognition in schizophrenia. True identity perception likely
reflects a judgement based on complex interactions between multiple facial features, so a task
assessing just one aspect of face perception (such as sex) may not be a valid indicator of a true deficit
in identity discrimination.
Identity recollection and familiarity
Finally, tasks that assess memory of – as opposed to discrimination between – faces have
largely revealed significant impairment in schizophrenia (Calkins, Gur, Ragland, & Gur, 2005; Chen
et al., 2009; Sachs, Steger-Wuchse, Kryspin-Exner, Gur, & Katschnig, 2004; Silver, Bilker, &
Goodman, 2009; Soria Bauser et al., 2012; Walther et al., 2009; Whittaker et al., 2001). In contrast,
patients have been shown to have an intact ability to recognise famous faces (Evangeli & Broks,
2000; Whittaker et al., 2001; Lee, Kwon, Shin, Lee, & Park, 2007; Joshua & Rossell, 2009; Walther
et al., 2009; Heinisch, Wiens, Grundl, Juckel, & Brune, 2013; Yun et al., 2014; although one study
reported significant impairment: Pomarol-Clotet et al., 2010). Three studies suggest that schizophrenia
35
patients tend to be less accurate in familiarity judgment of photographs of strangers and known people
(e.g., their doctor’s face) (Caharel et al., 2007; Irani et al., 2006; Kircher, Seiferth, Plewnia, Baar, &
Schwabe, 2007), while another reported intact performance (Delerue et al., 2010). However, these
tasks are not necessarily an indicator of pure face processing in schizophrenia because this disorder is
associated with general impairments in memory and new learning (Boyer, Phillips, Rousseau, &
Ilivitsky, 2007). Again, care should be taken when employing memory-based face recognition tasks to
ensure that general impairments in memory are taken into account.
The impact of expression variability on identity processing
Although models of face perception argue that facial identity processing is largely
independent from facial emotion processing, research clearly indicates that these two types of
information interact. Depending on conditions, emotional information may either interfere with or
enhance the perception of identity. For instance, Redford and Benton (2017b) examined the impact of
expression variability on identity discrimination in healthy controls using a card-sorting paradigm.
They found that participants made significantly more errors in grouping cards by identity when
emotionally-expressive faces were used, compared to cards which showed neutral faces. This suggests
that the additional variability introduced by emotional expressions interfered with the ability to
distinguish identity. In contrast, two studies (Redfern & Benton, 2017a, 2018) showed that increased
exposure to expression variability may enhance the learning of individual identities. In these studies,
participants who were exposed to neutral or low-expressiveness faces during the learning period were
less accurate at identifying identities when they were portrayed in high-expressiveness images.
However, participants who were exposed to high-expressiveness faces during learning were equally
accurate at identifying identities under low-expressiveness and high-expressiveness conditions. This
finding is in line with established literature which posits that greater exposure to variability
(expressive or otherwise) enhances our ability to learn new face identities (Andrews, Jenkins,
Cursiter, & Burton, 2015; Murphy, Ipser, Gaigg, & Cook, 2015; Ritchie & Burton, 2017). To date, no
published studies have specifically examined the impact of expression variability on identity learning
in schizophrenia. However, several have demonstrated that patients with schizophrenia are more
likely than healthy controls to show interference from identity information on an emotion recognition
task (Baudouin et al., 2002; Martin et al., 2005) and vice-versa (Zvyagintsev, Parisi, Chechko,
Nikolaev, & Mathiak, 2013). It is possible that impairments in the recognition of emotional
information may disproportionately impede face learning in day-to-day-life for people with
schizophrenia, despite intact processing of identity-specific information. Further research is needed to
clarify the impact of expression variability in populations with schizophrenia and other psychiatric
disorders.
36
Other aspects of impaired face processing in schizophrenia
Face information is extracted from the environment using both face-specific and more general
perceptual processes (McKone & Robbins, 2011). Some argue that face processing deficits in
schizophrenia indeed represent dysfunction in face-specific perceptual processes - generally referred
to as ‘holistic face processing.’ This represents a rapid, involuntary face-specific perceptual process
that integrates information across the face as a whole. It includes such information as the shapes of
individual features, the relative distances between them, and the contour of the cheeks and jaw
(Maurer, Grand, & Mondloch, 2002; McKone & Yovel, 2009). This process is specific to invariant
face information, and is therefore critical for perceiving identity (McKone & Robbins, 2011). For
instance, it has been demonstrated in healthy controls that holistic processing predicts an individual's
ability to remember and distinguish between faces (Richler, Cheung, & Gauthier, 2011). Similarly, it
has been shown that individuals with congenital prosopagnosia (inability to recognise faces) perform
poorly on holistic processing tasks (Palermo et al., 2011).
One common means of evaluating holistic processing is with the Face Inversion Effect -
observed as a reduction in face discrimination performance for inverted faces compared to upright
faces (see Figure 3D). The magnitude of this effect is thought to represent a loss of holistic
information crucial to face discrimination, and is disproportionately larger for faces compared to other
non-face stimuli (Yin, 1969). Studies of holistic processing in schizophrenia produced varied results,
with some studies reporting normal inversion effects for faces (Butler et al., 2008; Chambon,
Baudouin, & Franck, 2006; Schwartz, Marvel, Drapalski, Rosse, & Deutsch, 2002) while others
report reduced inversion effects compared to controls (Kim et al., 2010; Shin et al., 2008; Soria
Bauser et al., 2012). In particular, Shin and colleagues (2008) reported that patients with
schizophrenia were more impaired when discriminating faces that differed in configural information,
rather than featural information. An electrophysiological indicator of the face inversion effect is the
N170 (Eimer, 2000), a negative potential seen using electroencephalography (EEG). The N170 is
reduced in patients with schizophrenia while viewing inverted faces (Ibanez et al., 2012; Onitsuka et
al., 2006), and is associated with lower scores on measures of social functioning (Obayashi et al.,
2009; Tsunoda et al., 2012). These findings suggest there may be an underlying face processing
abnormality that may go undetected by commonly used behavioural measures.
In a related behavioural study, Schwartz and colleagues (2002) employed the composite face
task, which is considered to provide a more rigorous measure of holistic processing than other
inversion tasks (McKone, 2009). In this task, participants are required to make decisions about the
upper halves of faces while ignoring the lower halves. These face halves are either aligned to form a
complete face (producing an interference effect) or misaligned (removing the interference). When the
stimuli are inverted, however, the aligned faces no longer produce strong interference effects. It was
found that patients with schizophrenia showed typical patterns of interference for upright faces and
37
not inverted faces. While this study has not been repeated, it provides support for the argument that
holistic processing is largely preserved in schizophrenia and appears to contradict some of the results
using the face inversion effect.
Evidence for identity processing deficits using non-face stimuli
As outlined above, a number of studies have shown impaired performance on tasks designed
to assess face-specific processing. However, a number of similar deficits seem to be apparent on tasks
using non-face stimuli. For example, Soria Bauser and colleagues (2012) reported reduced inversion
effects for cars and bodies (see Figure 3D) that mirrored their findings using face stimuli, suggesting
an impairment that encompasses more than just face-specific holistic processing. An interesting
comparison is also provided by research looking at gait perception. Previous research has indicated
that the identity of an individual can also be extracted from an individual’s gait pattern (Cutting &
Kozlowski, 1977). Through the use of point-light displays (Johansson, 1973), visually impoverished
stimuli provide body form and structure solely through motion cues of coordinated dots (see Figure
3E). Similar to the ERP findings regarding face processing, the N170 component has also been found
in healthy individuals during visual processing of inverted point-light displays and static images of
bodies (Jokisch, Daum, Suchan, & Troje, 2005; Stekelenburg & de Gelder, 2004). Loula and
colleagues (2005) demonstrated that healthy subjects exhibited superior performance in identifying
self and friend’s movement when compared to a stranger’s movement. Furthermore, inverting the
point-light display resulted in chance performance across all three conditions.
There is some evidence to suggest that the ability to recognise identity cues from body
movements in impaired in schizophrenia. For instance, patients with schizophrenia are impaired in
discriminating point-light displayed body movements (biological motion) from scrambled point-light
display body movements (Kim, Doop, Blake, & Park, 2005). A study by Peterman and colleagues
(2017) used dynamic walking avatars (i.e. simplified 3D animated models) to investigate emotion and
gender recognition in patients with schizophrenia. It was found that patients were less able to extract
cues relating to gender and emotion from an avatars gait compared to healthy controls. They also
reported that patients were less likely to take the avatars emotion into account when making social
judgements, such as rating the angry avatar as more approachable compared to controls.
Does impaired identity processing reflect a generalised attentional deficit?
One possible account for face processing deficits in schizophrenia is that they are the result of a more
general impairment in allocating visuospatial attention (Baudouin et al., 2002). One suggestion is an
impairment in global versus local visual processing. ‘Global processing’ refers to the ability to attend
to any visual stimulus as a 'whole', as opposed to its component features (Tan, Jones, & Watson,
2009). Studies of schizophrenia have revealed impairments in global processing, but largely preserved
local processing both for static (Goodarzi, Wykes, & Hemsley, 2000; Johnson, Lowery, Kohler, &
38
Turetsky, 2005; Poirel et al., 2010; Silverstein, Kovacs, Corry, & Valone, 2000) and dynamic stimuli
(Chen, Nakayama, Levy, Matthysse, & Holzman, 2003). In addition, patients with schizophrenia
demonstrate a bias towards attending to the local level of a stimulus, even when task demands favour
a global strategy (Landgraf et al., 2011).
It is possible that a global processing deficit could contribute to impairments in identity
recognition because the important global-level information is not being processed efficiently. For
instance, it has been shown that identity recognition performance is improved when healthy
participants are primed to adopt a global processing strategy, and impaired when primed with a local
processing strategy (Macrae & Lewis, 2002; Perfect, 2003). Patients with schizophrenia similarly
showed less of a reduction in identity recognition performance compared to controls when configural
cues were removed from a face (Joshua & Rossell, 2009), indicating that these individuals relied more
strongly on local features when identifying famous faces. Global processing deficits could also
explain the expected deficits in identity recognition from gait in individuals with schizophrenia. Kim
and colleagues (2005) argued that deficits in biological motion perception in individuals with
schizophrenia may arise due to their well-documented difficulties in global motion perception (for
review see Chen, 2011) . This research question will be investigated further in Chapter 9.
Does impaired face identity processing reflect a general visual perceptual
difficulty?
Individuals with schizophrenia show a gamut of visual perceptual impairments (Butler et al., 2008).
These difficulties include form processing such as object recognition, grouping, perceptual closure,
and visual context (Brenner, Wilt, Lysaker, Koyfman, & O'Donnell, 2003; Doniger, Foxe, Murray,
Higgins, & Javitt, 2002; Kerr & Neale, 1993; Kohler et al., 2000; Kurylo, Pasternak, Silipo, Javitt, &
Butler, 2007; Place & Gilmore, 1980; Rabinowicz, Opler, Owen, & Knight, 1996; Rief, 1991;
Silverstein et al., 2000; Uhlhaas, Phillips, Mitchell, & Silverstein, 2006; Yang et al., 2013). Moreover,
neuroanatomical data indicate that the visual cortex in schizophrenia is abnormal with respect to the
density of neurons (Selemon, Rajkowska, & Goldman-Rakic, 1995), total number of neurons (Dorph-
Petersen, Pierri, Wu, Sampson, & Lewis, 2007) and GABA concentration in the visual cortex that is
associated with orientation-specific center-surround suppression (Yoon et al., 2010). Interestingly, the
face fusiform area (FFA) seems relatively intact, at least functionally (Yoon, D'Esposito, & Carter,
2006). Given the exhaustive list of basic visual perceptual deficits in schizophrenia, it seems likely
that processing of complex visual stimuli such as faces would also be compromised. Thus, it is likely
that at least some aspects of face processing deficits observed in schizophrenia arise from visual
cortical abnormalities.
Conclusions
39
Deficits in face processing have frequently been observed in patients with schizophrenia. In
order to fully understand the mechanisms underlying these impairments it is important to consider the
relative contribution of the multiple factors that may be involved. The fact that deficits have been seen
in face identity tasks without an emotional/ expression recognition component suggests that these
deficits are unlikely to be limited to emotion processing. Moreover, the observation of more
generalised impairments in visual and attentional function in these patients also raises questions about
whether there is indeed anything special about faces at all. Lastly, the potential role of medication in
these impairments has yet to be clearly determined. It is only through future controlled studies that
balance difficulty across memory, attentional and perceptual demands – or directly assess the
capacities – that we will begin to understand how face processing deficits emerge in these patients.
40
Chapter 4: What do we really want to know about face processing
in schizophrenia? Rationale for the Inpatient Study
Impairments in face processing have been demonstrated in patients with schizophrenia, and
are shown to correlate significantly with aspects of social functioning and overall quality of life
(Addington et al., 2006; Hooker & Park, 2002; Irani et al., 2012). Two components of accurate face
processing are the perception of emotion (perceiving transient emotional expressions) and identity
(perceiving the invariant features which distinguish an individual; Calder, 2011). Deficits in both of
these components have been demonstrated to a varying degree in schizophrenia (Chen et al., 2012;
Martin et al., 2005). However, it remains unclear whether these functional deficits reflect a single
face-specific deficit, distinct impairments in emotion and identity processing, or some other
generalised impairment. In healthy individuals, emotion and identity processing are thought to be
relatively independent processes which are mediated via different neural routes (Haxby & Gobbini,
2011; Haxby et al., 2000). It is therefore important to determine whether different subtypes of
impairments can be identified in schizophrenia, and whether distinct processes can be targeted for
intervention to improve functional outcomes (Marwick & Hall, 2008). The current experiment aims to
investigate these questions using dynamic (video-based) stimuli, which are shown to have greater
ecological validity and sensitivity to deficit compared to traditional static (image-based) emotional
stimuli.
Emotion processing deficits
There is extensive research demonstrating that patients with schizophrenia perform poorly on
tasks of facial emotion processing compared to healthy controls (see Kohler et al., 2010 for meta-
analytic review). This deficit has been demonstrated across a variety of tasks involving either
recognising or labelling the emotion portrayed (Bediou et al., 2007; Fakra et al., 2008; Johnston,
Devir, & Karayanidis, 2006; Kucharska-Pietura et al., 2005; Penn et al., 2000; Simpson, Pinkham,
Kelsven, & Sasson, 2013; Strauss, Jetha, Ross, Duke, & Allen, 2010; Whittaker et al., 2001) or
discriminating between pairs of static faces showing the same or different expression (Addington et
al., 2006; Fakra et al., 2015; Martin et al., 2005; Weniger et al., 2004). Together, these results suggest
that this deficit is robust and persists even when memory and language demands of the task are
minimised. Moreover, longitudinal studies suggest that these deficits are stable up to at least one year,
and do not improve with the resolution of symptoms (Addington & Addington, 1998; Kee, Green,
Mintz, & Brekke, 2003). Despite these impairments being demonstrated repeatedly across a range of
studies, it remains unclear whether these deficits in emotion processing exist above and beyond more
general perceptual or attentional impairments.
41
Identity processing deficits
Studies investigating identity processing in schizophrenia, however, have not been as
consistent (Darke, Peterman, Park, Sundram, & Carter, 2013). Significant deficits have been
demonstrated in schizophrenia across a variety of identity discrimination tasks (Butler et al., 2008;
Chen et al., 2009; Shin et al., 2008; Soria Bauser et al., 2012) but not others (Chen et al., 2012;
Johnston et al., 2010; Norton et al., 2009; Pomarol-Clotet et al., 2010). An important distinction can
be drawn between face discrimination tasks (considering overall identity of a face) and single-feature
labelling tasks (considering one aspect of identity, such as age or sex). Namely, studies assessing just
one feature of identity are less likely to show impairments in schizophrenia, particularly when the task
only requires broad judgments (e.g.: male or female?) rather than fine-grained judgements (e.g.:
estimate age by decade) (Bediou et al., 2007; Bediou et al., 2012; R. E. Gur et al., 2002).
Consequently, it has been suggested that the deficits shown in previous studies (Butler et al., 2008;
Chen et al., 2009; Shin et al., 2008; Soria Bauser et al., 2012) may reflect general task demands, rather
than actual impairment in face-specific processes (Marwick & Hall, 2008). For instance, face
measures with a considerable memory component typically show deficits (Caharel et al., 2007; Silver
et al., 2009; Soria Bauser et al., 2012) and may result from broader working memory impairments
associated with schizophrenia (Piskulic, Olver, Norman, & Maruff, 2007).
Are these face processing deficits specific to schizophrenia?
A further question is whether any face processing deficits observed are shared by other
psychiatric disorders. The diagnostic specificity of these deficits is unclear as the majority of face
processing studies compare only patients with schizophrenia to healthy controls. For the smaller
number of studies that have included other patient groups, individuals with schizophrenia have been
found to show significantly greater face processing impairments relative to individuals with
depression (Bediou, Krolak-Salmon, et al., 2005; Weniger et al., 2004), bipolar disorder (Addington
& Addington, 1998; Lembke & Ketter, 2002; Venn et al., 2004) and affective psychoses (Edwards et
al., 2001). However, other studies have reported equivalent impairments in emotion processing in
patients with bipolar disorder (Bozikas, Tonia, Fokas, Karavatos, & Kosmidis, 2006; Getz, Shear, &
Strakowski, 2003; Lembke & Ketter, 2002). Face processing impairments in disorders other than
schizophrenia, and the relevant results of the inpatient experiment, will be reviewed further in Chapter
10.
It is also possible that these deficits accompany specific symptoms which are shared across
disorders. For instance, emotion processing deficits have been reliably associated with both positive
and negative symptoms in schizophrenia, as well as other illness factors such as later age of illness
onset and inpatient status (Kohler et al., 2010; Martin et al., 2005; Sachs et al., 2004; van 't Wout et
al., 2007). However, the relationship between identity recognition impairments and symptoms has not
42
been widely studied, and evidence is mixed. Several studies have reported a negative correlation
between identity recognition performance in schizophrenia and negative symptoms (Addington &
Addington, 1998; Baudouin et al., 2002; Chen et al., 2012; Martin et al., 2005), while others report an
association between performance and overall symptom severity (Chen et al., 2009; Penn et al., 2000;
Sachs et al., 2004). This literature will be discussed further in Chapter 11.
The current study aimed to address these outstanding issues by assessing participants with a
range of psychiatric disorders (n=86) and healthy controls (n=20) using four different emotion and
identity processing tasks. Furthermore, to ascertain whether these deficits generalise beyond face
processing, performance was also compared on an equivalent task using non-face stimuli. Dynamic
(video-based) stimuli were used in all tasks, as these are shown to have better ecological validity than
traditional static face images (Atkinson, Dittrich, Gemmell, & Young, 2004; Kilts, Egan, Gideon, Ely,
& Hoffman, 2003). Chapter 5 will review the importance of dynamic versus static emotional stimuli
in greater detail, and will discuss the development and initial testing of the novel stimuli created for
this experiment. Additionally, Chapter 6 will investigate associations between psychosis-like
personality traits in healthy controls (i.e. schizotypy) and emotion-processing in this initial study.
For clarity, the results of the inpatient study have been divided into different chapters aligned
with the following research questions:
I. Characterising the deficit in schizophrenia: Can impairments in emotion-processing be
explained by more general deficits in non-emotional face processing or non-face processing?
The first aim focuses on the comparison of patients with schizophrenia and age-matched healthy
controls on a series of dynamic tasks designed to assess emotion-processing, identity-processing,
and non-face visual processing. These results will be discussed in Chapter 8.
II. Are face processing impairments specific to schizophrenia, or are they shared by other
psychiatric disorders?
The second aim is to compare face processing in patients with schizophrenia with other
psychiatric inpatient groups including bipolar disorder, other psychotic disorders (such as drug-
induced psychosis), and non-psychotic disorders (such as major depressive disorder). These
results will be discussed in Chapter 10.
III. Do specific symptoms correlate with face-processing impairments?
The third aim is to examine the relationship between face processing impairments and specific
symptomatology across psychiatric disorders. These results will be discussed in Chapter 11.
43
Justification of tasks used
To assess emotion processing ability, two different paradigms were employed: an emotion
discrimination task and an emotion labelling task. As mentioned in Chapter 2, these two tasks have
slightly different demands, and may be sensitive to different aspects of cognitive impairment. That is,
discrimination tasks are more dependent on working memory ability, while labelling tasks may be
affected by semantic memory deficits (Macmillan & Creelman, 1991). Furthermore, as these two
types of tasks may be correlated with different impairments in schizophrenia (Irani et al., 2012), both
paradigms were included to see if any kind of differential deficit is seen in patients.
To assess facial identity recognition, two tasks were created that were matched to the task
demands of the two emotion tasks: identity discrimination and sex labelling. Identity discrimination
involves making same-or-different judgements of serially presented faces. Patients with schizophrenia
have shown impairments in some studies (Butler et al., 2008; Martin et al., 2005; Shin et al., 2008;
Soria Bauser et al., 2012) but not others (Edwards et al., 2001; Johnston et al., 2010; Soria Bauser et
al., 2012). However, these studies used static face stimuli only. The current study is the first to use
dynamic face stimuli to assess facial identity processing in schizophrenia. The second task, sex
labelling, involves simply deciding whether a face looks more male or more female. Patients with
schizophrenia typically show no impairments on tasks such as this, which require judgements about
only a single aspect of facial identity (Bediou et al., 2007; Bediou, Franck, et al., 2005; Bediou,
Krolak-Salmon, et al., 2005). For this reason, some authors conclude that identity recognition is
intact in schizophrenia. However, it is unclear if this task is sensitive to true impairments in identity
recognition, which typically involves consideration of multiple aspects of facial identity. Both
paradigms have therefore been included in this experiment.
Finally, a non-face control task was included to determine whether any deficits shown on the
four face tasks may be better accounted for by a more general visual or cognitive impairment. In other
words, if impairments in emotion-processing and identity-processing are specific to faces, we would
expect to see comparatively intact performance on a non-face discrimination task. However, if
impairments are not specific to faces, we would expect to see similar deficits on the non-face task.
44
Chapter 5: Is our data only as good as our tools? Issues of Static
versus Dynamic Faces, and Development of a Novel Stimuli Set
Face-based tasks are used ubiquitously in the study of human perception and cognition.
Video-based (dynamic) face stimuli are increasingly utilised by researchers because they have higher
ecological validity than static images (Fiorentini & Viviani, 2011). However, there are few ready-to-
use dynamic stimulus sets currently available to researchers. This chapter will review the literature
surrounding the use of dynamic versus static stimuli in face research. It will then outline the
development of three original dynamic stimulus sets which were created as part of this PhD. Finally,
this chapter will present an experiment evaluating the utility of these new stimuli sets.
Criticisms of traditional static face stimuli sets
Face processing has been the focus of an enormous variety of research spanning a range of
disciplines, examining both human and non-human subjects. Not only is face processing a rapid,
involuntary, and highly specialised perceptual process, but it also plays a crucial role in
communication and social interaction. Consequently, face-based stimuli sets are used for a huge range
of applications in the study of human perception and cognition, such as investigating low-level visual
processing (Cauchoix, Barragan-Jason, Serre, & Barbeau, 2014), memory (Faces subtest; WMS-III
(Wechsler, 1997), Theory of Mind (van Veluw & Chance, 2014), different aspects of facial identity
perception (Fitousi & Wenger, 2013), and the perception of emotional expression (Fusar-Poli et al.,
2009).
Traditionally, face processing studies are typically carried out using behavioural or
neuroimaging measures that involve viewing and making judgements about static images of faces. For
example, one commonly used measure of identity recognition is the Benton Facial Recognition Task
(Benton, 1983), which requires participants to match the identity of a target face to one of several
possible options. The most widely-used stimulus set for examining emotion processing is the Ekman
faces (Ekman & Friesen, 1976), a set of 60 photographs demonstrating the six main emotions:
happiness, sadness, disgust, fear, anger, and surprise. Other standardised static face sets include the
Japanese and Caucasian Facial Expressions of Emotion (JACFEE; Biehl et al., 1997), the Montreal
Set of Facial Displays of Emotion (MSDEF; Beaupré & Hess, 2005) and the Nim Stim Face Stimulus
Set (Tottenham et al., 2009).
However, the use of stereotyped or exemplar faces such as these have been criticised for
lacking ecological validity. That is, they show only exaggerated, staged expressions which do not
reflect the subtler nuances that we experience in natural social interactions (Davis & Gibson, 2000).
Furthermore, it has been argued that tasks using exemplar faces are prone to ceiling effects and may
not be sensitive to subtler impairments in face-specific processes (Harms, Martin, & Wallace, 2010).
45
To address this, some researchers have utilised morphing software to create stimuli that show varying
intensities of emotion. Figure 5.1 shows an example of a set of morphed faces ranging from high to
low intensity, created by Norton and colleagues (2009). Stimuli such as these permit the study of
threshold differences between clinical populations, which reveal that patients with schizophrenia
require greater intensity to identify expressions such as disgust and fear (Bediou, Franck, et al., 2005;
Chen et al., 2012; Norton et al., 2009).
Another way in which researchers have attempted to improve the ecological validity of face
tasks is through the use of ambient stimuli. Ambient stimuli refer to naturalistic photographs or
videos, typically obtained from films or the internet. It has been argued that reliance on tightly-
controlled, artificially edited stimuli not only limits the generalisability of results to everyday
situations, but may even undermine our theoretical understanding of face recognition (Burton, 2013).
Burton argues that, as face processing involves deriving stable constructs (such as an emotion or an
identity) from constant variability, then by eliminating this variability we may no longer be assessing
the same perceptual process. For instance, research in healthy populations suggests that the inclusion
of body information (i.e.: not just an isolated face) has an effect on both emotion and identity
recognition, even if the participant is unaware of using these cues (O'Toole et al., 2011; Rice, Phillips,
Natu, An, & O'Toole, 2013; Van den Stock, Tamietto, Hervais-Adelman, Pegna, & de Gelder, 2015).
Other studies suggest that more naturalistic images may produce different patterns of deficits in
disorders such as schizophrenia. For example, Davis and Gibson (2000) found that patients with
schizophrenia were significantly impaired when discriminating posed facial expressions, but not
genuine facial expressions (e.g.: photos of surprise obtained by popping a balloon near the actor).
More recently, Regenbogen and colleagues (2015) found that patients with schizophrenia and
depression showed no deficits on an empathy task which involved viewing naturalistic stimuli. This
finding was in contrast to the majority of past literature using traditional stimuli (Li et al., 2010;
Stuhrmann, Suslow, & Dannlowski, 2011), suggesting that the use of non-naturalistic stimuli may be
artificially eliciting deficits that are unrelated to everyday perceptual functioning. Taken together, this
literature suggests that, although heavily edited and standardised stimuli are attractive because they
can be used to minimise unwanted variability, researchers should take extra care to consider the
greater generalisability of their tasks to everyday functioning.
46
In addition to varying intensity, researchers are increasingly utilising dynamic (video-based)
face sets to examine face processing impairments. Again, the ecological validity of static stimuli has
been questioned because they are not representative of the moving, changing facial expressions we
encounter in face-to-face interactions (Atkinson et al., 2004; Kilts et al., 2003). Discerning the
emotional state of an individual in daily life involves detecting and rapidly interpreting temporal
changes in facial movements such as a brief smile, or a narrowing of the eyes. It can be argued,
therefore, that expressions are inherently dynamic, and that static images may be too impoverished to
adequately tap into emotion processing mechanisms (Fiorentini & Viviani, 2011). Several dynamic
face sets have been made available for emotion research, including the Perception of Emotion Test
(POET; Kilts et al., 2003), the Cohn-Kanade Facial Expression Database (Kanade, Cohn, & Tian,
2000), the CMU-Pittsburgh AU-Coded Face Expression Image Database (Pantic, Valstar, Rademaker,
& Maat, 2005), and the Amsterdam Dynamic Facial Expression Set (ADFES; van der Schalk, Hawk,
Fischer, & Doosje, 2011). Research in both healthy and clinical populations suggests that there are a
range of differences in the ways that individuals respond to static and dynamic face stimuli.
Figure 5.1. Example of morphed stimuli created by Norton and colleagues (2009, p.1095) used in an
emotion discrimination task. Participants are asked, “Which face, the first or the second, looks more
afraid?”
Comparison between static and dynamic faces in healthy popu lations
A range of studies in healthy controls have investigated the role of dynamic motion in
emotion processing. These findings are summarised in Table 5.1. In particular, several behavioural
studies have reported advantages for recognising emotion from dynamic faces over traditional static
47
faces. For instance, two studies found increased accuracy rates for dynamic faces across all emotion
compared to matched static faces (Wehrle, Kaiser, Schmidt, & Scherer, 2000; Weyers, Muhlberger,
Hefele, & Pauli, 2006). Similarly, a study using dynamic stimuli that varied in the intensity of
expression found that dynamic stimuli were recognised more easily than static (Montagne, Kessels,
De Haan, & Perrett, 2007). Other studies, however, found no difference between static and dynamic
conditions (Fiorentini & Viviani, 2011) or only found significant effects for certain emotions. For
example, Recio, Schact and Sommer (2011) found that dynamic happy expressions were recognised
with greater accuracy compared to static faces, but reported no such effect for expressions of anger.
Kamachi and colleagues (2001) found that participants were better at recognising static faces of anger
and sadness compared to dynamic faces. When comparing the recognition of facial expressions
presented centrally versus peripherally, Fujimura and Suzuki (2010) reported that dynamic angry
faces were recognised more accurately than static faces in the periphery only, with no differences for
central presentation or other valences of emotion. In addition to differences in recognition between
dynamic and static stimuli, Yoshikawa and Sato (2006) found increased self-reported “emotional
experiences” in response to dynamic faces compared to matched static faces. Similarly, Biele and
Grabowska (2006) found that dynamic faces were perceived as more intense than static versions.
Using a cueing paradigm, Kaufman and Johnston (2014) found that dynamic cues had a greater
impact than static cues on a same-or-different emotion discrimination task.
Different responses to static compared to dynamic facial expressions have also been
demonstrated through imaging studies. In a PET study, Kilts and colleagues (2003) revealed
significantly different patterns of brain activation for dynamic happy and angry faces compared to
static, particularly involving area V5, the superior temporal sulcus, cerebellum, and temperomedial
cortical areas. LaBar and colleagues (2003) found increased fMRI activation in the amygdala and
fusiform gyrus for angry and fearful dynamic expressions compared to static equivalents, indicating
stronger emotional responses to moving stimuli. Similarly, Sato and colleagues (2004) found greater
activation in the fusiform gyrus, medial temporal gyrus, and inferior occipital gyrus in response to
dynamic happy and fearful expressions. More recent fMRI studies and a meta-analysis by Arsalidou,
Morris and Taylor (2011) have similarly reported substantial increases in activation to dynamic faces
in brain areas associated with the processing of emotion, biological motion, and social cues (Kessler
et al., 2011; Trautmann, Fehr, & Herrmann, 2009). Further evidence for a dissociation between static
and dynamic facial expressions has come from studies of clinical populations. At least two case
studies have been published of brain-injured patients who were unable to identify emotions in static
images, but could correctly identify emotions from dynamic faces, or from faces formed of moving
point-light displays (Adolphs, Tranel, & Damasio, 2003; Humphreys, Donnelly, & Riddoch, 1993).
Taken together, these studies suggest that dynamic faces more effectively tap into neural processes
relevant to emotion processing compared to static face images.
Not only do static and dynamic expressions elicit different patterns of behavioural and neural
48
responses, but they also appear to produce differences in viewers’ unconscious muscular reactions.
Electromyography (EMG) studies allow researchers to examine the movements of participants’ facial
muscles in response to viewing static and dynamic facial expressions. For instance, Sato and
colleagues (2008) found that dynamic happy faces prompted stronger activation of the zygomaticus
major muscle (involved in smiling), while dynamic angry faces prompted stronger activation of the
corrugator supercilii muscle (involved in frowning) compared to static faces. A related study using
discreet video recording found similar patterns of muscular movements in response to dynamic faces,
suggesting that moving stimuli are more likely to elicit facial mimicry than static images (Sato &
Yoshikawa, 2007). Two studies measuring EMG responses to happy, angry, and neutral faces asked
participants to rate the intensity of the emotion shown. In both studies, the dynamic stimuli were rated
as more intense, as well as more realistic than static equivalents (Rymarczyk, Biele, Grabowska, &
Majczynski, 2011; Weyers et al., 2006). However, while happy faces elicited stronger activation of
the zygomatic muscles and reduced activation of the corrugator supercilii, no significant EMG
differences were found for dynamic angry faces in either study (Rymarczyk et al., 2011; Weyers et
al., 2006). Overall, EMG studies suggest that dynamic facial emotions are more likely to prompt
spontaneous facial mimicry than static faces, however this finding appears to be more robust for
happy faces than for other emotions.
The studies above provide convincing evidence of an advantage for dynamic stimuli over
static in the investigation of emotion. However, is this advantage due to the presence of motion, or
due to some other characteristic of the stimulus? An obvious confound when comparing dynamic
stimuli to static is that they contain different quantities of visual information. That is, while dynamic
stimuli are comprised of multiple frames, static comprise only one. Therefore, it is possible that
dynamic stimuli simply provide a larger number of visual cues compared to static images, and that
this drives the recognition advantages seen in previous studies. To investigate whether differences
between static and dynamic stimuli still exist when the quantity of information is controlled for,
Ambadar, Schooler and Cohn (2005) conducted an emotion-recognition study using subtle (low
intensity) facial expressions. Performance was compared across four stimulus conditions: dynamic (3
to 6 frame videos, moving from a neutral to emotional expression), multi-static (the same 3 or 6
frames, with masking in between to eliminate the perception of motion), first-last (sequence showing
only the first and final frames of each video), and single-static (showing the final frame only). They
found that both accuracy and confidence ratings were significantly higher for the two moving
conditions (dynamic and first-last) than the two static conditions (single-static and multi-static). This
suggests that the advantage shown for dynamic stimuli is due to the presence of motion, rather than
the quantity of information. As the performance for the first-last sequence was equal to the dynamic
sequence, the authors argue that emotion recognition is likely tuned to the perception of change from
a neutral face to an expressive face, and is not dependent on cues relating to the actual temporal
unfolding of an expression (Ambadar et al., 2005).
49
Table 5.1. Behavioural studies comparing emotional processing of dynamic and static face stimuli in
healthy controls.
Study Task type Dynamic vs static
Kamachi et al. (2001) Rate intensity (7 choice) Static > dynamic
Ambadar, Schooler & Cohn
(2005) Labelling (7 choice) Dynamic > static
Biele & Grabowska (2006)
Rate intensity (4 choice) Dynamic > static
Yoshikawa & Sato (2006) Matching to same intensity
(sliding scale)
No difference, but rapid changes
were perceived as more intense
than slow changes.
Montagne et al. (2007) Labelling (6 choice) Dynamic > static
Fujimura & Suzuki (2010) Labelling (6 choice) Dynamic > static
(anger only)
Fiorentini & Viviani (2011) Labelling (2 choice) No difference
Recio, Schacht & Sommer (2011) Labelling (3 choice) Dynamic > static
(happiness only)
Kaufman & Johnston (2014) Same-or-different discrimination
(static only)
Dynamic cues produced faster
‘same’ responses than static cues
Studies using dynamic faces to examine emotion processing in
schizophrenia
Research in healthy populations suggests that using dynamic stimuli – rather than static
images – to investigate emotion processing confers a range of advantages. These include increased
recognition for both exemplar and more subtle expressions, increased ratings of emotional intensity,
greater neural activation, and higher rates of spontaneous facial mimicry. Despite this, dynamic
stimuli are still not widely used, especially in studies of clinical populations. Only a handful of studies
have employed dynamic stimuli to investigate emotion processing in schizophrenia. These studies are
summarised in Table 5.2.
50
Table 5.2. Behavioural studies using dynamic stimuli to investigate emotion processing in
schizophrenia.
Study Task type Dynamic vs static Group difference?
Archer, Hay & Young
(1994)
Match label to face
(2 choice)
Static > dynamic
HC > SZ
Depression > SZ
Match face to face
(3 choice) No difference
HC > SZ
Depression > SZ
Johnston et al. (2010) Labelling (2 choice: fear
or suprise) No difference HC > SZ
Mendoza et al. (2011) Labelling (6 choice)
Judge intensity only N/A HC > SZ
Behere et al. (2011) Labelling (7 choice) Dynamic > static HC > SZ
Souto et al. (2013) Labelling (6 choice) N/A No difference
Hargreaves et al. (2016) Labelling (6 choice) N/A HC > SZ
Note: HC = Healthy control, SZ = Schizophrenia.
One of the earliest studies to use dynamic faces to investigate schizophrenia was conducted
by Archer, Hay and Young (1994) , who compared inpatients and healthy controls using colour videos
of talking faces. They found that inpatients with schizophrenia showed significantly poorer accuracy
for identifying anger, surprise, disgust, sadness and fear compared to healthy controls. They were also
significantly poorer at identifying anger and surprise compared to inpatients with depression.
Another study by Johnston and colleagues (2010) compared dynamic video of facial
expressions with static images in patients with schizophrenia and healthy controls. Participants
viewed one-second videos of neutral expressions morphing into either fearful or surprised
expressions. Within a single block, these videos were randomly interspersed with static images of
these same expressions. Two control tasks were included that used static stimuli: an identity
discrimination task and a non-face discrimination task. It was reported that patients performed poorer
than controls for the emotion discrimination task only (for both static and dynamic stimuli),
suggesting an emotion-specific deficit. Of greater importance, however, was the unique finding that
static and dynamic stimuli correlated with different patterns of symptomatology. High negative
symptom scores were associated with poorer performance for static stimuli but not dynamic stimuli,
while high positive symptom scores were associated more strongly with dynamic stimuli. Links
between performance and symptomatology will be discussed in greater detail in Chapter 11. However,
this finding suggests that tasks using static and dynamic face stimuli may tap into different aspects of
impairment in schizophrenia, and further highlights the need to use ecologically valid stimuli when
investigating face processing deficits in this disorder.
51
A similar study by Behere and colleagues (2011) used dynamic and static expressions to
evaluate threat identification in schizophrenia. Consistent with the results seen in healthy populations
(e.g.: Weyers, Muhlberger, Hefele, & Pauli, 2006; Montagne, Kessels, De Haan, & Perrett, 2007),
patients with schizophrenia were more accurate at identifying dynamic emotions compared to static.
Moreover, they reported that patients experiencing first-rank symptoms (a subset of symptoms
strongly associated with paranoid schizophrenia) were more likely to misidentify non-threatening
emotions as threatening (e.g.: label a neutral face as angry) compared to patients without first-rank
symptoms.
Using a slightly different paradigm, Mendoza and colleagues (2011) compared patients with
schizophrenia, first degree relatives, and unrelated controls on a dynamic morphed stimuli task. In this
experiment, participants were shown a gradual sequence of frames morphing from a neutral face to
one of six emotions, and were required to press a button as soon as they could identify the emotion. It
was found that patients with schizophrenia required significantly more frames (i.e.: greater intensity)
to correctly recognise all six emotions. Unaffected relatives, however, required greater intensity to
recognise disgust and fear only.
A more recent study (n.b.: published since the collection of data for this PhD) examined
emotion processing in schizophrenia using dynamic stimuli which varied in emotional intensity
(Hargreaves et al., 2016). In this paradigm, videos of the six primary emotions were presented in a 6-
option labelling task. The first block showed emotions at 20% intensity, then increased in 20%
intervals until the fifth and final block showed expressions at 100% intensity. It was found that
outpatients with schizophrenia (n=47) performed significantly less accurately than healthy controls
across all emotions except surprise. Accuracy increased with emotional intensity for all participants.
Furthermore, performance in patients correlated with IQ, several measures of memory, attentional
control and social cognition, while performance in healthy controls correlated with a face memory
task and social cognition only. These results concur with research using static stimuli (e.g.: Bediou et
al., 2007) and lend weight to the argument that emotion-processing deficits are associated with the
core cognitive deficits seen in schizophrenia.
In an attempt to develop a paradigm with even greater ecological validity than 2-dimensional
video, Souto and colleagues (2013) created a virtual reality program to evaluate emotion recognition
using 3D computer-generated avatars. Interestingly, preliminary findings revealed no significant
differences in accuracy rates between healthy controls and patients with schizophrenia. However, this
may be due to insufficient sample sizes (n=12 per group). Regardless, the utility of virtual reality
alternatives to 2-dimensional stimuli remains a promising avenue for future study.
In addition to behavioural studies, to date only two imaging studies have been conducted
using dynamic stimuli to investigate emotion processing in schizophrenia. Russell and colleagues
(2007) collected fMRI data while patients with schizophrenia viewed videos of emerging or
dissipating fear (i.e.: neutral-to-fear vs fear-to-neutral). Compared to healthy controls, patients with
52
paranoid symptoms showed abnormal activity in the left and right amygdala while viewing emerging
fearful faces, while non-paranoid patients showed increased activity in the hippocampi only. These
findings are in line with previous imaging research suggesting that amygdala abnormalities may be
involved in the misattribution of emotional salience in schizophrenia (Anticevic et al., 2012). A more
recent fMRI study by Mothersill and colleagues (2014) examined activations in response to passive
viewing of angry and neutral face videos. They found that patients with schizophrenia showed less
deactivation of the medial prefrontal cortex (including the anterior cingulate cortex), and less
activation of the left cerebellum when viewing faces compared to healthy controls. Again, these
results are consistent with research using static face stimuli which have implicated the ACC in
emotion processing difficulties in schizophrenia (Habel et al., 2010; Hall et al., 2008).
In summary, the use of dynamic stimuli to investigate emotion processing in schizophrenia
remains sparse, but is gaining in popularity. In line with studies using static stimuli, dynamic studies
have demonstrated significant differences in emotion identification between patients with
schizophrenia and healthy controls. Furthermore, one study suggests that impairments in recognising
static and dynamic faces are associated with different patterns of symptomatology in schizophrenia.
Overall, the literature suggests that not only are dynamic stimuli more ecologically valid than static,
but they appear to involve quite different neural substrates and therefore may be sensitive to different
types of impairment. Given these differences, there is a strong rationale for employing dynamic
stimuli in future schizophrenia research.
Rationale for the development of original dynamic stimuli
Given the above criticisms of traditional static stimuli, it was decided that dynamic stimuli
would be utilised for this thesis. Several online databases of dynamic face videos already exist and
can be accessed without charge for research purposes (e.g.: ADFES; van der Schalk, Hawk, Fischer,
& Doosje, 2011). For the purpose of our experiments, however, these stimuli were not suitable
without further modification. For instance, the majority of videos contained non-face identifying
features such as hair, glasses, and clothing which could be used by participants to distinguish between
different identities. Given that we wished to evaluate facial identity processing, it was necessary to
eliminate these features to prevent participants from relying on these extraneous cues.
Furthermore, one of our goals was to introduce a new technique: applying morphing software
to video-based stimuli. Just as Norton and colleagues (2009) utilised morphing to modify static
stimuli, we aimed to modify the intensity of emotional faces (or the similarity between two faces) to
create a broader range of stimuli. By varying this intensity, it is expected that the stimuli will be less
sensitive to ceiling effects and more likely to detect subtle impairments in face processing ability
(Hargreaves et al., 2016; Norton et al., 2009).
Three matched dynamic stimuli sets were created: a set of emotional faces varying by
53
emotional intensity (fear and disgust); a set of non-emotional face animations varying by facial
similarity (morphing of same-sex or different-sex face pairs); and a set of rotating car animations.
This third dynamic set was created to serve as a non-face control stimulus that is matched to the same
task parameters as the face sets. All stimuli can be downloaded from http://go.unimelb.edu.au/e3t6.
Step by step creation of a new stimuli set
Preparation of face stimuli for morphing
Raw videos were sourced with permission from the MMI-Facial Expression Database
(https://mmifacedb.eu; Pantic et al., 2005; Valstar & Pantic, 2010) and the Facial Expressions and
Emotion Database (FEED; Wallhoff, 2006) . Each .AVI video file was then converted to a sequence
of frames using VirtualDub (Lee, 2012). From this sequence, 13 frames were selected that showed a
progression from a neutral resting face to a ‘completed’ facial movement or expression. Each frame
was then edited in Adobe Photoshop CS2 to isolate the face against a black background, stabilise any
head movements, and to remove non-face cues such as glasses, hair, and facial hair. All stimuli were
converted to greyscale to eliminate the possibility of participants using colour-matching as an
alternative strategy to discriminate between faces. Each face fitted within 200x200 pixels.
Morphing to vary the intensity of an expression
As a means to assess facial affect processing, videos of facial expressions were edited to vary
the intensity of each expression without altering the identity of the individual. To achieve this, the
first frame of each video (a neutral expression) was morphed with every subsequent frame using
Fantamorph 5 (Abrosoft, 2012). This was accomplished using the ‘Face Locator’ add-in to map out
the main features of each face. Once the maps were manually adjusted to indicate features as precisely
as possible, the morphing slider was used to select the ratio between the neutral and emotive faces,
e.g.: 33% neutral, 67% emotive. This produced an overall effect of ‘relaxing’ or ‘diluting’ the facial
movements in order to create a subtler facial expression in the final video. The final morphed frame
was then exported as a new file, and then the process was repeated for all remaining frames. This
method was used with videos showing fear and disgust to create a series of animations ranging from
100% intensity to 33% intensity (see Figure 5.2 for examples). Original animations lasted 1 second.
After piloting, however, stimuli were slowed down to 2 seconds in order to increase accuracy to an
acceptable level.
For this set, twelve videos of different individuals (6 showing fear, 6 showing disgust) were
morphed to create five levels of expression intensity: 33%, 50%, 67%, 83%, and 100%. A sixth
intensity level, 17%, was also piloted. These were not included in the final set, however, because
healthy controls could not reliably identify the emotion at such a low intensity. The final set
comprised 60 stimuli.
54
Figure 5.2. Partial image sequences (every other frame) from five of the video stimuli ranging in
intensity of emotion. The top video shows the original video of an individual’s face changing from
neutral to an expression of disgust. This video was then morphed with the neutral face frame (leftmost
frame) to reduce the intensity of the final expression (rightmost frame). Emotional intensity ranged
from 100% (unedited video) to 33% intensity.
Morphing to vary the identity of a face
To assess non-emotional aspects of face processing, videos were edited to vary the similarity
between two different individuals. To accomplish this, pairs of videos were selected that showed the
same non-emotive facial movement (e.g.: raising the eyebrows, opening the mouth, or sticking out the
tongue). The thirteen individual frames were then matched as closely as possible so that both videos
showed the movement at the same speed. From there, the first frame of one video was morphed with
the first frame of the second video using the ‘Face Locator’ add-in in Fantamorph 5. When repeated
for all frames, this produced a new video showing the new ‘morphed’ individual performing the full
55
facial action. This method was used to create a series of 1 second animations ranging from one
individual to the other via 20% increments.
Two different sets of stimuli were created: one set where the faces in each morphed pair were
the same sex, and one set where each morphed pair were opposite sex. For the same-sex set, six pairs
of unique individuals (3 male, 3 female) were morphed in pairs to create six sets of face animations
ranging from one identity to the other by increments of 20%. Thirty-six stimuli were created in total.
The stimuli for the opposite-sex set were created in the same way as above. Six pairs of
individuals of the opposite sex were morphed together to create six sets of animations ranging from
100% male to 100% female. Thirty-six stimuli were created. Examples from this set are shown in
Figure 5.3.
Figure 5.3. Partial image sequences from six different non-emotional face video stimuli. The top and
bottom videos show two different individuals making the same motion (eyebrow raise). These videos
were then morphed together to create four new videos which vary on a continuum from person A to
person B.
56
Dynamic Car Stimulus Set
In order to evaluate the specificity of face processing deficits, a set of non-face dynamic
stimuli were also created. Side-views of cars were selected because, like faces, they are composed of a
fixed configuration of features (e.g.: wheels, windows, headlights), but do not appear to invoke face-
specific processing networks nor tap into emotional responses in the same way as faces. Previous
research suggests that, unlike viewing faces, viewing cars is not associated with activity in the
fusiform face area (Grill-Spector, Knouf, & Kanwisher, 2004), nor does it elicit the elevated MEG
response component M170 (Xu, Liu, & Kanwisher, 2005).
To create dynamic car videos, 3D meshes of various car models were obtained online via a
free 3D modelling website (Studio ArchiDOM, 2011). These meshes were then recoloured to match
vehicle colour and animated using 3DS Max Design (Autodesk Inc., 2012). Each model was animated
rotating from a side-view to a 45-degree view. Attempts to use morphing to vary the similarity
between cars were unsuccessful. Instead, models were paired with similar looking models in order to
avoid ceiling effects in distinguishing different cars.
In total, six pairs of one-second rotating car animations were created. Each of the 12 cars
appears in two different animations, once rotating left, and once rotating right. Twenty-four stimuli
were created in total. See Figure 5.4 for examples.
Figure 5.4. Partial image sequences from two different 3D video stimuli used in the Car
Discrimination task. In each video, cars rotate from a side view to a 45-degree view. Car 1 and Car 2
are different models that are similar in appearance.
57
Experiment: Static versus Dynamic Emotion Stimuli
The aims of the current experiment were to determine a) whether the newly developed set of
emotional face stimuli will be identified more easily in dynamic form compared to static form, and b)
whether the type of paradigm used (either labelling or discrimination) will interact with the type of
stimulus viewed (dynamic or static). To assess this, a group of healthy undergraduate students
completed an emotion labelling task and an emotion discrimination task which were each composed
of randomly interspersed static and dynamic faces.
Method
Participants
Eighty-two first-year undergraduate psychology students were recruited through the Research
Experience Program (REP) at the Melbourne School of Psychological Sciences, University of
Melbourne. Written informed consent was obtained from all participants, who received course credit
in exchange for their participation. According to self-report all participants were free from
neurological injury, psychological disorder and substance use, and were not taking psychotropic
medications (See Appendices A & B). The study was approved by the University of Melbourne
Human Research Ethics Committee (No. 1135478.6). Due to incomplete or missing data, three
participants were excluded from analysis for the Discrimination task (n=79) and one was excluded
from analyses for the Labelling task (n=81). As all participants performed above 50% accuracy for
both tasks, no individuals were excluded on the basis of poor performance. Demographic data are
shown in Table 5.3.
Table 5.3. Participants’ (n=82) demographic information.
M SD
Age (years) 19.60 3.54 (Range 17-38)
Males/Females (% male) 18/64 (22%) -
Education (years) 12.23 0.76 (9 months)
Estimated FSIQ from NART
(n=54)* 112.88 5.43
Digit span – forward
(mean standard scores) 10.02 2.72
Digit span – backward
(mean standard scores) 11.65 2.92
*28 participants were excluded from this analysis due to poor English (All English as Second Language).
58
Questionnaires and demographics
Participants completed the National Adult Reading Test (NART): a brief reading task that
produces a broad estimate of general intellectual functioning (Crawford et al., 1992; See Appendix C)
. NART error scores were used to calculate Wechsler Adult Intelligence Scale Revised (WAIS-R) IQ
estimates, which are shown in Table 5.3. Participants also completed the Digit Span Forwards and
Backwards subtests from the WAIS-IV (Wechsler, 2008) as a measure of working memory ability.
Demographic information regarding age, sex, and years of education were also collected. As part of a
larger study, The Oxford-Liverpool Inventory of Feelings and Experiences (O-LIFE; Mason &
Claridge, 2006) was also administered to assess schizotypy. Correlations between these scores and
performance on these tasks will be described in Chapter 6.
Emotion Discrimination Task
This task was created using the 60 dynamic emotional face stimuli described above. Each 2-
second animation showed a neutral expression changing into either a fearful or disgusted expression.
Each final expression varied in intensity, showing either 33%, 50%, 67%, 83% or 100% intensity of
emotion. Half of the stimuli were presented in original dynamic form, and half were changed to static
images of the final frame only, presented for 2 seconds continuously. Stimuli were 5 x 4cm and were
viewed at a distance of approximately 50cm (5.7 x 4.6° of visual angle).
In each trial one expression was shown for 2 seconds, then a second face of a different
individual showing either the same or different expression was shown for 2 seconds, followed by a
blank screen with the words “Same or different?” (Figure 5.5A). Participants were instructed to press
‘S’ on a keyboard if the two faces showed the same emotion, and ‘D’ if the two faces showed a
different emotion. Pairs of expressions were always shown at the same intensity level, and always
showed the same stimulus type (either static or dynamic). Each stimulus was shown twice: once in a
‘same’ trial and once in a ‘different’ trial, for a total of 120 trials altogether.
To attempt to control for the possibility that certain stimuli might produce different effects in
the static and dynamic conditions, two different versions were created. In version A half of the faces
were presented as dynamic, and half as static. In version B, the stimuli were reversed, with the
dynamic stimuli shown as static and the static shown as dynamic. Half of the participants completed
version A and half completed version B.
Emotion Labelling Task
Stimuli used were the same as those used in Emotion Discrimination. In each trial an
expression was shown for 2 seconds, followed by a blank screen with the words “Fear or Disgust?”
(Figure 5.5B). Participants were instructed to press ‘A’ on a keyboard if the face showed fear and ‘S’
59
if the face showed disgust. Each stimulus was shown once, for a total of 60 trials altogether. Half of
the trials were static and half were dynamic. Participants completed either version A or B to coincide
with version A or B of the Discrimination task. That is, for each participant, the faces that appeared as
static for the Discrimination task were dynamic for the Labelling task, and vice versa.
Figure 5.5. Trial sequences for Emotion Discrimination (A) and Emotion Labelling (B). Static and
dynamic stimuli were interspersed randomly within each task. Note that for Emotion Discrimination,
both faces within each trial were always the same stimulus type (either static or dynamic). Correct
answers are: A: different, and B: disgust.
General procedure
Participants completed the two tasks in one of four counterbalanced orders (Labelling or
Discrimination first; either version A or B). Prior to each task, participants were shown two easy
practice trials with feedback. If the participant did not answer the two trials correctly, instructions
were repeated and the incorrect trial shown again until the participant understood the task.
Testing took approximately 30 minutes to complete, and participants were permitted to take
as many breaks as desired. Computerised tasks were completed on a laptop computer (60 Hz, 16-inch
screen size, resolution 1280 x 1024 pixels) at a comfortable viewing distance of approximately 50cm
in a quiet, distraction-free environment.
Analytical methods
Accuracy rates and reaction times were analysed for each group. Reaction times were
measure in milliseconds, where t=0 was the onset of the final frame of each trial (i.e.: “Same or
different” or “Disgust or fear”, see Figure 5.5). Participants had an unlimited time to respond. While
percentage correct is a straightforward measure of task performance, group differences may be
60
obfuscated by differences in response bias. For instance, it is often reported that clinical populations
show higher rates of false alarms compared to healthy controls. To minimise response bias, we
converted percent correct on each task to d’ scores using formulae recommended by MacMillan and
Creelman (1991). A higher d’ value indicates more accurate performance. Before analysis, all
participants performing below chance (50% overall accuracy) were excluded.
Prior to calculating d’, hit rates and false alarms were calculated using formulae suggested by
Corwin (1994), which are adjusted to avoid dividing by zero. Hit rates were calculated as: (Correct
hits +.05)/ (Total targets + 1); and false alarm rates were calculated as: (False alarm + .05)/(Total
distracters + 1).
For the Labelling task, d prime was simply calculated as: d’ = z(Hit rate) – z(False alarms).
For the Discrimination task, this value was then converted to a modified d’ using table A5.3 from
MacMillan and Creelman (1991). This was because same-or-different tasks are shown to contain an
inherent bias to say ‘same’ more often than ‘different’ (MacMillan & Creelman, 1991), therefore a
higher level of adjustment is necessary.
A measure of response bias, c, was also calculated using the formula: c = -0.5 [z(Hit
rate)+z(False alarms)](Macmillan & Creelman, 1991). A value of 0 indicates no bias, while a positive
or negative value of c indicates an increasing tendency to favour one response option over the other.
D prime and c values for each group were then compared using t-tests or Repeated-Measures
ANOVA, where appropriate. Uncorrected reaction times (in ms) for correct trials were also analysed.
Results
Impact of emotional intensity
Analyses were conducted to determine whether varying emotional intensity had a differential
impact on static and dynamic conditions in either task. A 2 x 2 x 5 Repeated-Measures ANOVA was
conducted with percent accuracy as the dependent variable and Task (Discrimination versus
Labelling), Stimuli (Dynamic versus Static) and Emotion Intensity (33%, 50%, 67%, 83%, and 100%)
as within-subjects factors. Shapiro-Wilks tests revealed that assumptions of normality were violated
(p<.05) for all 20 conditions. However, upon visual inspection of histograms these were not found to
be extreme violations. Given equal sample sizes and the robustness of ANOVAs to violations,
parametric tests were used for the following analyses. Mauchly’s test indicated that the assumption of
sphericity was violated for Emotion Intensity x2(9)=31.49, p<.001, therefore the Greenhouse-Geisser
correction was used for all analyses involving this factor. Significant main effects were found for
Task, F(1,77) = 138.1, p<.001, ŋp2=.64; Stimuli, F(1,77) = 36.6, p<.001, ŋp
2=.32; and Emotion
Intensity, F(4,261.3) = 51.3, p<.001, ŋp2=.40. Overall, accuracy was higher for the Labelling task
(M=87.4%) than the Discrimination Task (M=72.9%, difference=14.5%, see Figure 5.7A). For both
tasks, accuracy increased as Emotional Intensity increased (See Figure 5.6). Bonferroni-adjusted post-
61
hoc pairwise comparisons revealed a significant jump in accuracy from the 33% intensity condition to
the 50% intensity condition (p<.001) and an almost significant jump from the 50% intensity condition
to the 67% intensity condition (p=.056). There was no significant difference in accuracy between the
67%, 83%, and 100% intensity conditions (p values>.99).
A significant interaction between Task and Stimuli was also found, F(1,77) = 11.0, p=.001,
ŋp2=.13. Post-hoc pairwise t-tests revealed that accuracy was significantly higher for Dynamic stimuli
compared to Static stimuli (difference=7.7%) on the Discrimination Task only, t(78)=6.43, p<.001.
No significant difference in accuracy was found between Dynamic and Static stimuli conditions on
the Labelling task, t(78)=1.46, p=.15. No other interactions approached significance.
Figure 5.6. Comparison of percent accuracy by emotional intensity for dynamic and static conditions
on the Discrimination task (A) and the Labelling task (B). Error bars indicate 95% confidence
intervals around means.
62
Figure 5.7. Raw accuracy rates (A) and d prime scores (B) for Dynamic and Static stimuli for the two
tasks: Emotion Discrimination and Emotion Labelling. Error bars indicate 95% confidence intervals.
Response bias – c
Mean values of c (a measure of response bias) ranged from -.10 to .14 across the four
conditions (see Figure 5.8). One-sample t-tests revealed that c was significantly different from zero
for Static conditions on the Discrimination task, t(78)=3.42, p=.001, 95% CI [.06, .23], and the
Labelling task, t(80)=-2.60, p=.01, 95% CI [-.18, -.02]. Mean c values were not significantly different
from zero for either of the Dynamic conditions: Discrimination task, t(78)=-1.2, p=.22, 95% CI [-.13,
.03], Labelling task, t(80)=-1.34, p=.19, 95% CI [-.11, .02]. A 2 x 2 repeated-measures ANOVA with
c as the dependent variable and Task (Discrimination vs Labelling) and Stimuli (Dynamic vs Static)
as within-subject factors was conducted to ascertain whether c values differed significantly between
the four conditions. A significant main effect was found for Task, F(1,77) = 8.17, p=.005, ŋp2=.10; but
not Stimuli F(1,77) = 3.58, p=.06, ŋp2=.04. The interaction between Task and Stimuli was also
significant, F(1,77) = 12.24, p=.001, ŋp2=.14. Post-hoc pairwise t-tests revealed that response bias was
significantly greater for Static compared to Dynamic stimuli on the Discrimination task, t(78)=-3.87,
p<.001; but no difference was found between Stimuli conditions on the Labelling task, t(80)=1.06,
p=.29. For the Discrimination task, participants were more likely to say ‘different’ (rather than
‘same’) when viewing Static faces, but showed no such bias for Dynamic faces. In contrast, although
there was a slight tendency to say ‘disgust’ (rather than ‘fear’) when viewing Static faces on the
Labelling task, this was not significantly different for Dynamic stimuli, which showed zero bias.
63
These results indicate that response criteria differed significantly between conditions. This is
a problematic finding because unequal response criteria may differentially impact accuracy rates
between tasks. For instance, a large response bias (such as a greater tendency to say ‘same’ rather than
different) can cause overall percent accuracy to appear much lower than a task with very low response
bias (Harvey Jr, 2003). For this reason, d prime scores were calculated as an alternative to raw
accuracy rates, and included in all subsequent analyses. D prime is a sensitivity measure which takes
both true-positive and false-positive responses into account, and therefore provides a measure of
performance which is unaffected by response bias (Macmillan & Creelman, 1991).
Figure 5.8. Mean response bias as evidenced by c rates across the four task conditions. Error bars
indicate 95% confidence intervals.
Repeated-measures ANOVA – d prime performance
Given that c values differed between conditions, d prime scores were calculated as a measure
of performance because they are believed to be less affected by response bias than raw accuracy rates.
Therefore, to examine the impact of task and stimulus type on bias-corrected performance rates, a 2 x
2 Repeated Measures ANOVA was conducted with d prime scores as the dependent variable. Task
(Discrimination versus Labelling) and Stimuli (Dynamic versus Static) were included as within-
subjects factors. As no significant interactions were found involving Emotion Intensity, the data was
collapsed across all levels of intensity for the following comparisons. Shapiro-Wilks tests revealed
that assumptions of normality were violated (p<.05) for two of the four conditions, however, once
again visual inspection of histograms suggested that these were not extreme violations, and parametric
tests were continued.
64
Significant main effects were found for Task, F(1,77) = 14.5, p<.001, ŋp2=.16, and Stimuli,
F(1,77) = 38.2, p<.001, ŋp2=.33. Overall, d prime scores were significantly higher for the Labelling
task (M=2.38) compared to the Discrimination task (M=2.02), and were also significantly higher for
Dynamic stimuli (M=2.39) compared to Static stimuli (M=2.03; see Figure 5.7B). A significant
interaction between Task and Stimuli was also revealed, F(1,77) = 6.2, p=.015, ŋp2=.08. This was
explored further using post-hoc pairwise t-tests which showed that, while d prime scores were
significantly higher for Dynamic stimuli compared to Static stimuli on both tasks, this difference was
larger for the Discrimination task (difference = .53), t(78)=6.20, p<.001, than for the Labelling task
(difference = .21), t(80)=2.14, p=.04. There was no difference in performance between dynamic
conditions on the Discrimination and Labelling tasks, t(78)=1.60, p=.11.
Repeated-measures ANOVA – reaction time data
A 2 x 2 Repeated Measures ANOVA was conducted with reaction times in milliseconds as
the dependent variable, and Task (Discrimination versus Labelling) and Stimuli (Dynamic versus
Static) as within-subjects factors. No significant interactions were found involving Emotion Intensity,
therefore the data was collapsed across all levels of intensity. Shapiro-Wilks tests revealed that
assumptions of normality were violated (p<.05) for all four conditions, however, once again
parametric tests were continued because visual inspection of histograms revealed no extreme
violations. Reaction time data are shown in Figure 5.9.
Figure 5.9. Comparison of reaction times for static and dynamic conditions on the two tasks: Emotion
Discrimination and Emotion Labelling.
65
Main effects did not approach significance for either Task, F(1,77) = .32, p=.57, or Stimuli,
F(1,77) = 2.34, p=.13. However, a significant interaction was found between Task and Stimuli,
F(1,77) = 9.9, p<.001, ŋp2=.21. Post-hoc pairwise t-tests showed that reaction times were significantly
slower for Dynamic stimuli compared to Static stimuli on the Discrimination task (difference = 167
ms), t(78)=4.65, p<.001. In contrast, there was no significant difference between Dynamic and Static
conditions on the Labelling task, (difference = 69ms), t(80)=1.70, p=.09.
Taken together with the results of the d’ analyses, this pattern of results suggests that
performance on the Discrimination task was slower, but more accurate for Dynamic stimuli compared
to Static stimuli. To explore the possibility of a speed-accuracy trade-off, four uncorrected Pearson
correlations were conducted (see Figure 5.10). These revealed no significant linear relationship
between performance (d’) and reaction time for any of the four conditions (p values=0.7-75). Visual
inspection of scatterplots (see Figure 5.10) show no evidence of a non-linear relationship between
accuracy and speed on any of the four conditions. Therefore, it appears unlikely that participants who
were faster on the tasks were likely to make more errors.
Impact of task version
To assess whether the version of tasks completed had any impact on task performance (d’) or
reaction times, a series of independent t-tests were conducted. It was found that participants who
completed version B of the tasks were significantly more accurate at labelling static faces than
participants who completed version A, t(79)=3.77, p<.001. No other comparisons approached
significance (p values=.21-.97). This suggests that the faces shown as static in version A were more
difficult to label than the faces shown as static in version B. However, there was no difference in
version A or version B faces on the dynamic conditions, or in the Discrimination task for either
accuracy or reaction times.
66
Figure 5.10. Scatterplots plotting the relationships between accuracy (d’) and reaction times (ms) on
each of the four task conditions: Dynamic Discrimination (A), Static Discrimination (B), Dynamic
Labelling (C) and Static Labelling (D). Uncorrected Pearson correlations are shown in the upper right
corner of each plot.
Correlations with demographics
Pearson correlations were conducted to determine whether performance (d’) on either the
Discrimination or Labelling tasks were influenced by demographic factors. Correlations were initially
run separately for static and dynamic conditions. As coefficients did not differ, data was collapsed
across conditions. Coefficients are summarised in Table 5.4. No significant correlations were found
between age, years of education, FSIQ estimates, Digit Span scores, and performance on either of the
emotion tasks (p values=.28-.99). However, it was found that performance on the two tasks was
positively and significantly correlated, r=.30, p=.008, CI[.08, .49]. This finding indicates that
67
participants who performed well on the Emotion Discrimination task also performed well on the
Emotion Labelling task.
Table 5.4. Pearson correlations between task performance and demographic factors.
Emotion Discrimination (d’) Emotion Labelling (d’)
r (p) [95% CI] r (p) [95% CI]
Age -.07 (.53) [-.29, .15] .02 (.85) [-.20, .24]
Years of Education .004 (.97) [-.22, .22] .05 (.66) [-.17, .27]
Estimated FSIQ from
NART* -.002 (.99) [-.28, .28] .10 (.48) [-.18, .36]
Digit span - forward .08 (.47) [-.14, .30] .02 (.86) [-.20, .24]
Digit span - backward .005 (.96) [-.22, .23] -.12 (.28)
[-.33, .10]
Emotion Discrimination
(d’) - - .30 (.008)* [.08, .49]
Emotion Labelling (d’) .30 (.008)* [.08, .49] - -
*28 participants were excluded from this analysis due to poor English (All English as Second Language),
leaving n=54; **p<.01.
Discussion
The goal of this experiment was to investigate whether a newly developed face stimulus set
would be recognised more accurately in dynamic form compared to static form, using two different
types of emotion processing tasks: discrimination and labelling. Results showed that, as anticipated,
performance was significantly higher for dynamic stimuli across both tasks. While performance was
higher overall for the labelling task, there was no difference in performance between Labelling and
Discrimination when comparing the two dynamic conditions. Both tasks produced a similar increase
in accuracy for stimuli shown at a higher emotional intensity, and this effect was comparable for both
dynamic and static stimuli. Interestingly, task performance did not correlate with any demographic
factors such as age, years of education, working memory ability, or estimated IQ.
Dynamic versus static emotions
The finding that dynamic facial expressions were recognised more easily than static
equivalents is consistent with a number of past studies comparing these stimuli in healthy populations
(Ambadar et al., 2005; Biele & Grabowska, 2006; Fujimura & Suzuki, 2010; Montagne et al., 2007;
Recio et al., 2011). This result indicates that participants recognise fear and disgust more readily when
68
it is presented as a moving stimulus rather than a static image. This is in line with the notion that
perceived motion is a central component of emotion recognition, and that studies relying on static
stimuli alone may be overlooking the importance of this contribution (Alves, 2013). However, an
alternative explanation is that the dynamic advantage is not due to the motion itself, but the simple
fact that the dynamic stimulus provides more information than a single frame. While this explanation
cannot be disregarded, a study by Ambadar and colleagues (2005) suggests that the dynamic
advantage remains even when the quantity of information is controlled for (i.e. by comparing a video
to a sequence of static images interspersed with masks to disrupt the perception of motion). Therefore,
it does not seem likely that the results seen in the current study are due to factors unrelated to the
perception of motion.
Discrimination versus labelling paradigm
To the best of my knowledge, this is the first study to compare static and dynamic faces on an
emotion discrimination task. Interestingly, this task not only showed a dynamic advantage, but this
advantage was more pronounced than that shown in the labelling task. It could be argued that this
discrepancy is due to an underlying difference in difficulty between the two tasks, as accuracy was
higher for the labelling task overall. However, there was no difference in accuracy rates between the
dynamic conditions for each task, which suggests that the two tasks were reasonably well matched in
difficulty for the dynamic stimuli. One could argue, instead, that performance on the discrimination
task was more affected by a loss of motion than performance on the labelling task. In other words,
same-or-different discrimination tasks may simply be a better measure of motion-sensitive emotion
processing ability. Further research is clearly required to investigate this idea, and is beyond the scope
of this thesis. Nevertheless, it would be informative to know if this result is replicated in other studies.
For instance, the dynamic advantage has not been unanimously found across studies, and it could be
due to the paradigm used. In contrast to the experiments cited above, some studies using dynamic
labelling paradigms have reported no dynamic advantage (Fiorentini & Viviani, 2011) or only found
an advantage for certain emotions, such as anger or happiness (Fujimura & Suzuki, 2010; Recio et al.,
2011). It is possible that labelling tasks may simply be less effective at tapping into emotion
processing deficits than other paradigms, such as discrimination.
Impact of emotional intensity
This study showed that accuracy increased with higher emotional intensity for both the
discrimination and labelling tasks. Similar patterns were shown for both dynamic and static stimuli,
with no evidence of an interaction between intensity and type of stimulus. This finding is consistent
with the study by Hargreaves and colleagues (2016) who reported similar patterns using dynamic
expressions in an emotion labelling task. It is also consistent with studies using static stimuli of
varying emotional intensity (Bediou, Krolak-Salmon, et al., 2005; Chen et al., 2012; Norton et al.,
69
2009). On closer inspection, this increase in accuracy in the current study appears to be driven by
jumps between the 33% and 50% intensity conditions, and between the 50% and 67% conditions.
Accuracy did not differ between the 67%, 83% and 100% conditions for either static or dynamic
stimuli. This result is unlikely to be due to ceiling effects, as raw accuracy in these conditions was
below 80% for the discrimination task and below 90% for the labelling task. Therefore, it is possible
that participants did not obtain more useful visual information from the most intense expressions.
Study limitations: A possible speed-accuracy trade-off for emotion discrimination?
The results of this experiment revealed that participants were more accurate at discriminating
between dynamic facial expressions than static expressions. However, they were also significantly
slower at distinguishing dynamic expressions compared to static. This raises the possibility of a
speed-accuracy trade off; namely, that participants sacrificed accuracy for speed of response on static
trials. However, our analyses suggest that this was not the case: Individuals who responded faster
were not less accurate compared to individuals who responded more slowly. So, what could account
for the slower response times for dynamic stimuli? One possible explanation is that participants
simply required more time to process dynamic stimuli before making a decision. Static stimuli
presented identical information for the entire 2-second presentation time, and (while participants were
not able to respond until after the second stimulus had disappeared), they may well have already made
their decision before the presentation ended. In contrast, dynamic stimuli present their information
more gradually, and participants may have needed to view the entire video before obtaining enough
information to make their decision with certainty. This difference in the timing of information may
possibly lead to a significant delay in making same-or-different decisions about dynamic stimuli
compared to static stimuli.
Conclusions and implications
This experiment compared dynamic and static versions of face stimuli across two different
emotion processing tasks. Using a newly developed face stimulus set, it was found that dynamic
stimuli were processed more accurately than static faces for both the same-or-different discrimination
task and an emotion labelling task. These results showed that the lesser-utilised discrimination task
may in fact be a better measure of motion-sensitive emotional information than the typical emotion
labelling task. Given the demonstrated ecological validity of dynamic stimuli (e.g. Alves, 2013), there
is a strong motivation to use video-based stimuli in future emotion processing studies. The results of
the current experiment indicate a clear dynamic advantage in a newly-developed set of video-based
expressions that vary across emotional intensities. This suggests that these new stimuli are a more
effective and more ecologically-valid tool to assess emotion processing deficits compared to
traditional measures.
70
Chapter 6: Are these tasks sensitive to schizophrenia-like traits?
Schizotypy in Healthy Controls and the Relation to Emotion-
processing Performance
“Schizotypy” is a personality construct that indicates the presence and degree of sub-clinical
behavioural traits similar to those in schizophrenia – such as hallucinations, unusual beliefs or
affective difficulties – in otherwise healthy individuals (Mason, Claridge, & Jackson, 1995). These
schizophrenia-like traits have been associated with emotion-processing deficits in previous literature.
In the previous chapter, a new set of dynamic tasks were developed and verified in order to overcome
some of the limitations of traditional face tasks. A subsequent aim was to determine whether the
newly-developed dynamic stimuli are in fact sensitive to schizophrenia-like traits in healthy controls.
The following chapter will review studies examining face and emotion-processing impairments
associated with schizotypal traits. It will then present an experiment investigating whether different
aspects of schizotypy are associated with performance on the tasks introduced in the previous chapter.
Schizotypy in the general population
Considerable controversy surrounds the question of whether psychosis is better described as a
categorical or dimensional phenomenon. Studies show that the symptoms of psychosis, such as
hallucinations (Lincoln & Keller, 2008), delusions (Rossler et al., 2007), and social anhedonia
(Hanssen, Bak, Bijl, Vollebergh, & van Os, 2005), are not limited to patients, but are distributed
continuously throughout the general population. Therefore, many argue that schizophrenia represents
an extreme position on a continuum (or multiple continua) of normal experience (Claridge, 1997;
DeRosse & Karlsgodt, 2015). Others, however, argue for a quasi-dimensional view in which the
schizophrenia spectrum disorders exist as a discrete category separate from healthy individuals
(Lawrie et al., 2010; Meehl, 1990). These authors argue that the existence of a continuous distribution
of symptoms in the general populations does not erase the qualitative boundaries between clinical
illness and healthy functioning (Lawrie et al., 2010). Regardless of which standpoint is adopted, the
study of subthreshold symptoms in clinically healthy samples permits researchers to explore the
relationships between symptoms, and to identify possible predictors of conversion to schizophrenia
(Phillips & Seidman, 2008). Furthermore, studies of healthy populations can eliminate common
confounds in clinical research such as the impacts of antipsychotic and other psychotropic medication,
illness chronicity, adversity and disease stigma.
The presence of sub-clinical features of psychosis in otherwise healthy individuals has been
variously termed ‘schizotypy’, ‘psychosis-proneness’, and ‘psychoticism’ (Mason & Claridge, 2006).
A distinction can be drawn between three main types of samples: those with genetic or familial risk
(such as unaffected relatives of persons with schizophrenia), those exhibiting subthreshold
71
behavioural symptoms of psychosis (i.e.: schizotypal personality traits), and those that meet formal
diagnostic criteria for Schizotypal Personality Disorder (Kohler et al., 2014). Schizotypal personality
traits include a variety of characteristics such as magical ideation, superstitiousness, referential
thinking, paranoia, and social anxiety (Vollema & Hoijtink, 2000).
Schizotypy is typically measured using self-report questionnaires such as the Schizotypal
Personality Questionnaire (SPQ, Raine, 1991). On the basis of factor analytic studies, there is a
general consensus that schizotypy consists of three factors: positive schizotypy (cognitive-perceptual
features), negative schizotypy (affective and interpersonal difficulties), and cognitive disorganisation
(Venables & Rector, 2000; Vollema & Hoijtink, 2000). This mirrors the three-factor model of
schizophrenia (i.e.: positive, negative and disorganised symptoms; Mason & Claridge, 2006), and
lends further support to the idea that schizotypy and schizophrenia spectrum disorders exist on a
shared continuum. Furthermore, individuals that score highly on measures of schizotypy have a
heightened risk of developing a formal psychotic disorder (Kelleher & Cannon, 2011). For instance,
the Dunedin birth cohort study found that children who reported psychotic experiences at age 11 were
5-16 times more likely to be diagnosed with a schizophrenia spectrum disorder (Poulton et al., 2000).
Similarly, a large-scale longitudinal study by Hanssen and colleagues (2005) found that of the 2% of
people who reported a recent psychosis-like experience, 8% were diagnosed with a schizophrenia
spectrum disorder within the next two years. Such experiences alone cannot be considered a
prodromal stage for clinical psychosis, however, as the overwhelming majority of individuals do not
go on to meet diagnostic criteria (Hanssen et al., 2005). Rather, other factors must play a role in the
development of, or conversion to, clinical psychosis.
Emotion-processing impairments in schizotypy
Research suggests that impairment in emotion processing is associated with illness severity in
schizophrenia (Couture, Penn, & Roberts, 2006; Kohler et al., 2010). This raises the question, could
emotion processing impairments serve as an endophenotypic marker of future illness in healthy
individuals? A range of studies support the presence of an emotion processing deficit in individuals
who score highly on schizotypy measures, albeit to a lesser degree than in clinical populations
(Kohler et al., 2014). For instance, high schizotypy in adults has been associated with a reduced
ability to name or distinguish between facial expressions, particularly for “threat” expressions such as
anger and disgust (Brown & Cohen, 2010; van 't Wout, Aleman, Kessels, Laroi, & Kahn, 2004;
Williams et al., 2007). One large scale study (n>1500) found that scores on the Schizotypal
Personality Questionnaire-Brief predicted performance on a facial emotion labelling task, but not
performance on an identity discrimination task or sex labelling task (Germine & Hooker, 2011). This
negative relationship persisted even when the highest scorers (≥16 out of a maximum 22) were
excluded, indicating that the relationship was unlikely to be driven by a small group of individuals at
very high risk of psychosis. Overall, these findings suggest that trait schizotypy is related to facial
72
emotion processing, and cannot be accounted for by more general face processing abilities. Studies of
adolescent populations have shown a similar pattern, with high psychosis-proneness scores found to
be associated with poorer facial emotion recognition, as well as difficulties recognising their own
emotions, independent of individual differences in intellectual functioning (Roddy et al., 2012; van
Rijn et al., 2011). However, it should be noted that some studies failed to find any relationship
between schizotypy and emotional processing in healthy adults (Jahshan & Sergi, 2007; Shean, Bell,
& Cameron, 2007; Toomey, Seidman, Lyons, Faraone, & Tsuang, 1999) or children (Thompson et al.,
2011).
Aside from behavioural performance on emotion recognition tasks, schizotypy has also been
associated with a range of other indicators of anomalous emotion processing. For instance, adults with
high total schizotypy scores are more likely to show an unusual fixation on the left side of a face
during initial saccades (Leonards & Mohr, 2009) and show reduced right-hemisphere performance
during a lateralised emotion-naming task (Mason & Claridge, 1999). Seiferth and colleagues (2008)
found that high schizotypy individuals showed atypical patterns of neural activation during an
emotion recognition task, including increased activation for neutral faces. These differences were
significant despite normal behavioural performance in these participants. Similarly, abnormal
neurophysiological responses have also been reported. Batty and colleagues (2014) found
significantly reduced N170 amplitude in high schizotypy individuals when viewing inverted faces,
suggesting abnormalities in the holistic processing of face stimuli. Taken together, these studies offer
broad support for a schizophrenia-like neural processing deficit in healthy individuals with
subthreshold symptoms of psychosis.
Researchers have also examined the relationship between emotion processing and specific
factors of schizotypy. Results are highly varied, however, possibly due to a lack of consistency in the
measures used to assess both emotion processing and schizotypy. Three studies using static emotion
tasks reported a significant relationship between emotion processing deficits and the positive
schizotypy factor (Kerns, 2005; Shean et al., 2007; van 't Wout et al., 2004), while two others found
associations with the negative schizotypy factor only (Abbott & Green, 2013; Williams et al., 2007).
Another study by Brown and Cohen (2010) found that both disorganised symptoms and lower overall
quality of life were associated with reduced accuracy for identifying neutral faces. At present, only
one study has examined individual factors on schizotypy using a dynamic (video-based) emotion
processing task. Abbott and Byrne (2013) found that only positive schizotypy was associated with
overall task performance, however, negative schizotypy was associated with the number of errors
made when recognising positive emotions.
It has been suggested that positive (cognitive-perceptual), negative (affective and
interpersonal) and disorganised factors of schizotypy may impact emotion processing in two different
ways. That is, positive symptoms may produce impairment through misattribution of emotion (i.e.: a
73
bias towards perceiving negative emotions) while negative and disorganised symptoms may produce
errors of omission (i.e.: a failure to recognise positive emotions) (Abbott & Byrne, 2013).
The Oxford-Liverpool Inventory of Feelings and Experiences (O-LIFE)
As alluded to above, schizotypy has been previously measured using a range of different self-
report questionnaires. A likely contributor to the inconsistency in past studies is that these
questionnaires vary considerable in content and scope. For instance, the SPQ (Raine, 1991) covers a
broad range of items and is modelled closely on the DSM symptoms of schizotypal personality
disorder. In contrast, other scales such as the Perceptual Aberration Scale (Chapman, Chapman, &
Raulin, 1978) focus only on specific aspects of schizotypy. One criticism of the SPQ is that it relies
heavily on the reporting of clinical symptoms, and may be less sensitive to detecting trait-based
features of schizotypy (Mason & Claridge, 2006). In order to create a measure intended specifically
for use in non-clinical populations, Mason and colleagues (1995) developed the Oxford-Liverpool
Inventory of Feelings and Experiences (O-LIFE). The items of this questionnaire were based on
extensive factor analysis of fifteen pre-existing schizotypy or psychosis-proneness scales (Claridge et
al., 1996). The O-LIFE comprises four scales: Unusual Experiences (perceptual and cognitive
features, i.e. the positive schizotypy factor), Cognitive Disorganisation (relating to attention,
concentration and decision-making), Introvertive Anhedonia (broadly relating to the negative
schizotypy factor), and a new fourth factor, Impulsive Nonconformity (including features such as
disinhibited behaviour, recklessness, and a desire to avoid a conforming lifestyle) (Mason & Claridge,
2006).
The O-LIFE has been widely used in clinical and non-clinical samples across a range of areas
including neuropsychological functioning (Avons, Nunn, Chan, & Armstrong, 2003; Rawlings &
Goldberg, 2001), latent inhibition (Schmidt-Hansen, Killcross, & Honey, 2009; Shrira & Tsakanikos,
2009), and emotion processing (Hoshi, Scoales, Mason, & Kamboj, 2011; Najt, Bayer, & Hausmann,
2012). Other benefits of the O-LIFE scale are that it has well established normative data (Mason &
Claridge, 2006; Mason et al., 1995), high internal consistency (α=.72-.89 for all scales; (Claridge,
1997; Mason et al., 1995), and good test-retest reliability (>.70 for all scales; Burch, Steel & Hemsley,
1998) . For these reasons, the O-LIFE was considered the most appropriate instrument for this study.
Experiment: Schizotypy and processing of dynamic versus static faces
An additional goal of this thesis was to examine whether schizotypy was associated with face-
processing performance in healthy controls using our dynamic emotion processing tasks. To
accomplish this, data from the 82 healthy controls who participated in the experiment described in
Chapter 5 was analysed in relation to their scores on the O-LIFE scale. It was predicted that both
74
positive and negative aspects of schizotypy would be negatively correlated with emotion processing
performance.
Method
Protocol and method for this study can be found in Chapter 5. Demographics for the 82
healthy controls are repeated below in Table 6.1. After finishing the computerised tasks, participants
completed a slightly modified version of the O-LIFE (See Appendix D). This version consists of the
same 147 items introduced by Mason and colleagues (1995), but rather than responding “yes” or “no”,
participants were instructed to rate each item on a Likert scale ranging from 1 (“Not at all”) to 6
(“Very much”). The purpose of this modification is to increase the sensitivity of the measure, as it
allows participants to partially endorse some of the more unconventional items that they would
otherwise respond “no” to in a binary response format. For example, the item “Do you believe in
telepathy?” has a low rate of positive endorsements. This response format has been used successfully
in previous studies (Corcoran, Cropper, Wong, Rutherford, & Groot, 2016; Cropper, Groot, Corcoran,
Bruno, & Johnston, 2017; Cropper, Johnston, & Groot, 2015). Additionally, these Likert scores can be
easily converted to typical scores to be compared to Mason and colleagues’ (1995, 2006) normative
data.
Table 6.1. Healthy controls’ (n=82) demographics and mean raw scores for the O-LIFE scales.
M SD
Age (years) 19.60 3.54 (Range 17-38)
Males/Females (% male) 18/64 (22%) -
Education (years) 12.23 0.76 (9 months)
Estimated FSIQ* 112.88 5.43
Digit span - forward 10.02 2.72
Digit span - backward 11.65 2.92
OLIFE scales
Unusual Experiences 77.21 26.20
Cognitive Disorganisation 79.91 18.11
Introvertive Anhedonia 92.54 12.40
Impulsive Nonconformity 66.89 10.42
Total 316.55 52.50
*28 participants were excluded from this analysis due to poor English (All English as Second Language),
leaving n=54.
75
Results
Correlations with demographics
Mean scores for the O-LIFE and participant demographics are shown in Table 6.1. Responses
for two of the four O-LIFE scales, Unusual Experiences and Introvertive Anhedonia, were not
normally distributed (Shapiro-Wilks test p<.05) therefore Spearman rank correlations were conducted
in these instances. Bootstrapping was used to calculate bias corrected and accelerated (BCa)
confidence intervals using 1000 resamples (DiCiccio & Efron, 1996).
Spearman’s rank correlations revealed that none of the O-LIFE scales correlated significantly
with Age (rs=.-.17 to .02) or Years of Education (rs=-.04 to .10). However, Estimated IQ scores were
negatively correlated with both Unusual Experiences (rs=-.28, 95% CI[-.50, -.04], p<.05) and
Cognitive Disorganisation (rs=-.29, 95% CI[-.54, -.04], p<.05). As FSIQ was not significantly
associated with task performance (see Results in Chapter 5), however, this was unlikely to affect the
interpretation of these results.
Table 6.2. Spearman rank correlations between face-processing performance and O-LIFE scores for
healthy controls (n=82).
Unusual
Experiences
Cognitive
Disorganisation
Introvertive
Anhedonia
Impulsive
Nonconformity
rs
[95% CI]
rs
[95% CI]
rs
[95% CI]
rs
[95% CI]
Static Emotion
Discrimination
.-.05
[-.29, .18]
.10
[-.14, .32]
-.16
[-.38, .10]
.12
[-.12, .36]
Dynamic Emotion
Discrimination
-.02
[-.26, .21]
.04
[-.21, .27]
.03
[-.22, .30]
.14
[-.12 .39]
Static Emotion
Labelling
-.21
[-.42, .03]
-.20
[-.43,.04]
-.06
[-.31, .20]
-.18
[-.39, .05]
Dynamic Emotion
Labelling
-.31*
[-.49, -.11]
-.08
[-.30, .15]
-.06
[-.30, .18]
-.17
[-.40, .08]
Note: * indicates a significant correlation, where the 95% confidence interval for rs does not include 0.
Correlations with task performance
Pearson correlations revealed significant negative correlations between total schizotypy
scores on the O-LIFE and performance on the Emotion Labelling task, both for static (r=,-.28, 95%
76
CI[-.45, -.08]) and dynamic conditions (r=-.22, 95% CI[-.39, -.05]). Correlations with Emotion
Discrimination did not approach significance (rs =-.01 to .05).
Table 6.2 shows correlations between O-LIFE scales and performance across each of the four
task conditions. The only significant result was a negative correlation between Unusual Experiences
and Emotion Labelling for the dynamic condition only. No other associations approached
significance.
Discussion
The aim of this experiment was to determine whether certain aspects of schizotypy would be
associated with emotion processing ability in a sample of healthy individuals. It was found that total
schizotypy scores were associated with performance on Emotion Labelling, but not Emotion
Discrimination. Contrary to expectations, only the positive schizotypy subscale, Unusual Experiences,
was correlated with task performance, and only for one condition of the four: dynamic Emotion
Labelling.
The finding that individuals with higher total schizotypy scores performed more poorly on a
test of emotion perception is consistent with several past studies using a range of schizotypy scales
(Brown & Cohen, 2010; Germine & Hooker, 2011; van 't Wout et al., 2004; Williams et al., 2007).
Interestingly, this correlation was found for an Emotion Labelling task only, and was not reflected in
an Emotion Discrimination task. This discrepancy may be due to differences in task demand between
the Labelling and Discrimination paradigms. For instance, one previous study also reported
significant correlations between schizotypy and performance on a labelling task, but not a
discrimination task (Williams et al., 2007), while two prominent studies reporting significant findings
used labelling tasks only (Brown & Cohen, 2010; van 't Wout et al., 2004). Only one study could be
found which reported significant correlations with schizotypy using an emotion discrimination
paradigm (Germine & Hooker, 2010), although this may be due to the much larger sample size used
in their study (n>1500). Nevertheless, our data indicates that individuals high in schizotypy are more
likely to show impairments in the ability to correctly identify emotional expressions, but are less likely
to show an impaired ability to distinguish between two different expressions.
The current study showed that the positive schizotypy subscale (the Unusual Experiences
scale from the O-LIFE) was negatively correlated with performance on the Emotion Labelling task,
but only for dynamic faces, not static faces. This finding is broadly consistent with previous research
indicating associations between the positive schizotypy factor and emotion perception performance
(Abbott & Byrne, 2013; Kerns, 2005; Shean et al., 2007; van 't Wout et al., 2004), however the
majority of these studies used static stimuli. It is unclear why a difference would be found between
the static and dynamic conditions. A possible explanation is that, in line with the literature discussed
77
in the previous chapter, dynamic faces are simply a ‘purer’ measure of emotion processing ability, and
therefore were more sensitive to true emotion processing deficits. Contrary to previous research, the
current study found no evidence of a correlation between negative schizotypy scores (i.e.: the
Introvertive Anhedonia and Impulsive Noncomformity scales from the O-LIFE) and emotion
processing performance (Abbott & Green, 2013; Williams et al., 2007). However, this finding is in
line with other studies which similarly reported no correlation with either negative or cognitive
disorganisation factors of schizotypy (Kerns, 2005; Shean et al., 2007; van 't Wout et al., 2004).
In conclusion, the current study found a negative correlation between total schizotypy scores
and the ability to correctly label emotional expressions. In particular, the positive dimension of
schizotypy was associated with poorer performance for dynamic faces. However, schizotypy was not
associated with the ability to discriminate between different emotions (i.e.: indicate whether two
emotions are same or different). These findings are broadly in line with previous research using static
stimuli. Furthermore, it reaffirms the idea that individuals with high trait schizotypy tend to show
deficits in emotion processing similar to those shown in schizophrenia. These deficits may be more
closely associated with the positive schizotypy factor – analogous to the positive symptoms of
schizophrenia – rather than generalised cognitive symptoms, age, or intellectual ability.
78
Chapter 7: Method for Inpatient Study
The previous chapters outlined the development of a new, more ecologically-valid set of
video-based stimuli for examining face processing abilities in schizophrenia. They then presented a
study demonstrating their effectiveness and sensitivity to schizophrenia-like traits in healthy controls.
The current chapter will now present the Method section for a large inpatient study designed to
investigate face-processing in schizophrenia using these new stimuli. General results, including the
impact of patient demographics on task performance, will also be presented and discussed here.
Method
Participants
Eighty-six psychiatric inpatients were recruited from an acute psychiatry inpatient unit in
Melbourne, Australia. Participants were recruited over an 18-month period to ensure a wide range of
diagnoses were represented. Final diagnoses were obtained from separation reports provided by each
patient’s treating psychiatrist. Based on this diagnosis, patients were categorised into four groups: 36
schizophrenia-spectrum (including schizophrenia, schizoaffective disorder, and first episode
psychosis), 15 bipolar disorder (bipolar-affective disorder with a history of psychotic symptoms), 18
non-psychotic disorders (including bipolar II without psychotic symptoms, major depression,
generalised anxiety disorder, borderline personality disorder [PD], and situational crisis – none of
whom had ever previously experienced symptoms of psychosis) and 17 non-schizophrenia psychotic
disorders (including drug induced-psychosis, major depression with psychosis, borderline PD with
hallucinations, and schizotypal PD). A breakdown of each group by diagnosis is shown in Table 7.1.
Twenty non-clinical controls were recruited via online advertising. All were free from
neurological injury, psychiatric illness and substance use disorder by self-report, and were not taking
psychoactive medication.
Demographics
Prior to testing, each participant completed the Edinburgh Handedness Inventory (Oldfield,
1971), the National Adult Reading Test (NART; Crawford et al., 1992) , and a demographic
questionnaire with questions regarding age, gender, years of education, and additionally for patients,
current medication and illness duration (See Appendices C, E-G). Patients were also assessed using
the Positive and Negative Syndrome Scale (PANSS; Kay, Fiszbein, & Opler, 1987; see Appendix H) .
Healthy controls also completed the Oxford-Liverpool Inventory of Feelings and Experiences (O-
LIFE) as a measure of schizotypy (Mason & Claridge, 2006).
79
Table 7.1. Breakdown of patient diagnoses by group.
Schizophrenia
spectrum disorders
n=36
Bipolar affective
disorder
n=15
Non-schizophrenia
psychosis
n=17
Non-psychotic
disorders
n=18
Schizophrenia
n=20 Bipolar Type I, with
previous or current
psychotic symptoms
n=15
Drug-induced psychosis
n=9
Major depression
n=12
Schizoaffective
n=9
Depression with
psychotic features
n=6
Situational crisis
n=4
First-episode psychosis
n=7
Schizotypal personality
disorder n=1
Dysthymia
n=1
Atypical psychosis
n=1
Bipolar Type II
n=1
(Comorbid borderline
personality disorder n=4)
Tasks
Emotion Discrimination Task
Stimuli used were 2-second videos of faces changing from neutral expressions to expressions
of either disgust or fear. Development of these stimuli is explained in detail in Chapter 5. Original
videos consisted of five unique individuals (3 male, 2 female) each showing one expression of disgust
and one of fear. These ten videos were morphed to create five levels of expression intensity (33%,
50%, 67%, 83%, and 100%), totalling 50 stimuli.
For each trial one expression was shown, then followed by a second face of a different
individual showing either the same or different expression, then followed by a blank screen with the
words “Same or different?” shown until response (Figure 7.1A). Pairs of expressions were always
shown at the same intensity level. One hundred trials were shown in total.
Emotion Labelling Task
Stimuli used were the same as those in Emotion Discrimination, with an additional ten videos
to make a total of 60 animated stimuli. Each expression was shown for 2 seconds, then followed by a
blank screen with the words “Fear or disgust?” shown until response (Figure 7.1B). Sixty trials were
shown in total.
80
Identity Discrimination Task
Stimuli used were one-second videos of faces showing non-emotive facial movements, such
as opening the mouth, raising an eyebrow, or poking out the tongue. The development of these stimuli
was described in Chapter 5. Stimuli subtended approximately 5.72 x 4.58° of visual angle at a
viewing distance of approximately 50 cm. In order to vary the similarity between pairs of models,
video of different individuals (of the same sex) were “morphed” together to create new faces. Six
pairs of unique individuals (3 male, 3 female) were used. Each pair was morphed to create six new
animations ranging from one identity to the other at 20% increments, totalling 36 stimuli.
For each trial, a ‘pure’ face (either 100% person 1 or 100% person 2) was shown, followed by
a second face from the same set that was either 0%, 20%, 40%, 60%, 80%, or 100% different, then
followed by a blank screen with the words “Same or different?” shown until response (See Figure
7C). One hundred and twenty trials were shown in total.
Sex Labelling Task
Stimuli used were identical to Identity Discrimination above, with the exception that each of
the six identities was morphed with an opposite-sex identity instead of a same-sex identity. Six sets of
6 face animations were created, ranging from male to female. Half of the trials were “male” (i.e.:
60%, 80%, or 100% male) and half were “female” (0%, 20%, and 40% male). For each trial, a single
face was shown for 1 second, followed by a blank screen with the words “Male or female?” shown
until response (Figure 7D). Each of the 36 faces was shown twice, totalling 72 trials.
Car Discrimination Task
Stimuli used were 1-second videos of 3D car models rotating from a side view to a 45-degree
view (see Chapter 5 for more information on stimulus development). Twelve unique cars were
animated and paired with similar looking models. For each trial, one car video was shown, then a
second car video was shown, followed by a blank screen with the words “Same or different?” shown
until response (Figure 7E). One hundred and twenty trials were shown in total.
81
Figure 7. Example trials for each of the five tasks: Emotion Discrimination, Emotion Labelling,
Identity Discrimination, Sex Labelling, and Car Discrimination. Correct responses are A: different, B:
disgust, C: different, D: male, E: different.
82
General procedure
Participants completed all five computerised tasks in one of four counterbalanced orders.
Prior to completing each of these five tasks, participants were shown two practice trials with
feedback. If the participant did not answer the two trials correctly, instructions were repeated and the
incorrect trial shown again until the participant understood the task.
For the three Discrimination tasks, participants were instructed to say whether each stimulus
(either face, car or emotion, respectively) was the same or different (see Figures 7A, 7C, & 7E). Half
of these trials were ‘same’ and half were ‘different’. For the two Labelling tasks, participants were
instructed to state whether each face more closely resembled “male” or “female” (Sex) or “fear” or
“disgust” (Emotion), respectively (Figures 7B & 7D). The experimenter logged all verbal responses
using a keyboard, in order to reduce any impact of impulsive or impaired motor response mapping.
Testing took approximately two hours to complete, and participants were permitted to take as
many breaks as desired. Computerised tasks were completed on a laptop computer (60 Hz, 16-inch
screen size, 1280 x 1024 resolution) at a comfortable viewing distance in a quiet distraction-free
environment.
Results were analysed using the software package SPSS version 20. To control for response
bias, percentage correct on each task was converted to d’ scores using formulae recommended by
MacMillan and Creelman (1991). For Sex Labelling and Emotion Labelling (yes/no style) tasks, this
was calculated as: d’ = z(Hit rate) – z(False alarms). For the three Discrimination tasks this value was
then converted to a modified d’ using table A5.3 (Macmillan & Creelman, 1991). To eliminate
problems raised by dividing by zero, Hit Rate and False Alarms were adjusted (Corwin, 1994). A
higher d’ value represents better performance. A measure of response bias, c, was also calculated
using the formula: c = -0.5 [z(Hit rate)+z(False alarms)] (Macmillan & Creelman, 1991). A value of 0
indicates no bias, while a positive or negative value of c indicates increasing tendency to favour one
response option over the other.
Results
The main analyses of this inpatient study will be presented in Chapters 8, 10 and 11. First,
however, this section will compare participant groups on clinico-demographic factors such as age,
sex, and estimated intellectual ability, and illness factors such as duration of illness. This is to ensure
that differences in task performance between diagnostic groups cannot simply be attributed to
differences in these factors.
83
Demographics
Demographic information and symptom ratings are shown in Table 7.2. Pearson’s chi-square
test revealed that the gender makeup of groups did not differ significantly as a function of diagnosis,
X2(4) = 6.18, p=.186. One-way ANOVAs were performed with Group as a between-subjects factor
and Age, Years of Education, and estimated FSIQ as within-subjects factors. A significant effect was
found for Years of Education, F(4,101)=2.50, p=.048. Post-hoc t-tests (Bonferroni corrected) revealed
a significant difference of 2.11 years between the Control and Schizophrenia group (p=.02). FSIQ
estimates were also found to differ between groups, F(4,93)=2.52, p=.047. Post-hoc t-tests
(Bonferroni corrected) revealed a significant difference of 8.38 points between the Control and
Schizophrenia spectrum groups (p=.03). Age did not differ significantly between groups,
F(4,101)=.75, p=.56.
One-way ANOVAs conducted with the four inpatient groups only revealed no significant
group differences in mean duration of illness, F(3,82)=1.42, p=.25, mean daily dose of antipsychotics,
F(3,61)=1.68, p=.18, or daily benzodiazepine dose, F(2,13)=.63, p=.55. Medication status for each
group is shown in Table 7.2.
PANSS subscales
One-way ANOVAs were run with group as IV (excluding healthy controls), and Positive,
Negative, and General Psychopathology scores as DVs. A significant main effect was found for
Positive Symptoms, F(3,82)=18.76, p<.001. Bonferroni corrected post hoc tests revealed, not
surprisingly, that the Non-psychosis group had significantly lower Positive symptom scores than all
other groups (p=.002 to <.001). The Other group trended towards having significantly lower Positive
symptom scores compared with the bipolar group (p=.055). No other group differences approached
significance.
Correlations between performance and demographic factors
Pearson correlations were conducted to determine if age, years of education, or FSIQ
estimates predicted performance on any of the tasks. It was found that increasing age correlated with
worsening d’ for Emotion Labelling (r=-.197, p=.04), but not other tasks. FSIQ estimates produced no
significant correlations. Years of education correlated positively with performance on Emotion
Discrimination (r=.260, p=.007) only. To determine if years of education could account for
differences in performance on the Emotion Discrimination task, the analyses were re-run excluding
participants with fewer than 11 years of education (eliminating 13 patients in the schizophrenia
spectrum group, 5 Bipolar disorder, 6 Other and 6 Non-psychosis). One-way ANOVA revealed a
84
significant mean effect of group F(4,70), p<.001. Bonferroni-corrected post hoc tests showed that the
schizophrenia spectrum group performed significantly lower than the control (p<.001), Non-psychosis
(p=.001) and Other groups (p=.02). The control group also outperformed the Bipolar group (p=.01).
Thus, limited years of education was unlikely to account for the group differences shown on these
tasks.
Table 7.2. Mean participant demographics and questionnaire scores by group (SD in parentheses).
Schizophrenia
spectrum
n = 36
Bipolar
disorder
n = 15
Other
psychotic
disorders
n = 17
Non-psychotic
disorders
n = 18
Healthy
controls
n = 20
Age (years)
34.44 (9.44)
range 19-53
36.60 (14.80)
range 19-65
30.65 (7.61)
range 18-43
36.17 (13.57)
range 19-59
34.05 (10.72)
range 18-56
Males/females
(% male)
23/13
(65%)
9/6
(60%)
12/5
(71%)
6/12
(33%)
12/8
(60%)
Education
(years) 10.89 (2.55)* 11.40 (1.76) 11.65 (3.28) 11.83 (2.48) 13.00 (1.59)*
Premorbid IQa 100.33
(10.93)* 103.62 (8.83) 102.65 (9.94) 105.69 (10.73) 108.70 (6.57)*
Illness duration
(years) 9.12 (7.56) 8.87 (11.07) 4.15 (6.31) 8.45 (9.97) -
Antipsychotic
daily doseb
367.2 mg
(276.2)
358.9 mg
(52.6)
225.9 mg
(115.6)
155.6 mg
(212.4) -
Benzodiazepine
daily dosec
35.00 mg
(26.30)
44.00mg
(35.25)
-
23.13mg
(17.72) -
PANSS
Positive scale 17.28 (5.45) 18.47 (4.17) 14.24 (4.35) 8.56 (1.76)d
Negative scale 11.53 (3.72) 9.13 (1.89) 9.94 (3.46) 12.28 (4.85)
General
psychopathology 31.53 (6.47) 31.60 (4.98) 31.53 (7.05) 31.67 (4.51)
*Bonferroni-corrected t-tests revealed a significant group difference at p<.05; aDue to dyslexia, illiteracy, or poor English proficiency, IQ estimates were not available for four patients in the
schizophrenia spectrum group, two in the other group, and two in the non-psychosis group. bChlorpromazine equivalent dose. cDiazepam equivalent dose. dPositive symptom scores for the non-psychosis group were significantly lower than all other groups (p
values=.002 to <.001).
85
Table 7.3. Medication status for inpatient groups.
Neither
medication
Antipsychotics
only
Benzodiazepines
only
Both Antipsychotics
and Benzodiazepines
Schizophrenia spectrum 1 28 0 7
Bipolar disorder 0 10 0 5
Other psychotic disorders 5 12 0 0
Non-psychotic disorders 12 2 3 1
Correlations between performance and illness factors
Pearson correlations revealed that illness duration correlated negatively with task
performance on Emotion Discrimination (r=-.337, p=.002), Emotion Labelling (r=-.224, p=.04) and
Car Discrimination (r=-.298, p=.005), and trended negatively with Identity Discrimination (r=-.204,
p=.06). Hierarchical regression analyses revealed that, after controlling for Age, Illness Duration
continued to significantly predict performance on Emotion Discrimination (change in R2=.09,
F(1,83)=8.37, p=.005) and Car Discrimination (change in R2=.13, F(1,83)=12.27, p=.001), but no
longer predicted performance on Emotion Labelling (change in R2=.01, F(1,83)=1.27, p=.26). This
suggests that patients with a longer illness duration performed more poorly on all tasks except for Sex
Labelling, regardless of group, however this cannot account for the group differences observed.
Benzodiazepine daily dose (n=16) correlated negatively with performance on Emotion
Labelling (r=-.519, p=.04) and Car Discrimination (r=-.535, p=.033) and trended towards significance
for Identity Discrimination (r=-.458, p=.07). Mean antipsychotic daily dose (n=84) produced no
significant correlations.
To determine if benzodiazepine use could account for the group differences in task
performance, a MANOVA was run excluding the 17 inpatients who had taken benzodiazepines. As
before (see Chapter 10 for full task analyses), a significant main effect was found for group, F(4,83) =
2.50, p<.001 (Pillai’s Trace) and univariate tests still revealed significant, or trending towards
significant effects of group for four of the five tasks: Identity Discrimination (F(4,83) = 2.07, p=.09),
Emotion Discrimination (F(4,83) = 11.22, p<.001), Emotion Labelling (F(4,83) = 7.01, p<.001), and
Car Discrimination (F(4,83) = 3.96, p=.005). This suggests that although benzodiazepine use was
related to poorer performance on some tasks, it cannot account for the group differences reported in
this study.
Correlations with PANSS symptom scores will be presented with further discussion in
Chapter 11.
86
Discussion
The purpose of this section was to explore the impact of demographic and illness factors on
task performance, and to investigate whether these factors could possibly account for differences
between diagnostic groups. Results indicated that the five participant groups did not differ
significantly in age, and the four patient groups did not differ significantly in duration of illness or
medication dose. Despite attempts to match groups on all factors, however, the schizophrenia group
completed significantly fewer years of education, and achieved lower IQ estimates compared to the
healthy control group. No other group differences approached significance.
This discrepancy is not unusual in studies comparing patients with schizophrenia and
unaffected controls (e.g.: Hargreaves et al., 2016; Mothersill et al., 2014; Bediou, Franck et al., 2005;
Bediou et al., 2007), as lower levels of academic achievement and lower performance on measures of
premorbid IQ are both common features of schizophrenia (Morgan et al., 2012; Woodberry, Giuliano,
& Seidman, 2008). In fact, it has been argued that attempts to IQ-match healthy controls to patients
may be creating a false equivalence, as lower academic achievement due to illness factors may cause
IQ tests to underestimate true premorbid IQ scores (Meehl, 1970; Walker & Standen, 2011). To avoid
this, some authors choose to match groups on parental years of education instead, as this is believed
to be a reliable alternative indicator of premorbid IQ (Silverstein et al., 2012). Again, however, this
approach is less effective if parents of patients are also affected by psychiatric illness (Keefe, Eesley,
& Poe, 2005).
Regardless of the matching techniques used, the issue remains that significant group
differences in education and premorbid IQ may undermine or mimic true group differences on
performance-based tasks. Fortunately, the results of this study suggest that premorbid IQ was not
significantly associated with performance on any of the dynamic tasks. Although years of education
was significantly associated with performance on one task (Emotion Discrimination), further analyses
revealed that group differences remained unchanged when patients with lower education levels were
excluded. Therefore, it is highly unlikely that differences in education level could be driving
differences in task performance between groups.
In conclusion, although the schizophrenia group completed fewer years of education and
received lower premorbid IQ scores compared to the healthy control group, these differences are
unable to fully account for the group differences in performance presented in the following chapters.
Furthermore, there were no significant differences in age, education, premorbid IQ, illness duration,
or medication dosage between the four inpatient groups.
87
Chapter 8: So what kind of deficit is it, really? Characterising
Face Processing Deficits in Inpatients with Schizophrenia
As reviewed in Chapter 2, the ability to recognise and interpret facial expressions has been
reliably shown to be impaired in schizophrenia. These impairments have been demonstrated across a
wide range of paradigms, and likely contribute to poorer functional outcome in this disorder.
However, it is unclear whether this deficit is truly specific to emotion-processing, or if it can be better
accounted for by impairments in other domains. For instance, deficits in non-emotional face
processing (identity processing) may also be impaired in schizophrenia, suggesting that the “emotion-
processing” deficit may be better represented as a general deficit in perceiving faces. Alternatively,
deficits on emotion-processing tasks may be due to more basic impairments in visual attention. This
chapter will address the first aim of this thesis: Can we better characterise face processing deficits in
schizophrenia? In other words, can impairments in emotion-processing be explained by more general
deficits in non-emotional face processing or non-face processing? This question will be addressed by
comparing performance in schizophrenia inpatients and healthy controls on emotion-processing,
identity-processing, and non-face processing (car discrimination) using the dynamic tasks introduced
in previous chapters.
Experiment: Main inpatient study
The method for Experiment 1 is described in detail in Chapter 7. This chapter will focus on
results comparing the performance of patients with schizophrenia (n=36) to healthy controls (n=20).
Results including all other patient groups will be presented in Chapters 10 and 11.
Results
Demographics
Demographic information is repeated in Table 8. Detailed analyses regarding demographics
can be found in the previous chapter.
Group differences in task performance
To compare performance of healthy controls and patients with schizophrenia, d’ scores for the five
tasks were compared in a 2 x 5 multivariate ANOVA. As groups showed similar patterns of
performance across morphing levels in the previous analyses, scores were collapsed across morphing
levels for all tasks. A significant multivariate effect of group was found, F(5,48) = 12.01, p<.001,
ŋp2=.56 (Pillai’s Trace). Performance (d’) for patients and healthy controls are shown in Figure 8.1.
Univariate tests revealed significant effects of group for all tasks except Sex Labelling, F(1,52) = .42,
88
p=.52. Healthy controls outperformed patients with schizophrenia on Emotion discrimination, F(1,52)
= 39.46, p<.001, ŋp2=.43, Emotion labelling, F(1,52) = 39.23, p<.001, ŋp
2=.43, Identity discrimination,
F(1,52) = 7.26, p=.009, ŋp2=.12, and Non-face discrimination, F(1,52) = 12.37, p=.001, ŋp
2=.19.
Table 8. Mean participant demographics by group (SD in parentheses).
Schizophrenia spectrum
n = 36
Healthy controls
n = 20
Age (years) 34.44 (9.44)
range 19-53
34.05 (10.72)
range 18-56
Males/females
(% male)
23/13
(65%)
12/8
(60%)
Education (years)* 10.89 (2.55) 13.00 (1.59)
Premorbid IQa* 96.44 (10.55) 104.65 (9.32)
Illness duration (years) 9.12 (7.56) -
Antipsychotic daily dose
(Chlorpromazine equivalent) 367.2 mg (276.2) -
Benzodiazepine daily dose
(Diazepam equivalent) 35.00 mg (26.30) -
*Significant group difference at p<.05; aDue to dyslexia, illiteracy, or poor English proficiency, IQ estimates
were not available for two patients in the schizophrenia spectrum group.
Figure 8.1. Performance (d’) of healthy controls and patients with schizophrenia across the five
dynamic tasks. Error bars indicate 95% confidence intervals. *p<.01; ** p<.001.
Task performance in healthy controls
89
Figure 8.2 shows the mean performance of healthy controls for the five tasks. To determine
whether difficulty varied across the five dynamic tasks in healthy controls, a repeated-measures
ANOVA was conducted with accuracy (d’) as the dependent variable and task as a within-subjects
factor. Mauchly’s test showed that the assumption of sphericity was not violated, X2(9)=15.64, p=.08.
A significant main effect of task was found, F(4,76)=7.50, p<.001, ŋp2=.28, indicating that despite
attempts to match task demands, difficulty was not uniform across all tasks. Bonferroni-corrected
post-hoc tests revealed that performance on Identity discrimination was significantly higher than Sex
labelling (p<.001, mean difference =.92), and Emotion discrimination (p=.004, mean difference=.79).
No other comparisons were significant (p values=.07-.99). This suggests that the Identity recognition
task was slightly less difficult compared to the Sex labelling and Emotion discrimination tasks,
however performance across all other tasks was of a comparable level.
Task performance in the schizophrenia group
Mean performance across tasks for the schizophrenia spectrum group are shown in Figure 8.3.
As for healthy controls, a repeated-measures ANOVA was conducted with accuracy (d’) as the
dependent variable and task as a within-subjects factor. According to Mauchly’s test, the assumption
of sphericity was not violated, X2(9)=11.29, p=.26. A main effect of task was found, F(4,132)=53.29,
p<.001, ŋp2=.62. Bonferroni-corrected post-hoc comparisons revealed that performance on the two
emotion tasks was not significantly different from one another (p>.99) but were both significantly
lower than the three remaining tasks (p values>.001, mean differences = .76 – 1.58). The identity
discrimination task was significantly higher than all other tasks (p values<.001, mean differences =
.46 – 1.58). Finally, the car discrimination and sex labelling tasks were not significantly different from
one another (p>.99).
Response bias – c
Mean values of c for healthy controls and patients across tasks are shown in Figure 8.4. Mean
values of c ranged from .04 to .81 across groups and tasks. Independent t-tests revealed no significant
differences between healthy controls and patients for any task (p values=.30-.76), suggesting that
response bias did not differ between groups, and are therefore unlikely to account for differences in
task performance in these groups.
90
Figure 8.2. Performance (d’) of healthy controls across the five dynamic tasks. Error bars indicate
95% confidence intervals. Dots indicate the performance of individual participants. *p<.01, **p<.001.
Figure 8.3. Performance (d’) of patient with schizophrenia across the five dynamic tasks. Error bars
indicate 95% confidence intervals. Dots indicate the performance of individual participants. *p<.001.
91
Figure 8.4. Mean response bias as evidenced by c rates across the five tasks for schizophrenia patients
versus controls. Error bars indicate 95% confidence intervals. For the three Discrimination tasks,
negative values indicate a tendency to say ‘same’. For Emotion Labelling, a positive value indicates a
tendency to say ‘fear’ while a negative value indicates a tendency to say ‘disgust’. For Sex Labelling,
positive values indicate a tendency to say ‘male’.
Impact of morphing: Varying emotional intensity
To examine the impact of varying emotional intensity on the ability to recognise emotions,
and whether this impact differed between groups, a separate 2 x 5 mixed-model ANOVA was
conducted on raw accuracy for each emotion task. Group (control or schizophrenia) was included as a
between-subjects factor and intensity level (33%, 50%, 67%, 83% and 100%) as a within-subjects
factor. Mauchly’s test showed that the assumption of sphericity was violated for both tasks, therefore
the Greenhouse-Geisser correction was used, X2(9)=131.66, p<.001; X2(9)=19.16, p=.02.
Emotion discrimination task
Raw accuracy rates across intensity levels on the emotion discrimination task are shown in
Figure 8.5.A. A significant main effect was found for group, F(1,53)=41.31, p<.001, ŋp2=.44, but not
emotional intensity, F(1.68, 89.40)=2.79, p=.08. The group by intensity interaction did not approach
significance either, F(1.69, 89.40)=.22, p=.93. Bonferroni-corrected post-hoc comparisons revealed
that the control group outperformed the schizophrenia group at every intensity level (p values=.03-
.005, mean differences = 12.9-18.5%).
92
Emotion labelling task
Unlike emotion discrimination, the labelling task permits comparison between different
emotions. To determine whether the valence of emotion affected performance on this task, a 2 x 2 x 5
mixed-model ANOVA was conducted including emotion (fear or disgust) as an additional within-
subjects variable. As no significant main effect (p=.87), or interactions involving this variable (p
values =.26-.98) were found, all subsequent analyses were collapsed across fear and disgust trials.
Raw accuracy rates for emotion labelling are shown in Figure 8.5.B. A 2 x 5 mixed-model
ANOVA was run comparing group and intensity level. The main effects of group, F(1,54)=26.07,
p<.001, ŋp2=.33, and intensity of emotion, F(3.37,182.20)=7.13, p<.001, ŋp
2=.12, were both
significant, but not the interaction, F(3.37, 182.20)=.08, p=.98. Unlike the emotion discrimination
task, accuracy for naming emotions increased somewhat with increasing intensity. Bonferroni-
corrected post-hoc comparisons showed that accuracy was significantly lower for 33% intensity faces
compared to both 67% (p=.01, mean difference =9.3%) and 83% intensity faces (p=.001, mean
difference = 10.6%). Additionally, 50% intensity faces were less accurate than 83% faces (p=.04,
mean difference = 6.4%). No other comparisons approached significance. Like the emotion
discrimination task, post-hoc comparisons showed that the control group outperformed the
schizophrenia group at every level of intensity (p values<.001, mean differences = 16.7-19.1%).
Figure 8.5. Mean accuracy performance for the schizophrenia spectrum (SZ) and control groups for
(A) emotion discrimination and (B) emotion labelling. Emotional intensity is presented on the y axis,
where 100% indicates an unedited expression and 50% indicates an expression morphed 50% with a
neutral expression. Error bars indicate 95% confidence intervals around the mean. *Significant group
difference at p<.05; **p<.001.
93
Impact of morphing: Varying facial identity
To determine whether varying facial identity (including gender) affected performance on the
non-emotional face tasks, a separate 2 x 6 mixed-model ANOVA was conducted on raw accuracy for
each task. Group (control or schizophrenia) was included as a between-subjects factor and morphing
level (0%, 20%, 40%, 60%, 80% and 100%) as a within-subjects factor. As Mauchly’s test of
sphericity was violated for both tasks, Greenhouse-Geisser corrected values were reported,
X2(14)=56.88, p>.001; X2(14)=190.68, p>.001.
Identity discrimination task
Raw accuracy rates for patients and healthy controls on the identity discrimination task are
shown in Figure 8.6A. Significant main effects were found for morphing level, F(3.58,
193.13)=332.13, p<.001, ŋp2=.86, and group, F(1,54)=5.60, p=.02, ŋp
2=.09. The interaction between
morphing level and group was also significant, F(3.58, 193.13)=4.98, p=.001, ŋp2=.08. Pairwise
comparisons indicated that accuracy for distinguishing between faces increased as facial similarity
increased. Bonferroni-corrected comparisons showed that nearly all morphing levels were
significantly different from one another (p values>.001, mean differences = 11.0-85.4%), with the
exceptions that the 80% and 100% conditions did not differ from each other (p<.001), nor did they
differ from the 0% condition (p values=.09-.30). Both groups performed above chance when faces
were at least 60% different, and at or below chance at 40% different. Bonferroni-corrected post-hoc
independent comparisons showed that the healthy controls outperformed the schizophrenia group
when faces were 60% different (p=.02, mean difference = 14.4%) and 100% different (p=.006, mean
difference = 13.6%). There were no significant differences between groups on any other conditions (p
values=.07-.99).
Sex labelling task
Raw accuracy for sex labelling across different morphing levels are shown in Figure 8.6B. A
repeated-measures ANOVA revealed a significant main effect for morphing level, F(1.71,
90.70)=40.72, p<.001, ŋp2=.43, but not group, F(1, 53)=.03, p=.86. The interaction did not approach
significance, p=.59. Unexpectedly, accuracy was reliably higher for male faces than female faces (see
Figure 8.6B). Pairwise comparisons revealed that accuracy for 60% male faces was significantly
lower than 80% male faces (p<.001, mean difference = 16.4%), which was in turn lower than 100%
male faces (p=.04, mean difference = 3.3%). In contrast, accuracy for 60% female faces was
significantly lower than 80% female faces (p<.001, mean difference = 15.8%) there was no difference
between 80% and 100% female faces (p=.20). These results suggest that, as anticipated, faces that
were a morphed mix of male and female faces were correctly identified less frequently than un-
morphed faces (100% male or female). However, there was no significant difference in accuracy
between healthy controls and patients with schizophrenia at any level of morphing.
94
Figure 8.6. Mean accuracy performance for the schizophrenia spectrum (SZ) and healthy control
groups for (A) identity discrimination and (B) sex labelling. Morphing level is shown on the y axis,
where 50% indicates an equal morph between Face 1 and Face 2. Note that for A, accuracy is
calculated as correct ‘different’ responses for all conditions except the 0% condition, where the
correct response is ‘same’. Error bars represent 95% confidence intervals. *Significant group
difference at p<.05.
Discussion
The aim of this experiment was to better characterise face-processing deficits in
schizophrenia using a newly-developed set of dynamic tasks. Specifically, to answer the question:
Can impairments in emotion-processing be explained by more general deficits in non-emotional face
processing or non-face processing? This question was explored by comparing performance of
schizophrenia inpatients and healthy controls on emotion-processing, identity-processing, and non-
face processing (car discrimination) using the dynamic tasks introduced in previous chapters. This
study found that patients with schizophrenia were impaired not only on dynamic emotion-processing
tasks, but also on an identity discrimination task and a non-face discrimination task. In contrast,
performance on a sex labelling task was unimpaired.
Emotion-processing deficits
95
Consistent with previous studies using dynamic face stimuli, this study found significant
impairments in the ability to correctly label emotions in patients with schizophrenia (Archer et al.,
1994; Behere, Venkatasubramanian, Arasappa, Reddy, & Gangadhar, 2011; Hargreaves et al., 2016;
Johnston et al., 2010; Mendoza et al., 2011). It was also found that patients were significantly
impaired on a same-or-different emotion discrimination task compared to healthy controls. Although,
to my knowledge, results have not been previously published using dynamic faces in this task, this
finding is in line with studies using static faces to investigate emotion discrimination in schizophrenia
(Addington et al., 2006; Hooker & Park, 2002; Martin et al., 2005; Penn et al., 2000; Sachs et al.,
2004; Weniger et al., 2004). This suggests that our patients were impaired in both consciously naming
and distinguishing between different emotional expressions.
Identity-processing deficits
For the first time, this study examined facial identity processing in schizophrenia using
dynamic face stimuli. It was found that performance was significantly poorer in patients compared to
controls on a same-or-different discrimination task. This is consistent with several studies which
reported identity discrimination impairments using posed static stimuli (Butler et al., 2008; Martin et
al., 2005; Shin et al., 2008; Soria Bauser et al., 2012), however, others found no difference between
patients and healthy controls (Edwards et al., 2001; Johnston et al., 2010; Soria Bauser et al., 2012).
While it is not clear why this variation exists, it is possibly due to differences in the stimuli used or in
the samples tested (such as differing symptoms or duration of illness). It should be noted, however,
that this difference is no longer significant when analysed with all five groups (results in Chapter 10).
Therefore, this difference may be described as marginal at best.
In contrast, the current study found no performance differences between patients and controls
on a sex labelling task. As average accuracy was approximately 81% for both groups, this lack of
difference cannot be attributed to ceiling effects. Tasks such as this, which require participants to
attend to just one aspect of identity, typically do not show impairments in schizophrenia (Bediou et
al., 2007; Bediou, Krolak-Salmon, et al., 2005; Chen et al., 2012). In their original model of face
processing, Bruce and Young (1986) refer to information such as age or sex as visually derived
semantic codes. These codes are readily extracted from unfamiliar faces and may be derived from
featural cues (such as nose shape), configural cues (such as spacing of features), or a combination of
both. In healthy individuals, decisions about sex tend to prioritise the shape of the eyebrows, nose,
and overall face outline (e.g.: “fleshiness” of the face) (O'Toole et al., 1998). In contrast, decisions
regarding identity draw from a wider range of visual cues, including perceiving the face as a holistic
‘whole’ (Richler et al., 2011). It is possible that distinguishing the sex of a face is simply a less
demanding perceptual task and is unimpaired in schizophrenia despite significant deficits in
distinguishing the identity of a face. An alternative interpretation, however, is that difficulties shown
96
on the identity discrimination task are driven by the same-or-different paradigm itself, rather than a
true identity-processing deficit. The same-or-different paradigm involves serial presentation of two
stimuli and, therefore, requires the viewer to hold the first stimulus in memory for several seconds in
order to mentally compare it to the second stimulus. Therefore, this paradigm may be more sensitive
to more general cognitive deficits, such as working memory, than labelling paradigms which simply
show one stimulus at a time. Consistent with this idea is the finding that patients were also impaired
on a same-or-different task involving non-face stimuli (cars). Therefore, it is likely that impairment on
these tasks reflect cognitive impairments in areas such as attention, processing speed, or working
memory – all of which are commonly affected in schizophrenia (Keefe & Harvey, 2012). Future
studies might aim to explore this further using a same-or-different paradigm that uses simultaneous
presentation, rather than sequential, in order to reduce the demand on working memory.
Is the emotion-processing deficit in schizophrenia simply the result of more general cognitive deficits?
To better characterise face-processing deficits, it is necessary to establish whether identity
processing is specifically impaired in schizophrenia. Although patients with schizophrenia showed
significant performance deficits on an identity discrimination task, the finding that patients were
similarly impaired on a car discrimination task (yet unimpaired at recognising the sex of a face),
strongly suggest that these deficits do not represent specific impairments in processing face
information. Rather, performance on these tasks may be affected by more general cognitive deficits. It
is therefore possible that previous studies reporting identity-processing deficits in schizophrenia may
have similarly been impacted by more general attentional or perceptual impairments which are not
specific to faces. One suggestion is that these impairments are driven by deficits in global visuospatial
processing: the ability to view the gestalt of a visual object which is critical for face recognition
(Macrae & Lewis, 2002) and is reliably shown to be affected in schizophrenia (Poirel et al., 2010).
This question will be further investigated in Chapter 9.
In contrast, this study found significant impairments in both recognising and distinguishing
between emotions in schizophrenia. As performance was impaired even on a simple labelling
paradigm (i.e.: not comparing serially presented stimuli), these deficits cannot be accounted for by the
greater working memory demands of the same-or-different paradigm. Similarly, these deficits are
unlikely to be explained by more general impairments in processing non-emotional aspects of faces,
such as identity. Therefore, these findings agree with past literature (Edwards, Jackson, & Pattison,
2002; Johnston et al., 2010) that indicates a specific deficit in processing emotional expressions in
schizophrenia.
Conclusions
97
This study examined facial processing deficits using a series of novel dynamic tasks. Patients
with schizophrenia showed the expected impairments in both naming and distinguishing between
dynamic emotions. Patients were also poorer at non-emotional face discrimination and non-face
discrimination, but were unimpaired at distinguishing the sex of a face. These findings hint that the
emotion-processing deficits reported in previous studies may represent a more general cognitive or
perceptual impairment that affects performance regardless of stimulus type. Chapter 11 will examine
this idea further by exploring correlations between performance and patterns of symptomatology. The
following section will investigate whether performance on dynamic face tasks could be affected by
more general impairment in the global allocation of visuospatial attention.
98
Chapter 9: Could face-processing impairments in schizophrenia be
mediated by deficits in allocating visuospatial attention?
Patients with schizophrenia are shown to have reduced performance on tasks of emotion
recognition as well as non-emotional face recognition (Darke et al., 2013; Kohler et al., 2010).
However, whether these performance deficits represent a face-specific deficit, or a more general
deficit in cognition or perception remains unclear. One possible account for identity and expression
recognition deficits in schizophrenia is that they are the result of a more general impairment in
allocating visuospatial attention. ‘Global’ versus ‘local’ visual attention refers to the ability to attend
to the overall configuration of a stimulus or its individual details, respectively (Navon, 1977).
Previous studies suggest that patients with schizophrenia show impairments in global attention.
Therefore, a corollary question raised in the course of this PhD is: could deficits in face-processing
performance be explained by difficulties in allocating visuospatial attention? This chapter will present
the results of an additional inpatient study (n=45) designed to address this question.
The role of global and local visuospatial attention in face processing
Global and local visuospatial attention is typically evaluated using compound stimuli, which
consist of a large 'global' letter made up of many small 'local' letters (see Figure 9.1; Navon, 1977) .
Studies of schizophrenia using this, and similar paradigms have revealed impairments in global
processing, but largely preserved local processing both for static (Goodarzi et al., 2000; Johnson et al.,
2005; Poirel et al., 2010; Silverstein et al., 2000) and dynamic stimuli (Chen et al., 2003). In addition,
patients with schizophrenia demonstrate a bias towards attending to the local level of a stimulus, and
are more likely to persist with a local processing strategy even when task demands favour a global
strategy (Landgraf et al., 2011).
It is possible that a global processing deficit could contribute impairments in identity
recognition because the important global-level information is not being processed efficiently. For
instance, it has been shown that identity recognition performance is improved when healthy
participants are primed to adopt a global processing strategy, and impaired when primed with a local
processing strategy (Macrae & Lewis, 2002; Perfect, 2003). In a study by Joshua and Rossell (2009) it
was found that patients with schizophrenia showed less of a reduction in identity recognition
performance compared to controls when configural cues were removed from a face, indicating that
these individuals relied more strongly on local features when identifying famous faces, even when
configural information was available.
99
Figure 9.1. Example of hierarchical stimuli used by Johnson and colleagues (2005, p.938). In this
version of the task, participants are required to respond to the target 'H' whenever it appeared at the
global or local level of a series of hierarchical figures.
Although global processing appears to play an important role in face identity recognition,
there is evidence to suggest that the recognition of emotional expression is less reliant on configural
information. For instance, individuals with congenital prosopagnosia (‘face blindness’) have been
shown to perform normally on emotion recognition tasks and are able to identify the basic expressions
of happiness, anger, disgust, fear sadness and surprise, despite demonstrating severe impairments in
the ability to process configural face information (Kress & Daum, 2003; Palermo et al., 2011;
Schmalzl, Palermo, & Coltheart, 2008). While it has been argued that these individuals may rely more
heavily on compensatory strategies than healthy controls – for example, identifying anger based on
eye shape alone (Palermo et al., 2011) – intact performance has been reported even for subtle
expressions that are difficult to categorise (Duchaine, Parker, & Nakayama, 2003; Humphreys,
Avidan, & Behrmann, 2007).
It has been argued that healthy individuals rely primarily on a few feature-based cues when
identifying emotions (Morris et al., 2009). For example, viewing the eyes alone is a reliable indicator
of sadness, fear and anger (Calder et al., 2000; Martin et al., 2012). Similarly, viewing only the mouth
of a face is adequate for identifying basic expressions of happiness and disgust, and results in greater
accuracy than viewing the eye region (Calder et al., 2000). A study by Bimler and colleagues (2013)
demonstrated that inversion has little impact on discriminating between subtly morphed facial
expressions, suggesting that while holistic processing is critical to identity recognition, it is not
required for distinguishing complex expressions. Similarly, a study by Martin and colleagues (2012)
showed that expression recognition performance was enhanced when participants were primed to
attend to individual features using a locally-directed hierarchical figures task and was impaired when
primed to attend globally.
Taken together it appears that, unlike identity recognition which requires the integration of
global information, expression recognition may be improved when a piecemeal, local processing
100
strategy is employed. However, it must be noted that although a local strategy is advantageous for
certain experimental tasks with very specific instructions (e.g.: is this face happy or angry?), it does
not mean that global processing is not engaged in expression recognition. Martin and colleagues
(2012) note previous research that suggests that local strategies are more advantageous in situations
where conditions are especially challenging, such as when faces are presented very briefly (e.g.: 125
ms) or when image quality is degraded (Macrae & Martin, 2007; Martin & Macrae, 2007). When
these conditions are not imposed however, global processing plays a greater role (Macrae & Martin,
2007). For this reason, the authors state that expression recognition is likely not a solely feature-based
process. A similar stance was expressed by Calder and colleagues (2000). They reported that
interference effects were found when participants were asked to identify the expression shown in the
top half of a face which was joined with a lower face half showing a different expression. In contrast,
when the bottom face half showed a different identity but the same expression, no interference was
found. The authors argue that this interference occurs because each facial expression is associated
with its own average configuration that is extracted in addition to featural cues (Calder et al., 2000).
From the research discussed above, it appears that a feature-based local processing strategy is
adequate, and even advantageous, for many expression recognition tasks. However, this does not
necessarily reflect normal or complete functioning. It is likely that global information also plays an
important role in recognising expressions, particularly for stimuli with longer presentation times
(Martin et al., 2012). At this point, it is unclear whether global processing dysfunction could
adequately account for expression recognition deficits in schizophrenia.
The aim of the present study is to investigate whether global visual processing correlates with
performance on facial identity or emotion recognition in schizophrenia. A group of inpatients with
(n=17) and without schizophrenia (n=14) and a group of healthy controls (n=14) completed three
tasks: an identity discrimination task, an emotion discrimination task, and a Navon task to assess
global and local visual attention. It was hypothesised that performance on the global attention task
would correlate with performance on the facial identity task, while performance on the local task
would correlate with performance on the emotion task. It was also predicted that, in line with previous
studies, patients with schizophrenia would show reduced performance on the global attention task, but
not the local attention task.
Method
Participants
A total of 31 inpatients (12 female, 19 male) were recruited from a psychiatric unit in
Melbourne, Australia. Demographic information is shown in Table 9.1. Based on final diagnoses
made by each patient’s treating psychiatrist, patients were grouped into a schizophrenia spectrum
101
group (comprising 14 with schizophrenia, 1 with schizoaffective disorder, and 2 first-episode
psychosis) and a psychiatric control group (comprising 9 with bipolar disorder, 4 with depression and
1 dysthymia). An additional 14 healthy controls, all without past or present psychiatric illness, were
also tested. All participants were free from visual impairment, traumatic brain injury, or neurological
disease.
Emotion discrimination task
Emotion discrimination was assessed using an identical task and procedure to that described
in the main inpatient experiment (see Chapter 7).
Identity discrimination task
See the Chapter 7 for method and procedure.
Navon task (global vs local attention)
Stimuli
Stimuli were used with permission from McKone and colleagues (2010) . Sixteen
hierarchically-arranged letter stimuli created, each consisting of a large global letter (2.3°x 3.9°)
composed of 7 – 14 small, local letters (0.4°x 0.5°). Target letters were H and T, and non-target letters
were D, E, U, and V. Half of the targets were valid, with a target letter at either the global or local
level, and half were invalid, without a target letter at either the global or local level (see Figure 9.2).
Each stimulus had different letters at the global and local levels, and none contained both targets. A
further six hierarchical stimuli, not presented in the test phases, were shown to participants in a
familiarisation task prior to testing.
Stimuli were presented in black against a white background. The medial edge of each
stimulus appeared at an eccentricity of 0.5° from a fixation cross (0.2°x 0.2°) shown in the centre of
the screen throughout testing. A post-stimulus mask was constructed from small letters (0.25° x 0.25°)
(see Figure 9.2C). The letters were presented in black on a white background. These stimuli were
arranged in a ‘brick wall’ (2.8°x 4.2°). This mask was presented simultaneously to the left and right
portions of the screen, with the medial edges at an eccentricity of 0.5° from the fixation cross.
Design
Participants completed two blocks of 256 trials: one global block and one local block. Block
order was counterbalanced across participants. A short break was permitted halfway through each
block. Each block was preceded by 16 practice trials, which included feedback. The practice trials
were presented with a long stimulus exposure (500 ms), to ensure compliance with task instructions.
Prior to the practice trials, participants completed a familiarisation task. This involved
viewing a Powerpoint presentation in which the experimenter introduced the six different stimuli for
102
each task, with examples demonstrating the global and local form of each. Six example stimuli were
then presented, and the participant was required to identify the two letters present in each stimulus.
Procedure
Participants were instructed to respond, with the index finger of the right hand, every time one
of the targets appeared at either the global or local level of a stimulus. The experimenter emphasised
that the participant was not to respond if a target was not presented. Each trial sequence began with
the appearance of a fixation cross (plus sign, ‘+’) for 1 s, a randomly lateralised stimulus for 150 ms,
followed by the post-stimulus mask for 1.5 s (see Figure 9.2C). The next trial began instantly, and
with no intertrial interval.
Figure 9.2. The Navon task. A: An example of a valid stimulus. The target H is present at the global
level. B: An example of an invalid stimulus. C: Post-stimulus mask used in the Navon task. D:
Sample trial sequence.
General procedure
Prior to testing each participant completed the National Adult Reading Test (NART;
Crawford et al., 1992) , and a demographic questionnaire with questions regarding age, gender, years
of education, and additionally for patients, current medication and illness duration (See Appendix A).
Patients were also assessed using the Positive and Negative Syndrome Scale (PANSS; Kay, Fiszbein,
& Opler, 1987) .
103
Participants completed the computerised tasks in one of four counterbalanced orders. Stimuli
were presented in E-Prime 2.0 on a 14 inch (viewing size 13.3 inch) laptop screen (1366 x 768, 60
Hz) at a comfortable viewing distance. This study received ethics approval from the Melbourne
Health Human Research and Ethics Committee (HREC Project no. 2008.657), and the University of
Melbourne Human Research Ethics Committee (Ethics ID no. 1135478.4).
Results
Demographics
Demographic information for the three groups are shown in Table 9.1. A Pearson’s chi-square
test showed that the groups did not differ in gender makeup, X2(2) = .60, p = .74. One-way ANOVAs
revealed significant group differences for age, F(2,42)=4.48, p=.02, years of education, F(2,40)=8.15,
p=.001, and estimated FSIQ, F(2,41)=12.65, p<.001. Post-hoc Bonferroni-corrected comparisons
showed that psychiatric controls were significantly older than healthy controls (p=.02, mean
difference = 11.9 years), but neither group was significantly different from the schizophrenia group (p
values=.08-.99). Furthermore, healthy controls completed significantly more years of education (p
values=.002-.006, mean differences = 2.6-2.9 years) and had higher estimated IQ scores (p
values<.001, mean differences = 15.1-15.7 points) than the two patient groups, which did not differ
from each other (p values>.99).
To investigate whether differences in age, education, or estimated IQ could account for group
differences in task performance, a series of bivariate Pearson correlations were conducted (see Table
9.2). It was found that age correlated negatively with performance on the two same-or-different
discrimination tasks (identity discrimination: r=-.35, p=.02; emotion discrimination: -.46, p=.002).
Years of education and IQ both correlated positively with emotion discrimination (rs=.42-.46, p
values=.01-.001) and global attentional performance (p values=43-.46, rs=.002-.003). In contrast, no
correlations approached significance for local attentional performance.
Independent samples t-tests were used to compare disease-specific factors for the two patient
groups. These revealed that the schizophrenia group had significantly higher positive symptom scores
on the PANSS compared to psychiatric controls, t(29)=2.18, p=.04, mean difference = 4.1. Groups did
not differ in duration of illness (p=.42), daily antipsychotic dosage (p=.73), daily benzodiazepine
dosage (p=.82), negative PANSS scores (p=.41), or general psychopathology PANSS scores (p=.35).
Uncorrected Pearson correlations revealed that duration of illness was negatively associated with
identity discrimination (r=-.49, p=.005), emotion discrimination (r=-.36, p=.046), and local task
performance (r=-41, p=.02), but not global performance (r=-.16, p=.39). None of the tasks correlated
significantly with antipsychotic dosage (p values=.53-.92), benzodiazepine dosage (p values=.32-82),
104
positive symptoms (p values=.15-.63), negative symptoms (p values=.19-.99), or general
psychopathology (p values=.21-72).
Table 9.1. Mean participant demographics and questionnaire scores by group (SD in parentheses).
Schizophrenia
n = 17
Psychiatric controls
n = 14
Healthy controls
n = 14
Age (years)*
37.65 (9.45)
range 25-60
40.29 (12.29)
range 26-66
28.36 (11.86)
range 20-55
Males/females
(% male)
10/7
(59%)
9/5
(64%)
7/7
(50%)
Education (years)* 11.12 (2.55) 10.64 (1.45) 13.57 (1.91)
Premorbid IQa* 99.21 (11.00) 98.63 (8.94) 114.28 (7.81)
Illness duration (years) 10.12 (6.85) 12.79 (11.21) -
Antipsychotic daily doseb 301.13 mg (212.68) 338.83 mg (257.78) -
Benzodiazepine daily
dosec 11.43 mg (10.69) 10.36 mg (5.83) -
PANSS
Positive scale* 17.00 (5.55) 12.93 (4.66)
Negative scale 15.47 (6.66) 13.71 (4.60)
General psychopathology 32.12 (5.68) 29.93 (7.14)
*Groups are significantly different at p<.05; aDue to illiteracy, an IQ estimate was not available for one patient in the schizophrenia spectrum group. bChlorpromazine equivalent dose. cDiazepam equivalent dose.
105
Table 9.2. Uncorrected Pearson correlations between demographic factors and task performance.
Age Years of Education Estimated IQ
d’ r (p)
[95% CI]
r (p)
[95% CI]
r (p)
[95% CI]
Identity Discrimination -.35, (.02)*
[-.58, -.06]
.21, (.17)
[-.10, .48]
.12, (.46)
[-.18, .40]
Emotion Discrimination -.46, (.002)*
[-.66, -.19]
.41, (.005)*
[.14, .64]
.46, (.001)*
[.19, .67]
Global task -.27, (.08)
[-.27, .03]
.43, (.003)*
[.15, .65]
.46, (.002)*
[.19, .67]
Local task -.28, (.07)
[-.53, .02]
.05, (.77)
[-.25, .35]
.17, (.28)
[-.13, .44]
*Significant correlation at p<.05.
Response bias – c
Values of c were calculated as a measure of response bias for each task. Mean c values for all
groups are shown in Table 9.3. One-way ANOVAs revealed no significant group differences for any
of the tasks, suggesting that differences in response bias cannot account for differences in task
performance between groups. Paired t-tests revealed significantly higher response bias for identity
discrimination compared emotion discrimination, t(44)=13.3, p<.001 (mean difference = .76), but no
difference in response bias between the global and local tasks, t(44)=.03, p=.98 (mean difference =
.01). In order to directly compare performance on the two discrimination tasks, d prime values were
calculated according to formulae by Macmillan and Creelman (1991). These values were used in all
subsequent analysis as they are believed to be less affected by differences in response bias than raw
accuracy rates (Macmillan & Creelman, 1991).
Table 9.3. Mean c scores for patients with schizophrenia and healthy controls on the five tasks.
Schizophrenia
spectrum
n = 17
Psychiatric
controls
n = 14
Healthy
controls
n = 14
M (SD) M (SD) M (SD) One-way ANOVA
c Identity Discrimination -.71 (.40) -.85 (.40) -.77(.40) F(2,42)=.49, p =.62
Emotion Discrimination .03 (.18) -.02 (.11) -.05 (.20) F(2,42)=1.05, p =.36
Global attention .19 (.28) .05 (.30) .14 (.24) F(2,42)=.97, p =.39
Local attention .16 (.18) .14 (.23) .09 (.16) F(2,42)=.43, p =.65
106
Face processing tasks
To compare d’ performance across groups on the two face processing tasks, a 2 x 3 mixed
ANOVA was conducted with task as a within-subjects factor (identity vs emotion) and group as a
between-subjects factor (see Figure 9.3). Significant main effects were found for task, F(1,42) =
83.53, p<.001, ŋp2=.67 , and group, F(2,42) = 9.56, p<.001, ŋp
2=.31, while a task x group interaction
trended towards significance, F(2,42) = 3.02, p=.059, ŋp2=.13. Bonferroni-corrected pairwise
comparisons showed that performance was significantly higher for identity discrimination than
emotion discrimination overall (p<.001, mean difference=1.02).
Post-hoc one-way ANOVAs revealed that groups differed significantly on emotion
discrimination, F(2,42) = 18.06, p<.001, but not identity discrimination, F(2,42) = 1.58, p=.22.
Bonferroni-corrected multiple comparisons showed that healthy controls were significantly more
accurate than schizophrenia patients (p<.001, mean difference = 1.1), or psychiatric controls (p=.001,
mean difference = .73), who did not differ from one another (p=.17, mean difference= .35). No other
comparisons approached significance.
Figure 9.3. Performance (d’) of healthy controls, psychiatric controls and schizophrenia spectrum
patients for the identity discrimination and emotion discrimination tasks. Error bars indicate 95%
confidence intervals. *p<.01, **p<.001.
107
Navon task (global vs local attention)
Performance (d’) on the two attentional tasks was compared across groups using a 2 x 3
mixed ANOVA with task as a within-subjects factor and group as a between-subjects factor (see
Figure 9.4). Significant main effects were found for task, F(1,42) = 5.64, p=.02, ŋp2=.12, and group,
F(2,42) = 6.39, p=.004, ŋp2=.23. The interaction did not approach significance (p=.12). Overall,
performance was significantly higher for the local task compared to the global task (mean
difference=.32). Post-hoc pairwise comparisons revealed that, across tasks, healthy controls
outperformed both schizophrenia patients (p=.01, mean difference=1.0) and psychiatric controls
(p=.008, mean difference=1.1), who did not differ from one another (p>.99).
To determine whether patients showed a specific global processing deficit, performance for
global and local tasks was compared within each group using paired-samples t-tests. Performance did
not differ significantly between tasks in the healthy control group, suggesting that this group had no
bias to either the global or local level, t(13)=.35, p=.74. Patients with schizophrenia showed
significantly reduced performance on the global task compared to the local task, indicating a global
processing deficit, t(16)=2.56, p=.02. Like healthy controls, psychiatric controls showed no
significant difference between tasks, t(13)=1.46, p=.17, however it is unclear if this was due to the
smaller sample size for this group.
Figure 9.4. Performance (d’) of healthy controls, psychiatric controls and schizophrenia spectrum
patients for the global and local attentional tasks. Error bars indicate 95% confidence intervals.
*p<.01, n.s. = not significant.
108
Figure 9.5. Mean local processing bias for healthy controls, psychiatric controls and schizophrenia
spectrum patients. Bias was calculated by subtracting individual performance (d’) on the global task
from performance on the local task. Error bars indicate 95% confidence intervals.
An alternative method to evaluate a global processing deficit is to calculate individual bias
scores. These scores were calculated for each participant by subtracting performance (d’) on the
global task from performance on the local task. Scores are shown in Figure 9.5. A positive score
indicates a local processing bias, while a negative score indicates a global processing bias. One-
sample t-tests revealed that while healthy controls, t(13)=.35, p=.74, and psychiatric controls,
t(13)=1.46, p=.17, showed no significant bias one way or the other, the schizophrenia group showed a
significant local processing bias, t(16)=2.56, p=.02. However, a one-way ANOVA showed no
significant effect of group on bias scores, F(2,42) = 2.26, p=.12. Taken together, there is some
evidence to suggest that the schizophrenia patients tended to perform better on the local task
compared to the global task. However, the group samples may be too small to detect a group level
difference.
109
Figure 9.6. Reaction times (ms) of healthy controls, psychiatric controls and schizophrenia spectrum
patients for the global and local attentional tasks. Error bars indicate 95% confidence intervals.
Unlike the two face tasks, which have unlimited responding time, the go/no go design of the
Navon tasks requires rapid responses. Therefore, reaction time data was also analysed to determine
whether the results above were affected by a speed-accuracy trade-off. Reaction times were examined
using a mixed model ANOVA with task as a within-subjects factor (global vs local) and group as a
between-subjects factor. Group means are shown in Figure 9.6. Significant main effects were found
for task, F(1,42) = 6.06, p=.02, ŋp2=.13, and group, F(2,42) = 5.72, p=.006, ŋp
2=.21. The interaction
did not approach significance (p=.47). Results mirrored those of the accuracy data, with significantly
faster performance on the local task compared to the global task (mean difference = 124.7 ms). Post-
hoc pairwise comparisons showed that healthy controls were significantly faster than both psychiatric
controls (p=.007, mean difference = 301.0 ms), and schizophrenia patients (p=.047, mean difference =
223.8 ms), who were not different from one another (p>.99). To examine the possibility of a specific
global processing deficit, paired-samples t-tests were conducted for each group separately. No
significant differences in reaction time for the global or local task were found for either healthy
controls, t(13)=1.58, p=.14, or psychiatric controls, t(13)=1.53, p=.15, although a non-significant
trend was found for schizophrenia patients, t(16)=1.78, p=.09. Overall, the results of the reaction time
analyses closely followed the same patterns as the accuracy analyses, suggesting that groups that were
less accurate were also slower to respond. Therefore, there is no evidence to suggest that the results
reported above were distorted by a speed-accuracy trade-off.
110
Correlations between tasks
To determine whether performance on the two face tasks could be accounted for by global or
local performance, a series of Bonferroni-corrected Pearson correlations were conducted (Table 9.4).
Significant positive correlations were found between all tasks (p values=.02-.006) with the exception
that identity discrimination performance was not significantly associated with emotion discrimination
performance (p=.06).
Given that the demographic variables age, years of education, and estimated IQ were all
correlated with task performance, a second analysis was conducted to control for these variables (see
Table 9.5). A series of partial Pearson correlations revealed that, when age, education and IQ were
taken into account, identity discrimination performance was positively correlated with both global
(r=.34, p=.03) and local (r=.53, p<.001) task performance. In contrast, emotion discrimination was
positively correlated with local performance only (r=.35, p=.03). Performance on the global and local
tasks was also positively associated (r=.72, p<.001). Together, these results suggest that reduced
performance on the local task affected performance on both face tasks, whereas reduced performance
on the global task affected identity discrimination only, and not emotion recognition.
Pearson correlations between the face tasks and individual local bias scores were also
calculated. However, neither identity discrimination (r=.14, p=.25) nor emotion discrimination (r=-
.14, p=.37) performance was significantly associated with degree of local bias.
Table 9.4. Bonferroni-corrected Pearson correlations for performance across different tasks.
Identity
Discrimination
Emotion
Discrimination Global task Local task
d’ r (p)
[95% CI]
r (p)
[95% CI]
r (p)
[95% CI]
r (p)
[95% CI]
Identity Discrimination - .37, (.06)
[.09, .60]
.39, (.02)*
[.11, 61.]
.56, (.006)**
[.31, .73]
Emotion Discrimination - .48, (.006)**
[.22, .68]
.42, (.02)*
[.14, .63]
Global task - .68, (.006)**
[.48, .81]
Local task -
*Significant correlation at p<.05., ** p<.01.
111
Table 9.5. Uncorrected partial correlations for performance across different tasks, controlling for
Age, years of Education, and IQ.
Identity
Discrimination
Emotion
Discrimination Global task Local task
d’ r (p)
[95% CI]
r (p)
[95% CI]
r (p)
[95% CI]
r (p)
[95% CI]
Identity Discrimination - .26, (.11)
[-.06, .53]
.34, (.03)*
[.03, .59]
.53, (>.001)**
[.26, .72]
Emotion Discrimination - .28, (.08)
[-.04, .55]
.35, (.03)*
[.04, .60]
Global task - .72, (<.001)**
[.52, .84]
Local task -
*Significant correlation at p<.05., ** p<.001.
Discussion
The goal of the present study was to examine whether deficits in global visual attention are
associated with impairments in facial identity or expression recognition in schizophrenia. Consistent
with our predictions, the schizophrenia group, but not healthy controls or psychiatric controls, showed
evidence of a selective deficit in global processing. However, our predictions regarding associations
between facial processing and global/local attentional ability were only partially supported. As
predicted, global processing ability was associated with identity discrimination but not emotion
discrimination. However, local processing ability was associated with both emotion and identity
discrimination.
Identity and emotion recognition in schizophrenia
Consistent with the results of the main inpatient study (Chapter 8), patients with
schizophrenia were significantly impaired on a dynamic emotion discrimination task compared to
healthy controls. This result is also consistent with previous studies using similar paradigms
(Addington et al., 2006; Hooker & Park, 2002; Martin et al., 2005; Penn et al., 2000; Sachs et al.,
2004; Weniger et al., 2004). Performance in psychiatric patients without schizophrenia also showed
reduced performance, suggesting that this deficit is not specific to schizophrenia. As the psychiatric
112
control group comprised a wide range of diagnoses, however, it is not possible to comment on the role
of emotion processing in specific mental illnesses. Moreover, the possibility that psychotropic
medication may be playing a role in this deficit in patients cannot be ruled out.
However, while the previous study found a significant difference in identity discrimination
for patients compared to healthy controls, no difference in performance was reported in the current
study. Given that mean scores are similar between studies, this likely reflects the smaller sample size
in the current study, resulting in a lack of power to detect more subtle group differences. It is possible
that any true impairments in facial identity processing are simply less pronounced compared to
emotion processing deficits in schizophrenia, and therefore require comparatively larger samples to
detect. This could possibly explain the considerable variation in studies that report (Butler et al., 2008;
Martin et al., 2005; Shin et al., 2008; Soria Bauser et al., 2012) or fail to report (Edwards et al., 2001;
Johnston et al., 2010; Soria Bauser et al., 2012), identity processing impairments in schizophrenia.
Navon tasks (global vs local attention)
In line with previous data in healthy individuals (McKone et al., 2010), the current study
found that healthy controls performed equally well in the global and local tasks, in both accuracy and
reaction time data. This indicates that the tasks employed ‘equal salience’ stimuli, meaning that the
hierarchical stimuli do not promote a bias to attend to either the global or local level (Yovel, Levy, &
Yovel, 2001). Confirmation of this property is important, as an inherent bias to attend globally or
locally may obfuscate individual differences in performance (Aimola Davies, personal
communication, 2010).
It was also found that although patients with schizophrenia performed less accurately than
healthy controls on both the global and local tasks, they showed a relative deficit on the global task.
This finding replicates the global deficit shown in previous similar studies (Goodarzi, Wykes, &
Hemsley, 2000; Johnson, Lowery, Kohler, & Turetsky, 2005; Poirel et al., 2010; Silverstein, Kovacs,
Corry, & Valone, 2000), and provides support for the idea that schizophrenia is associated with a
specific impairment in processing configural or global-level information from a visual stimulus.
Interestingly, while the psychiatric control group showed reduced performance on both tasks, similar
to the schizophrenia group, they did not show evidence of a relative global deficit. Were this finding
robust, it would suggest that the relative global processing deficit may in fact be specific to
schizophrenia. However, the psychiatric control group did show a non-significant trend towards lower
performance in the global task compared to local which means, given the smaller number of patients
in this group (n=14), that the study may have lacked sufficient power to show a significant result.
Furthermore, it is likely that the presence of mood symptoms would have affected performance, as
research suggests that negative affective states temporarily bias participants to preferentially attend to
local details, rather than global (Basso, Schefft, Ris, & Dember, 1996; Gasper & Clore, 2002) . Future
113
research is needed to investigate the role of patient symptomatology in global-local attentional
processing.
Is a global-local attentional bias associated with face processing performance?
As predicted, correlational analyses revealed that poorer global processing was associated
with reduced performance on an identity discrimination task, but not an emotion task. This suggests
that the ability to attend to global information affects our ability to discriminate between faces of
different identities, but is not necessary for distinguishing between emotions. This finding agrees with
previous studies that indicate enhanced non-emotional face processing when attention is directed
globally, and poorer performance when attention is directed locally (Macrae & Lewis, 2002; Perfect,
2003), as well as studies that suggest that emotion processing is less reliant on global-level cues
(Bimler et al., 2013; Calder et al., 2000; Martin et al., 2012). Furthermore, this result suggests that the
global processing deficit seen in schizophrenia may partially account for the deficits in facial identity
recognition shown in previous studies (Hooker & Park, 2002; Martin et al., 2005). However, there is
no indication that a global processing deficit could account for the deficits in facial emotion
recognition that are reliably associated with schizophrenia.
In contrast, performance on the local processing task correlated positively with both the
identity discrimination task and the emotion discrimination task. This finding suggests that the ability
to attend selectively to details (such as individual facial features) while ignoring overall configural
details is important for distinguishing both identity and emotion. While previous studies have
highlighted the importance of using feature-based cues in emotion recognition (e.g.: the eyes, Calder
et al., 2000; Martin et al., 2012), the association between local processing and identity discrimination
was unexpected. It is possible that the same-or-different paradigm itself favours a strategy that
involves comparing individual details, such as eyebrow shape, in order to decide if the two videos are
subtly different.
Although patients with schizophrenia were outperformed by healthy controls in the local task,
they still showed a bias to attend locally overall (i.e.: a global deficit). Therefore, it appears that the
profound deficits in emotion discrimination in this group cannot be (fully) explained by a poorer
ability to attend to local details.
Study limitations
The results of this study must be considered in the context of several limitations. First, the
sample sizes for the three groups were small and uneven, which limited the power to detect smaller
group differences and within-group effects. Furthermore, demographic factors such as age and years
of education were not matched between groups. Age, in particular, has an impact on cognitive factors
114
such as processing speed (Salthouse, 2000), which may differentially affect performance on tasks with
limited presentation times or limited response windows. Finally, due to the small patient sample, all
inpatients without a diagnosis of schizophrenia were categorised into a single psychiatric control
group. The heterogeneity of this group, which included both psychotic (e.g.: bipolar disorder with
delusions) and non-psychotic patients (e.g.: depression), made it impossible to draw any conclusions
about specific diagnoses.
Conclusions
In line with previous research, the schizophrenia group showed a relative deficit in global
attention compared to local. Both the schizophrenia and psychiatric control groups showed reduced
performance compared to healthy controls on an emotion discrimination task, but not an identity
discrimination task, suggesting an emotion deficit that is not specific to schizophrenia. Correlational
analyses suggest that the ability to attend to global information is an important skill for distinguishing
facial identity, but cannot account for deficits in emotion processing. While the ability to attend to
local details correlates with performance for both identity processing and emotion processing, this
deficit cannot fully explain the profound impairment in emotion discrimination shown by the patient
groups, either. Taken together, these findings indicate that impairments in the allocation of global or
local attention may be adequate to explain deficits in non-emotional facial processing in
schizophrenia. However, deficits in emotion processing may be partly, but not fully, explained by a
reduced ability to attend to local details. Overall, this study suggests that the emotion-processing
deficits seen in schizophrenia cannot be accounted for by impairments in allocating visuospatial
attention.
115
Chapter 10: Are these deficits specific to schizophrenia? Overlap
with Face Processing Deficits in Other Psychiatric Disorders
Deficits in emotion processing have been well established in schizophrenia and may even
constitute a core feature of this disorder (Kohler et al., 2010). However, whether these deficits are
diagnostically specific to schizophrenia remains a point of contention. This chapter will address the
second aim of this thesis: To explore whether face processing impairments are specific to
schizophrenia, or whether they are associated with psychosis more generally. This chapter will begin
by reviewing the literature examining face processing deficits in non-schizophrenia psychosis and
non-psychotic disorders. It will then present further results of the inpatient study, which compares
facial emotion and identity processing in schizophrenia with inpatients with bipolar disorder, other
causes of psychosis (such as drug-induced psychosis), and non-psychotic disorders. Finally, these
results will be discussed in terms of theories suggesting a shared aetiology between schizophrenia and
other psychotic disorders.
The current study
The purpose of the Results section in this chapter is to compare face processing in patients
with schizophrenia with other psychiatric inpatient groups, including bipolar I disorder. Unlike
previous studies, the current study included two additional groups: a non-schizophrenia ‘other’
psychosis group (predominantly drug-induced psychosis and psychotic depression), and a non-
psychosis psychiatric control group (predominantly major depressive disorder). Although under-
studied, disorders such as drug-induced psychosis are increasingly viewed as the less extreme end of a
continuum leading to schizophrenia. Drug-induced psychosis is thought to involve similar
neurochemical pathways to schizophrenia (Paparelli, Di Forti, Morrison, & Murray, 2011), with an
estimated 46% of patients with cannabis-induced psychosis and 20-30% of patients with
amphetamine-induced psychosis converting to schizophrenia within 8 years (Niemi-Pynttari et al.,
2013; Starzer, Nordentoft, & Hjorthoj, 2018). Despite this, no studies could be found which
specifically examined emotion processing in non-schizophrenia psychosis. Therefore, the first
research question is: is emotion processing disproportionately impaired in schizophrenia, or are these
deficits seen in other disorders such as a) bipolar I disorder, b) non-schizophrenia psychosis, or c)
non-psychotic disorders? Second, presuming these deficits are found, are these indeed specific to
emotion, or do they generalise to a) non-emotional face processing or b) non-face visual
discrimination?
116
Face processing deficits in bipolar disorder
Compared to the extensive schizophrenia literature, far fewer studies have examined face
processing abilities in other disorders, such as bipolar disorder. Nevertheless, there is substantial
evidence pointing to impaired face processing in bipolar disorder, particularly for recognising
emotion, although the degree and persistence of these deficits remain contentious (Van Rheenen &
Rossell, 2013b). Patients with bipolar disorder have shown impaired performance compared to
healthy controls in correctly labelling facial emotions (Getz et al., 2003; Lembke & Ketter, 2002; Van
Rheenen & Rossell, 2014) as well as discriminating between different expressions (Addington &
Addington, 1998; Bozikas, Tonia, et al., 2006; Rossell et al., 2013). Other studies using morphed
static faces have shown that patients with bipolar disorder require greater intensity to recognise
emotions compared to controls (Brotman et al., 2008; Schaefer, Baumann, Rich, Luckenbaugh, &
Zarate, 2010). This emotion processing deficit has not always been found consistently, however, with
several studies reporting intact emotion recognition in bipolar disorder (Edwards et al., 2001; Harmer,
Grayson, & Goodwin, 2002; Vaskinn et al., 2007) or in certain patient subsets, such as euthymic
patients (Lembke & Ketter, 2002).
Comparisons between diagnostic subtypes of bipolar disorder have also produced variable
results. While Summers and colleagues (2006) found significant impairments in labelling dynamic
emotions for both bipolar I and bipolar II patients, a different study using static face stimuli reported
impairments in bipolar I, but intact performance in bipolar II patients (Derntl, Seidel, Kryspin-Exner,
Hasmann, & Dobmeier, 2009). Finally, Lembke and Ketter (2002) found impaired emotion labelling
ability in manic patients with bipolar I, but intact performance in euthymic patients with bipolar I and
II.
One possibility is that emotion processing deficits in bipolar disorder represent a state-
dependent dysfunction. Evidence for this view comes from studies indicating greater impairments in
symptomatic patients compared to remitted or euthymic patients. For instance, poorer performance
has been reported in currently manic patients (Lembke & Ketter, 2002), depressed patients
(Langenecker, Saunders, Kade, Ransom, & McInnis, 2010; Schaefer et al., 2010), and those
experiencing mixed episodes (Gray et al., 2006). However, substantial impairments have also been
reported in remitted patients (Bozikas, Kosmidis, et al., 2006; Vederman et al., 2012), with one study
reporting no difference between remitted and symptomatic patients (Van Rheenen & Rossell, 2013a).
Moreover, emotion recognition deficits have also been reported in individuals at risk of developing
bipolar disorder (Brotman et al., 2008; Rock, Goodwin, & Harmer, 2010). Together, these results
suggest that emotion processing deficits in bipolar disorder may represent a trait marker of the illness
which is likely exacerbated by the resurgence of acute symptoms.
117
In addition to behavioural deficits, bipolar disorder has been associated with abnormal
patterns of neural activation. Functional imaging studies have generally shown reduced activation in
regions of the orbitofrontal cortex and ventrolateral prefrontal cortex, accompanied by increased
activation in the amygdala, parahippocampal gyrus, and medial prefrontal gyrus during emotion
processing (Brotman et al., 2014; Delvecchio, Sugranyes, & Frangou, 2013; Keener et al., 2012;
Surguladze et al., 2010; Vizueta et al., 2012). Abnormal functional connectivity has also been
reported in fronto-temporal circuits during an emotion recognition task in bipolar disorder (Versace et
al., 2010; Vizueta et al., 2012). These differences in functional activation broadly point towards
dysfunction in systems typically implicated in social cognition (Kret & Ploeger, 2015).
Several studies have also sought to disentangle trait abnormalities from state-dependent
abnormalities. For instance, Hulvershorn and colleagues (2012) reported that bipolar patients in
depressed mood states showed increased amygdala activity, whereas patients in manic mood states
showed increased insula activity and decreased activity of the left lateral orbitofrontal cortex. In
contrast, Liu and colleagues (2012) found an increase in left orbitofrontal activity in depressed
patients and reduced right anterior prefrontal activity in manic patients, whereas functioning of the
right anterior cingulate cortex, orbitofrontal cortex, and ventral striatum were all found to be abnormal
regardless of mood state. In a functional connectivity study, Versace and colleagues (2010) suggested
that abnormal connectivity between the left amygdala and orbitofrontal cortex may indicate a state
marker of depression. Overall, it is clear that further investigation is required to elucidate the neural
underpinnings of these deficits.
Face processing deficits in major depressive disorder
As for bipolar disorder, a range of studies indicate some degree of impairment in recognising
facial emotion in major depression (MDD), although the extent of these deficits remains unclear.
Several reviews suggest that, on balance, there are mild but significant impairments in emotion
processing in MDD (Bourke, Douglas, & Porter, 2010; Kohler, Hoffman, Eastman, Healey, &
Moberg, 2011; Leppanen, 2006). In particular, MDD patients tend to show a bias towards perceiving
neutral or ambiguous expressions as sadness, and show a reduced ability to recognise all basic
emotions except for sadness (Dalili, Penton-Voak, Harmer, & Munafo, 2015). These deficits have
been demonstrated using emotion labelling tasks (Douglas & Porter, 2010; Joormann & Gotlib, 2006;
Langenecker et al., 2005; Leppanen, Milders, Bell, Terriere, & Hietanen, 2004; Mah & Pollock, 2010;
Surguladze et al., 2004) as well as emotion discrimination or matching tasks (Asthana, Mandal,
Khurana, & Haque-Nizamie, 1998; Gur et al., 1992; Levkovitz, Lamy, Ternochiano, Treves, &
Fennig, 2003; Rubinow & Post, 1992). In contrast, a number of studies have reported no difference in
emotion processing ability between patients with MDD and healthy controls (Anderson et al., 2011;
118
Bediou, Krolak-Salmon, et al., 2005; Derntl, Seidel, Schneider, & Habel, 2012; Edwards et al., 2001;
Gaebel & Wolwer, 1992; Gollan, Pane, McCloskey, & Coccaro, 2008; Kan, Mimura, Kamijima, &
Kawamura, 2004; Schaefer et al., 2010). This discrepancy is possibly due to a lack of power in these
studies, as a recent meta-analysis suggests that the overall effect size for these deficits in MDD is
quite small (Hedges g= -.16; Dalili et al., 2015).
Just as for bipolar disorder, these emotion recognition deficits appear to have both state and
trait components. For instance, several studies have shown abnormal emotion processing in MDD
patients in remission (Bouhuys, Geerts, Mersch, & Jenner, 1996; Leppanen et al., 2004; Suslow et al.,
2004), and even in healthy individuals who are genetically at-risk of depression (Le Masurier, Cowen,
& Harmer, 2007). However, other studies have shown significant improvements in expression
recognition following the improvement of symptoms (Bouhuys, Geerts, & Gordijn, 1999; Hale, 1998;
Mikhailova, Vladimirova, Iznak, Tsusulkovskaya, & Sushko, 1996), suggesting that this deficit is
exacerbated by acute symptoms of depression.
Given that bipolar disorder and MDD are both affective disorders with some overlap in both
genetics and symptom expression (Huang, Hsiao, Hwu, & Howng, 2013; Mitchell et al., 2011), could
these emotion processing deficits indicate a shared impairment in these disorders? A meta-analysis by
Kohler and colleagues (2011) found no difference between bipolar disorder and MDD in the extent of
emotion processing deficits on behavioural tasks. However, two individual studies have highlighted
specific differences between these disorders. Schaefer and colleagues (2010) found reduced accuracy
for labelling emotions in both bipolar disorder and MDD, but only the bipolar group required greater
intensity of an expression to identify it correctly. Another study reported that patients with bipolar
disorder were poorer at labelling fear compared to MDD patients, but not other emotions (Vederman
et al., 2012). Neural activations to emotional stimuli in MDD may be different to those seen in bipolar
disorder, with meta-analyses pointing to increased activation of the ventro-rostral anterior cingulate
cortex (ACC), decreased dorsal ACC activation, and increased amygdala activity (Jaworska, Yang,
Knott, & MacQueen, 2015). A meta-analysis by Delvecchio and colleagues (2012) compared fMRI
activations during emotion processing in MDD and bipolar disorder patients. They found that both
disorders showed increased activation of limbic regions, including the amygdala and parahippocampal
gyrus, possibly indicating heightened salience of emotional stimuli in both of these disorders.
However, while bipolar disorder was associated with decreased ventrolateral PFC activity and
increased activity of areas of the thalamus and basal ganglia, MDD was associated with reduced
activation of the sensorimotor cortices. These differences may represent reduced inhibitory control
and higher facial mimicry in bipolar disorder, but reduced emotional reactivity in MDD. Whether
these abnormalities underlie the dysfunction shown on behavioural measures or simply reflect
different patterns of symptoms (such as reduced facial affect in depression versus lability in mania),
however, requires further investigation.
119
One study used dynamic causal modelling to investigate differences in MDD and depressed
bipolar patients during an emotion labelling task (Almeida et al., 2009). They found distinct patterns
of connectivity between the amygdala and orbitomedial prefrontal cortex during viewing of happy
faces that distinguished bipolar disorder and MDD. While MDD patients showed increased negative
left-sided top-down connectivity, bipolar disorder patients showed reduced positive left-sided top-
down connectivity and increased negative right-sided bottom-up connectivity. These findings suggest
that, even when bipolar and MDD patients are experiencing similar acute symptoms, they show
different areas of pathophysiology during emotion processing.
Face processing deficits in other non-psychotic disorders
As part of a growing interest in social cognition in psychiatric disorders, a variety of studies
have examined facial emotion processing in disorders such as anxiety and borderline personality
disorder. One meta-analysis examined emotion processing impairments in different anxiety disorders
across 40 studies (Plana, Lavoie, Battaglia, & Achim, 2014). Interestingly, a large weighted mean
effect size was found for post-traumatic stress disorder (Cohen’s d=-1.60), indicating substantial
deficits in emotion recognition. However, only small or negligible effects were found for social
phobia (d=.12), obsessive-compulsive disorder (d=-.16), panic disorder (d=-.25) and generalised
anxiety disorder (d=-.12). However, whether these deficits indicate specific impairments in
recognising emotion, or simply reflect more general difficulties with attentional control that typically
accompany anxiety disorders (Eysenck & Derakshan, 2011), remains to be established. Predictably,
functional imaging studies have showed similar patterns of neural activation in anxiety disorders to
those seen in bipolar disorder and MDD (Kret & Ploeger, 2015). Meta-analyses point towards
increased activity of the amygdala and reduced activity of the ventromedial PFC and thalamus during
emotion processing in patients with anxiety disorders (Etkin & Wager, 2007; Hattingh et al., 2012;
Hayes, Hayes, & Mikedis, 2012) as well as in anxiety-prone healthy individuals (Kret, Denollet,
Grezes, & de Gelder, 2011).
Another disorder of interest is borderline personality disorder (PD). Despite being categorised
as a personality disorder rather than an affective disorder, borderline PD is characterised by powerful
mood swings and emotional instability and has considerable clinical overlap with both bipolar II
disorder and the atypical presentation of major depressive disorder (Perugi, Fornaro, & Akiskal,
2011). Some authors suggest that these three disorders may share a similar ‘cyclothymic’ diathesis
which is supported by substantial familial clustering (Zanarini, Barison, Frankenburg, Reich, &
Hudson, 2009). A review by Kret and Ploeger (2015) suggested that borderline PD is associated with
subtle but significant impairments in recognising emotions, with a particular bias towards
misidentifying neutral emotions as negative. To date, only two meta-analyses have been conducted.
Daros and colleagues (2013) found that patients with borderline PD were less able to recognise
120
disgust, fear, and neutral expressions compared to healthy controls. In contrast, Mitchell and
colleagues (2014) found no significant evidence of recognition deficits for negative emotions
compared to healthy controls, but instead found a significant bias to label neutral or ambiguous faces
as negative, consistent with an enhanced sensitivity to threatening stimuli. Functional imaging studies
have produced similarly inconsistent results. Ruocco and colleagues (2013) reported increased
activation of the insula and posterior cingulate cortex and decreased activation of the amygdala and
parts of the anterior cingulate and dorsolateral prefrontal cortex in borderline PD during an emotion
recognition task. In contrast, the meta-analysis by Mitchell and colleagues (2014) reported increased
activation of the amygdala while viewing emotional expressions, a finding consistent with the
amygdala hyperactivation reported in depression (Peluso et al., 2009), bipolar disorder (Surguladze et
al., 2010) and anxiety disorders (Monk et al., 2008). Additionally, they reported altered activation of
areas of the anterior cingulate cortex, inferior frontal gyrus, and superior temporal sulcus, with an
inconsistent mix of hyperactivation and hypoactivation found across individual studies.
Overall, studies of non-psychotic disorders such as anxiety disorders (with the exception of
PTSD) and borderline PD show minor but often significant deficits in recognising facial emotions
compared to healthy controls. These subtle and often inconsistent deficits appear to be similar to those
reported in depression and in bipolar II disorder (without manic episodes). Imaging studies point to
similar areas of altered functioning of the amygdala and areas of the prefrontal cortex, among others,
however it is likely that these differences relate to an attributional bias towards threat sensitivity rather
than dysfunction of areas involved in emotion recognition specifically.
Results: Face processing across psychiatric disorders
Note that the Method for this study can be found in Chapter 7.
Group differences in task performance
Repeated –measures ANOVAs were conducted on raw accuracy data across morphing levels.
As all groups showed the same pattern of performance, data for each task was collapsed across
morphing levels for subsequent analyses.
Graphs showing d’ scores for the five tasks are shown in Figure 10.1. A 5 x 5 MANOVA
revealed a significant main effect for group, F(4,99) = 2.93, p<.001 (Pillai’s Trace). Univariate tests
revealed significant effects of group for all tasks except Sex Labelling. Mean d’ scores for each group
on the five tasks are shown in Table 10.
Emotion Discrimination (Figure 10.1A) also showed a univariate effect of group (F(4,99) =
14.18, p<.001). The Schizophrenia spectrum group performed significantly more poorly compared to
all other groups (p values=.02-<.001) except Bipolar disorder. The healthy control group also
121
significantly outperformed the Other psychosis (p=.03) and Bipolar disorder (p<.001) groups. The
Non-psychosis group trended towards significantly outperforming the Bipolar group (p=.07).
A similar pattern of results were found for Emotion Labelling (Figure 10.1B; Univariate
effect of group: F(4,99) = 9.19, p<.001). The healthy control group outperformed the Schizophrenia
spectrum (p<.001), Bipolar (p<.001) and Other psychosis groups (p=.02). The Non-psychosis group
also outperformed the Schizophrenia spectrum (p=.02) and Bipolar (p=.03) groups.
Identity Discrimination (Figure 10.1C) had a main effect of group, (F(4,99) = 2.47, p=.049).
There was a trend for the Schizophrenia spectrum group to perform more poorly than the Control
group (p=.09), however Bonferroni-corrected pairwise comparisons revealed no significant
differences between any groups (p values=.09-.99).
No effect of group was found for Sex Labelling, (F(4,99) = .363, p=.84). An effect of task
difficulty was seen with lowest performance at 40% male faces, however overall performance was
consistent across all groups (see Figure 10.1C).
Unexpectedly, a significant univariate effect of group was also found for Non-face
Discrimination (Figure 10.1E; F(4,99) = 4.31, p=.003). Bonferroni-corrected pairwise comparisons
revealed that this was driven by the Control group significantly outperforming the Schizophrenia
spectrum (p=.01) and Bipolar groups (p=.03). No other comparisons approached significance.
Table 10. Mean d’ scores for each group on the five tasks.
Schizophrenia
n = 36
Bipolar
disorder
n = 15
Other psychotic
disorders
n = 17
Non-psychotic
disorders
n = 18
Healthy
controls
n = 20
M (SD) M (SD) M (SD) M (SD) M (SD)
d’ Emotion
Discrimination .99 (.57)a 1.28 (.52)a 1.55 (.58)a,b 1.86 (.50)b 2.16 (.79)b
Emotion Labelling 1.14 (.73)a 1.00 (.83)a 1.51 (1.19)a 1.96 (1.02)b,c 2.44 (.73)b
Identity
Discrimination 2.57 (.54) 2.67 (.54) 2.59 (.51) 2.87 (.49) 2.95 (.40)
Sex Labelling 1.90 (.65) 2.12 (.60) 1.93 (.71) 2.00 (.36) 2.03 (.70)
Car Discrimination 2.11 (.64)a 2.04 (.79)a 2.45 (.61) 2.61 (.73) 2.73 (.62)b
aSignificantly different from healthy controls, p<.05; bSignificantly different from schizophrenia group, p<.05; cSignificantly different from bipolar disorder group, p<.05. (All bonferroni-corrected post-hoc comparisons).
122
Figure 10.1. Mean d’ performance for the schizophrenia spectrum (SZ), bipolar disorder (BPAD),
other psychosis (Other), non-psychosis (NP) and healthy control groups across the five tasks: Emotion
Discrimination (A), Emotion Labelling (B), Identity Discrimination (C), Sex Labelling (D), and Car
Discrimination (E). Significant differences between groups are indicated with dotted lines, **p<.01;
*p<.05. Error bars represent 95% confidence intervals.
123
Task performance in the bipolar disorder group
Mean performance across tasks for the bipolar group are shown in Figure 10.2. A repeated-
measures ANOVA was conducted with accuracy (d’) as the dependent variable and task as a within-
subjects factor. According to Mauchly’s test, the assumption of sphericity was violated therefore the
Greenhouse-Geisser correction was used, X2(9)=20.74, p=.02. A main effect of task was found,
F(2.18, 30.55)=26.02, p<.001, ŋp2=.65. Bonferroni-corrected post-hoc comparisons revealed that
performance on the two emotion tasks was not significantly different from one another (p>.99) but
were both significantly lower than the three remaining tasks (p values<.02, mean differences = .76 –
1.67). Performance on the identity discrimination task was significantly higher than all other tasks
except sex labelling (p values<.01, mean differences = .64 – 1.67). Finally, the car discrimination and
sex labelling tasks were not significantly different from one another (p>.99).
Figure 10.2. Performance (d’) of patients with bipolar disorder across the five dynamic tasks. Error
bars indicate 95% confidence intervals. Dots indicate the performance of individual participants.
*p<.01; **p<.001.
124
Task performance in the non-schizophrenia psychosis group
Mean performance across tasks for the other-psychosis group are shown in Figure 10.3. A
repeated-measures ANOVA was conducted with accuracy (d’) as the dependent variable and task as a
within-subjects factor. According to Mauchly’s test, the assumption of sphericity was violated
therefore the Greenhouse-Geisser correction was used, X2(9)=21.97, p=.01. A main effect of task was
found, F(2.18, 34.80)=9.45, p<.001, ŋp2=.37. Bonferroni-corrected post-hoc comparisons revealed that
performance on the identity discrimination task was significantly higher than the two emotion tasks
and the sex labelling task (p values <.03, mean differences = .66 – 1.08). Performance on the non-face
task was also significantly higher than the emotion discrimination task (p=.006, mean difference=.91).
No other comparisons approached significance.
Figure 10.3. Performance (d’) of patients with non-schizophrenia psychosis across the five dynamic
tasks. Error bars indicate 95% confidence intervals. Dots indicate the performance of individual
participants. *p<.05; **p<.001.
125
Task performance in the non-psychosis group
Mean performance across tasks for the non-psychosis group are shown in Figure 10.4. A
repeated-measures ANOVA was conducted with accuracy (d’) as the dependent variable and task as a
within-subjects factor. According to Mauchly’s test, the assumption of sphericity was not violated,
X2(9)=14.29, p=.11. A main effect of task was found, F(4, 68)=13.34, p<.001, ŋp2=.44. Bonferroni-
corrected post-hoc comparisons revealed that performance on the identity discrimination task was
significantly higher than the two emotion tasks and the sex labelling task (p values<.01, mean
differences = .87 – 1.01). Performance on the non-face task was also significantly higher than the
emotion discrimination task and the sex labelling task (p values<.02, mean difference=.61-.75). No
other comparisons approached significance.
Figure 10.4. Performance (d’) of patients with non-psychotic disorders across the five dynamic tasks.
Error bars indicate 95% confidence intervals. Dots indicate the performance of individual participants.
*p<.05, **p<.001.
Response bias – c
Mean values of c ranged from .02 to .88 across groups and tasks. Multivariate ANOVA
revealed no significant main effect of group. Pairwise comparisons showed no differences between
groups on any task, suggesting that response bias did not differ between groups, and are therefore
unlikely to account for differences in task performance in these groups.
126
Impact of morphing: Varying emotional intensity and facial identity
Emotion Discrimination. Repeated-measures ANOVA was conducted for raw accuracy across
the different intensities of expression. The main effect of intensity was significant, F(1.76, 183.44)=
33.13, p=.005. Contrary to expectations, accuracy was lowest for expressions at 67% intensity
compared to all other intensities (p values=.01 to .001), which did not differ significantly from one
another. This finding indicates that, for all groups, decreasing the intensity of a moving expression did
not significantly affect performance, with one exception. The finding that accuracy was lowest for
expressions at 67% intensity likely reflects the increase in visual artefact or “graininess” of the video
as a result of the morphing process at this intensity (Figure 10.5A).
Emotion Labelling. A Repeated measures ANOVA on raw accuracy revealed a significant
main effect for emotional intensity, F(3.46, 363.26)=17.48, p<.001, but not emotion (disgust vs fear).
Performance is shown in Figure 10.5B (disgust faces) and 9.5C (Fear faces). Unlike the Emotion
Discrimination task, accuracy for naming emotions increased somewhat with increasing intensity.
Performance at 33% intensity was significantly lower than all other levels (p values=.006 - .001), and
performance at 50% intensity was lower than 67% intensity (p=.04) and 83% intensity (p=.001). No
other differences in intensity approached significance.
Identity Discrimination. Raw accuracy across different levels of morphing are shown in
Figure 10.5D. Repeated-measures ANOVA revealed a main effect of morphing, F(5,525)= 599.34,
p<.001. Accuracy increased significantly with each level of increasing difference (p values = .04 -
.001) up to 80% difference, which did not differ significantly from 100% difference (p>.999). All
groups performed above chance when faces were 60% different, and at or below chance at 40%
different.
Sex Labelling. Repeated-measures ANOVA revealed significant main effects of sex,
F(1,104)= 33.13, p<.001 and morphing level, F(2,208)= 405.05, p<.001. Accuracy was lowest for the
60/40% morphed faces and highest for 100% (un-morphed) faces. Unexpectedly, accuracy was
reliably higher for identifying male faces than female faces (Figure 10.5E).
127
Figure 10.5. Mean accuracy performance for the schizophrenia spectrum, bipolar disorder, non-
schizophrenia psychosis (Other psychosis), non-psychotic disorders and control groups across the four
morphed face tasks: Emotion Discrimination (A), Emotion Labelling for disgust faces (B), Emotion
Labelling for fear faces (C), Identity Discrimination (D), and Sex Labelling (E). For A, B, and C,
morphing level is on the y axis, where 100% indicates an unedited expression and 33% indicates an
expression morphed 50% with a neutral expression. For D and E, morphing level is shown on the y
axis, where 50% indicates an equal morph between Face 1 and Face 2.
128
Discussion
The goal of this section was to investigate the extent of facial affect and identity processing
impairments (if any) using dynamic stimuli in patients with a variety of psychiatric disorders. The
purpose of this was to determine whether these deficits are specific to schizophrenia, or if there is
evidence of a shared mechanism of impairment underlying such deficits. The results revealed that
groups with bipolar disorder, schizophrenia, and ‘other’ (non-schizophrenia) psychotic disorders were
all significantly impaired on the facial emotion tasks compared to healthy controls, while patients with
non-psychotic disorders were unimpaired. In contrast, all patient groups showed relatively intact
performance on the two facial identity processing tasks. Unexpectedly, patients with schizophrenia
and bipolar disorder also showed deficits on a non-face comparison task.
Emotion-processing deficits in bipolar disorder
The results of the current experiment revealed significant facial emotion processing deficits in
patients with bipolar I disorder which were comparable to those seen in patients with schizophrenia.
Compared to healthy controls, these groups were impaired both in naming, and distinguishing
between different facial expressions of disgust and fear. These findings are consistent with previous
studies of patients with bipolar disorder using static stimuli, which typically report emotion
processing deficits regardless of task design (Addington & Addington, 1998; Bozikas, Tonia, et al.,
2006; Getz et al., 2003; Lembke & Ketter, 2002; Rossell et al., 2013; Van Rheenen & Rossell, 2014).
Unexpectedly, however, the current study showed no significant difference in performance between
the bipolar and schizophrenia groups. This finding is contrary to previous research indicating that
patients with bipolar disorder have less severe deficits than patients with schizophrenia (Addington &
Addington, 1998; Goghari & Sponheim, 2013; Kohler et al., 2011; Ruocco et al., 2014; Vaskinn et al.,
2007; Wynn et al., 2013), although several other studies have also produced a null result (Bellack,
Blanchard, & Mueser, 1996; Derntl et al., 2012; Edwards et al., 2001). This finding may reflect a lack
of power due to the small number (n=16) of bipolar disorder patients in this sample. However, it is
also possible that this sample was more impaired due to the exclusion of bipolar II disorder (non-
psychotic) patients from this group, a subtype that in some cases has been shown to be less impaired
in emotion processing compared to bipolar I disorder patients (Derntl et al., 2009; Lembke & Ketter,
2002).
Emotion-processing impairments in patients with and without psychotic features
At present, this is the first study to show that facial emotion processing is impaired in non-
schizophrenia psychoses. This finding supports the idea that non-schizophrenia forms of psychosis
produce similar deficits to those seen in schizophrenia, and likely involve similar neural mechanisms
(Paparelli et al., 2011). Moreover, the finding that emotion recognition was unimpaired in patients
129
with non-psychotic disorders is consistent with studies reporting broadly intact performance in major
depression (Anderson et al., 2011; Derntl et al., 2012; Gollan et al., 2008; Schaefer et al., 2010),
certain anxiety disorders (Plana et al., 2014), and borderline personality disorder (Mitchell, Dickens,
& Picchioni, 2014). Note that, although some meta-analyses suggest that these disorders are
associated with mild deficits in emotion recognition (particularly misinterpretation of threatening
faces) these are typically of a much smaller effect size than those seen in schizophrenia, and it is
possible that the current study lacked the sample size to detect these subtler deficits (Dalili et al.,
2015; Plana et al., 2014).
Identity-processing deficits and non-face processing performance: Are these deficits specific to
emotion?
The current study revealed no significant impairments in either discriminating between
different identities, or labelling the sex of a face, in any of the groups examined. Although the
analyses conducted in Chapter 8 revealed significantly reduced performance in the schizophrenia
group compared to healthy controls, this result was no longer significant when all groups were
analysed in a multivariate ANOVA. Nevertheless, the finding of intact identity processing is
consistent with previous studies in bipolar disorder (Bozikas, Tonia, et al., 2006; Getz et al., 2003;
Venn et al., 2004), major depression (Bediou, Krolak-Salmon, et al., 2005), and the majority of
studies of schizophrenia (Bediou et al., 2007; Bediou, Krolak-Salmon, et al., 2005; Chen et al., 2012).
An unexpected finding, however, was that the schizophrenia and bipolar I groups were both
significantly impaired on a non-face discrimination task compared to healthy controls. As mentioned
in Chapter 8, this finding suggests that any reduced performance on the identity discrimination task is
unlikely to represent true deficits in mental processes specific to face perception. Rather, this likely
indicates a more generalised deficit – perhaps in attention or working memory – that affects
performance regardless of perceptual category.
Taken together, these results suggest that although some groups showed mild deficits on a
non-face discrimination task, the substantial emotion processing deficits shown in bipolar I disorder,
schizophrenia, and non-schizophrenia psychosis cannot be fully explained by broader deficits in face
processing. However, they may be explained by a more general cognitive impairment. It is unlikely,
therefore, that these disorders have specific impairments in recognising and distinguishing between
dynamic emotions of disgust and fear. This finding is congruent with the idea that emotion processing
deficits are underlain by disturbance to specific frontotemporal networks, which are affected in
psychiatric disorders such as schizophrenia (Hall et al., 2008; Mukherjee et al., 2012), bipolar disorder
(Versace et al., 2010; Vizueta et al., 2012), as well as neurodegenerative disorders involving these
frontotemporal structures (Bediou et al., 2012).
130
Conclusions
The aim of this section was to determine whether emotion processing deficits are
disproportionately impaired in schizophrenia compared to other psychiatric disorders. The results
suggest that emotion processing is substantially impaired in disorders with psychotic features –
including schizophrenia spectrum disorders, bipolar I disorder, and non-schizophrenia psychosis – but
not in patients without psychotic symptoms. The finding that emotion-processing impairments
accompany psychosis rather than specific diagnostic categories provides some support for the idea
that the deficits seen in schizophrenia may fall on a continuum with other forms of psychosis.
Associations between specific psychotic symptoms and face processing deficits will be explored
further in Chapter 11.
131
Chapter 11: What else accompanies these deficits? Symptom
Correlates of Face-processing Deficits in Schizophrenia and
Other Disorders
In terms of clinical presentation, schizophrenia is a highly heterogeneous disorder. Despite
sharing the same diagnosis, any two patients may present with vastly different constellations of
symptoms which themselves may vary considerably over time (Tandon et al., 2009). In recognition of
this heterogeneity, there is an increasing research focus on associations between symptomatology and
neurocognitive deficits, such as face processing. This chapter will address the third aim of this study:
to explore whether face-processing deficits are associated with specific symptoms. It will begin by
reviewing the literature examining links between clinical symptoms and face processing deficits. It
will then present additional results from the inpatient study, which tested associations between
symptoms and face processing across a varied sample of psychiatric inpatients.
The liability spectrum model of emotion processing in psychiatric disorders
A national survey conducted in the United States indicated that little over half of people with
a mental illness held a single DSM-IV diagnosis; 25% met criteria for two, and 23% had three or
more disorders (Kessler, Chiu, Demler, Merikangas, & Walters, 2005). In recognition of these high
rates of comorbidity among psychiatric disorders, some authors are moving towards a dimensional
approach for conceptualising psychopathology. This approach may be particularly useful for
elucidating trends in otherwise heterogeneous disorders which have substantial symptom overlap with
other disorder categories, such as schizophrenia.
Noting that facial emotion recognition deficits have been reported to varying degrees across a
range of disorders, Kret and Ploeger (2015) proposed their liability continuum model to explain these
deficits. They argue that disrupted emotional processing, including recognition of facial expressions,
forms the basis of a broad range of disorders including affective disorders and psychosis. It is
suggested that disrupted emotional processing creates a vulnerability to mental illness by biasing
individuals to detect fewer positive emotional cues and more negative or emotional cues, leading to
more aversive emotional experiences which may culminate in either internalising or externalising
disorders. This liability spectrum could help to explain the high comorbidity between disorders, and
places importance on the role that emotion perception plays in the precipitation and maintenance of
clinical symptoms.
A major criticism, however, is that this model assumes that emotion perception difficulties
and associated brain changes precede the emergence of psychiatric symptoms. It is not yet clear
whether these deficits genuinely underlie clinical symptoms, or if they simply co-occur as part of the
132
disorder presentation. Kret and Ploeger (2015) note that attempts to clarify the causality of these
deficits are challenged by the effects of medication, improper diagnosis, and the reality that
individuals typically experience symptoms for many years prior to receiving a diagnosis.
Clinical symptoms in schizophrenia
Schizophrenia spectrum disorders are characterised by a range of diverse symptom such as
delusions, cognitive impairment, motor symptoms, apathy and perceptual disturbance. Various factor
models have been proposed to classify these symptoms, but the most widely supported of these is the
three-factor model (Blanchard & Cohen, 2006). Although the precise clustering of symptoms for each
factor varies somewhat across studies, it typically comprises a positive factor, a negative factor, and a
disorganised factor. Positive symptoms include distortions of reality such as delusions and
hallucinations (Tandon et al., 2009). The most common delusions are persecutory delusions and
delusions of reference, while the most common hallucinations are auditory in nature (i.e.: voices),
although hallucinations may occur in any modality. In clinical practice, positive symptoms are treated
as the formal onset of a schizophrenia spectrum disorder, however other symptoms often precede
these for many months or years. In contrast, negative symptoms involve a loss or reduction of
function, such as restricted affect, lack of initiative, poverty of speech, social withdrawal, and
anhedonia (Tandon et al., 2009). Unlike positive symptoms, negative symptoms are not ameliorated
by antipsychotic medication, and tend to persist outside acute episodes of psychosis (Stahl & Buckley,
2007). The final factor, disorganised symptoms, refers to both disorganisation of thought and
behaviour. Sometimes referred to as ‘formal thought disorder’, these include symptoms such as
circumstantiality, tangentiality, thought derailment, neologisms (inventing new words), loosening of
associations, and poverty of thought content (Tandon et al., 2009). Behaviourally, disorganised
symptoms can present as inappropriate affect (e.g.: crying in response to a joke), bizarre rituals or
behaviours, and strange attire. Disorganised symptoms are most prominent during acute episodes, and
are more strongly correlated with cognitive impairment than either positive or negative symptoms
(Harvey, Patterson, Potter, Zhong, & Brecher, 2006; Ventura, Thames, Wood, Guzik, & Hellemann,
2010).
Commonly used instruments for assessing symptoms
One of the most popular measures for assessing clinical symptoms in schizophrenia is the
Positive and Negative Syndrome Scale (PANSS) developed by Kay, Fiszbein and Opler (1987). This
scale contains 30 different items rated on a 7-point Likert scale (such as excitement, hallucinations,
and conceptual disorganisation) which are then organised into three subscales: positive symptoms,
negative symptoms, and general psychopathology. Although not directly corresponding to the three
133
factor model described above, these subscales are shown to be largely normally distributed,
independent from one another, and robust to the effects of medication, transient mood states, and
cognition (Mortimer, 2007). Due to its sensitivity to symptom changes over time, the PANSS is
considered by some to be the gold standard for measuring treatment efficacy in schizophrenia
(Kumari, Malik, Florival, Manalai, & Sonje, 2017).
Two other widely used instruments are the Scale for the Assessment of Negative Symptoms
(SANS) and the Scale for the Assessment of Positive Symptoms (SAPS) developed by Andreasen
(1984a, 1984b). These were the first standardised scales developed to measure positive and negative
symptoms separately, consisting of 34 (SAPS) and 19 (SANS) items, respectively. Despite the
popularity of these scales, they have been criticised for condensing symptoms into just two scales:
positive and negative. However, subsequent authors have published modified scoring systems for the
SAPS and SANS to include a separate ‘disorganised’ scale, to conform with three-factor models of
schizophrenia (Klimidis, Stuart, Minas, Copolov, & Singh, 1993).
Another instrument that has been used in research with a wide variety of psychiatric
populations is the Brief Psychiatric Rating Scale (BPRS; Overall & Gorham, 1962) . Depending on
the version used it includes either 16, 18, or 24 items covering various psychotic and affective
symptoms. The BPRS had often been favoured for its brief administration time and high sensitivity to
change over time (Mortimer, 2007). However, it lacks subscales and has been criticised for
insufficient coverage of negative symptoms (Mortimer, 2007).
A less frequently used measure is the Krawiecka-Manchester Scale (K-MS; Krawiecka,
Goldberg & Vaughan, 1977), a brief 8-item scale designed for screening for psychosis. While easy to
administer, this scale lacks reliability and validity estimates and was standardised based on a mere 10
psychiatric patients.
The final measure mentioned here is the Royal Park Multidiagnostic Instrument for Psychosis
(RPMIP), developed by McGorry (1989). Developed primarily as a diagnostic tool, the RPMIP
consists of multiple interviews and questionnaires conducted over time. Factor analysis of 92
symptoms drawn from the RPMIP revealed a four factor model, which some researchers have used to
examine correlations with face processing impairments in schizophrenia (McGorry, Bell, Dudgeon, &
Jackson, 1998).
134
Associations between symptom ratings and emotion processing in
schizophrenia
Table 11.1 summarises the results from 78 symptom correlation experiments published since
1991. This section will review these findings according to the type of task used to measure emotion
processing. All associations mentioned refer to negative correlations unless otherwise specified.
Emotion labelling tasks using five or more basic emotions
The most widely used emotion labelling task involves simply presenting one image at a time
and requiring participants to select one of five (or more) options that matches the expression shown.
Of the 32 experiments included here, 19 reported no significant correlations with symptoms, or
declined to report due to “no interpretable pattern” of results.
Of those studies employing the PANSS, one reported correlations with both positive and
negative symptoms (Addington & Addington, 1998), one found a correlation with positive symptoms
only (Sachse et al., 2014), and one found a correlation with anxiety only, but also reported lower
accuracy for a group of patients with positive symptoms compared to a group without (Hall et al.,
2004). Of those experiments using the SAPS and SANS, three reported correlations with negative
symptoms, particularly Alogia and Affective Flattening, (Edwards et al., 2001; Johnston et al., 2006;
Sergi et al., 2007), while another reported correlations with positive symptoms and poor attention
(Chambon et al., 2006). One study using the BPRS reported a correlation with Anergia (Mueser et al.,
1996), while another found a correlation with Conceptual Disorganisation only (Kee et al., 2003). A
single experiment employing the K-MS reported no significant correlations with task performance
(Evangeli & Broks, 2000). However, a study using the RPMIP found an association between reduced
accuracy for sad and fearful faces and the “Bleulerian” factor, a factor which incorporates a
combination of negative, disorganised, and catatonic symptoms (Edwards et al., 2001).
In addition, three studies were included here which did not examine correlations with
symptom measures, but instead compared groups of patients with specific symptoms. Strauss and
colleagues (2011) found that patients with high negative symptoms (termed ‘deficit syndrome’)
performed less accurately than patients with minimal negative symptoms. Finally, two older studies
found that patients with paranoid schizophrenia were better at recognising emotions than a non-
paranoid group (Kline, Smith, & Ellis, 1992; Lewis & Garver, 1995), although it is unclear if this was
due to the presence of fewer negative symptoms in the paranoid group.
Overall, there was a high percentage of studies reporting non-significant correlations. This is
unlikely to be due to low power alone, as many of these studies had substantial sample sizes (e.g.;
n=30 or more) which exceeded those used in studies reporting significant findings. Significant
correlations were found for both inpatients and outpatients, with most studies reporting an association
135
with negative symptoms rather than positive or disorganised. It is possible that the high number of
null results is due to the majority of these studies employing Ekman faces, which have been criticised
for lacking ecological validity and having poor sensitivity to subtler face processing impairments
(Harms et al., 2010).
Restricted-choice emotion labelling tasks
Experiments discussed in this section employed tasks which offered only two or three cued
choices per trial (e.g.: anger or disgust?), rather than selecting from the full range of basic emotions.
Only four of the fifteen experiments in this category reported null findings. One study using the
PANSS reported significant correlations with negative symptoms, positive symptoms, and two PCA
factors termed ‘Psychomotor Poverty’ (encompassing blunted affect, social withdrawal, and poor
rapport) and ‘Disorganisation’ (thought disorder) for happy faces only (Loughland et al., 2002b). Of
studies using the SAPS and SANS, six reported correlations with negative symptoms, especially
Alogia and Affect, (Baudouin et al., 2002; Gur et al., 2006; Kohler et al., 2000; Sachs et al., 2004;
Schneider et al., 1995; Turetsky et al., 2007), two found correlations with positive symptoms (Kohler
et al., 2000; Silver, Shlomo, Turner, & Gur, 2002), and two found correlations with disorganised
symptoms (Kohler et al., 2000; Schneider et al., 1995). Finally, one study using the BPRS found a
correlation with “schizophrenia-specific” symptoms only (Schneider et al., 1995). Of those studies
which compared groups based on symptoms, one reported poorer performance in patients with
schizophrenia-specific affective symptoms (compared to high non-specific affective symptoms)
(Heimberg, Gur, Erwin, Shtasel, & Gur, 1992), while another reported poorer performance in a
negative symptom group compared to a positive symptom group (Mandal, Jain, Haque-Nizamie,
Weiss, & Schneider, 1999). Overall, the majority of studies reported correlations between negative
symptoms and face processing deficits, regardless of the symptom scale used.
Same-or-different emotion discrimination tasks
Same-or-different tasks require participants to compare two faces, either serially or
simultaneously, and decide if they are showing the same emotion. Unlike labelling tasks, they do not
require overt categorisation of the emotion and are therefore thought to be less impacted by
impairments in semantic memory. Five of the 15 experiments reviewed here reported no significant
associations with symptoms. Two studies reported correlations between discrimination accuracy and
negative symptoms on the PANSS (Addington & Addington, 1998; Fakra et al., 2015). Two studies
using the SAPS and SANS reported correlations with negative symptoms (Doop & Park, 2009;
Edwards et al., 2001), and one reported a positive correlation with the Delusions item (Martin et al.,
2005). Of studies using the BPRS, one reported correlations with both positive symptoms and
cognitive disorganisation (Leitman et al., 2005) and one found a correlation with overall BPRS scores
only (Doop & Park, 2009). One study utilising the RPMIP reported a correlation between
136
performance and a factor comprised of a combination of negative, disorganised, and catatonic features
(Edwards et al., 2001). Finally, one experiment found that a group of patients with high negative
symptoms was less accurate than a group with low negative symptoms (Strauss et al., 2010), however
no other symptoms were analysed. Overall, the patterns seen in emotion discrimination tasks appear
to mirror those seen in labelling tasks, with poorer performance most frequently associated with worse
negative symptoms.
Judgements of emotional intensity
A handful of studies have examined the relationship between symptoms and performance on
tasks that require judgements of emotional intensity, such as showing two faces and instructing
participants to indicate the face that looks more afraid. It has been argued that these tasks may be
more sensitive to subtle impairments in emotion processing compared to more straightforward
labelling tasks (Norton et al., 2009). One of the four studies reported no significant associations with
scores on the SAPS, SANS or BPRS (Sachs et al., 2004). One study found correlations with both
negative symptoms and general psychopathology on the PANSS, although these were significant for
fearful faces but not happy faces (Norton et al., 2009). Finally, of studies using the SAPS and SANS,
one reported a correlation with negative symptoms and Anhedonia (Silver et al., 2002), while another
found a correlation with Flattened Affect only (Gur et al., 2006). Although only four studies were
included here, they again show a tendency to show more associations with negative symptoms than
other types of symptoms.
Tasks using digitally generated or modified faces
The vast majority of experiments in this area have used photographs to assess emotion
processing. However, three studies have examined associations with symptoms in schizophrenia using
faces that were either entirely computer generated, or digitally ‘degraded’ in order to increase the
difficulty of the task. Hall and colleagues (2004) found that a group of patients with positive
symptoms were less accurate than a group without. However, the only item from the PANSS that
significantly correlated with performance was Anxiety. Van’t Wout and colleagues (2007) reported a
correlation with PANSS negative symptoms only, and only for fearful faces. However, a large-scale
study by Fett and colleagues (2013) found significant correlations with Disorganised, Negative, and
Positive symptoms on the PANSS.
Tasks using dynamic faces
As discussed in Chapter 5, dynamic (video-based) face stimuli provide more emotional
information than traditional static stimuli and are associated with different patterns of neural
activation (Ambadar et al., 2005; Arsalidou et al., 2011). It is therefore possible that these stimuli
would show correlations with different patterns of symptomatology as well. In support of this idea,
137
Johnston and colleagues (2010) found that the ability to distinguish between static faces was
correlated with negative symptoms on the PANSS, while performance for dynamic faces was
correlated with positive symptoms only. In contrast, a study by Mendoza and colleagues (2011) found
significant associations between accuracy for dynamic faces and both positive and negative symptom
scales on the PANSS. Two studies using the SAPS and SANS reported no significant associations
with symptoms (Behere et al., 2011; Bellack et al., 1996). However, one study found a correlation
between performance and conceptual disorganisation on the BPRS (Kee et al., 2003). Finally, a study
by Davis and Gibson (2000) compared two tasks using posed dynamic expressions (i.e. actors) and
genuine dynamic expressions (i.e. elicited by stimuli). They found that a group of patients with
paranoid symptoms were better at distinguishing genuine faces compared to a group without paranoid
symptoms. However, both groups were equally impaired at distinguishing posed emotions. This result
suggests that tasks using more ‘artificial’ expressions may be more difficult to distinguish, and as a
result may be less able to differentiate between the effects of different symptoms. Clearly, more
research is required to explore the relationship between dynamic emotion recognition and
symptomatology in schizophrenia.
Conclusions and summary
Overall, the most frequently reported correlate with emotion processing deficits in studies of
schizophrenia was severity of negative symptoms. This correlation was found across all task types,
and across different symptom measures. Studies using instruments specialised for schizophrenia, such
as the PANSS and SAPS/SANS, produced more significant correlations than more generalised
instruments, such as the BPRS and K-MS, although it is possible that this simply reflects the greater
use of these specialised instruments. Finally, studies using labelling tasks that involved a broad range
of expressions (five or more) appeared to produce null results more often than other task types. While
the reason for this is unclear, it is possible that tasks with only a few response options (e.g.: fear or
disgust?) are more suited to differentiating between levels of ability than tasks with a wider range of
response options. No clear patterns emerged regarding inpatient or outpatient status of the patients
tested.
A smaller number of studies reported significant correlations with positive or disorganised
symptoms. Again, no clear pattern was found regarding task type, symptom measure, sample size, or
population tested. The only possible exception to this was studies using dynamic face stimuli, with
one study suggesting that recognising static faces is affected by negative symptoms, while dynamic
faces is affected by positive symptoms (Johnston et al., 2010). However, this finding has not been
replicated.
138
Table 11.1. Studies reporting symptom correlates with reduced performance on emotion tasks in schizophrenia spectrum disorders.
Symptom Measure Task Author (year) Participants Results
Emotion Labelling Tasks – select 1 from 5+ choices (e.g.: happiness, sadness, anger, fear, disgust, surprise, neutral)
PANSS 7-choice, Ekman faces (Ekman &
Friesen, 1976)
Hall et al. (2004) 20 SZ* (-) correlation with Anxiety.
Group with positive symptoms less accurate than group
without.
Lewis & Garver (1995) 18 SZ* N. S.
Addington & Addington (1998) 40 SZ (inpatient, then 3
months later)
(-) correlation with Negative Symptoms and Positive
Symptoms (outpatient phase only).
Amminger et al. (2011) 30 first episode SZ
(outpatient)
79 ultra-high risk of psychosis
N. S.
7-choice task (Bölte et al., 2002) Sachse et al. (2014) 19 paranoid SZ* (-) correlation with Positive Symptoms.
9-choice task (Kucharska-Pietura et
al., 2005)
Kucharska-Pietura et al. (2005) 100 SZ (inpatients) N. S.
5-choice Morphed-intensity task
(Bediou, Krolak-Salmon, et al.,
2005)
Bediou, Krolak-Salmon et al.
(2005)
29 SZ* N. S.
Bediou et al. (2007) 40 drug-naïve first-episode
psychosis (inpatient)
N. S.
SANS/SAPS 7-choice, Ekman faces (Ekman &
Friesen, 1976)
Wölwer et al. (1996) 36 remitted SZ, 32 acute SZ
(inpatient)
No consistent relationship found (SANS only).
Streit et al. (1997) 16 SZ (tested as inpatient and
4 weeks later)
N. S. (SANS only).
139
Green et al. (2007) 20 SZ/ SZA (outpatient) N. S.
Edwards et al. (2001) 29 SZ, 23 affective psychoses,
28 other psychosis (inc. SZA
and schizophreniform)
(first episode, inpatient).
(-) correlation with Negative Symptoms (but
inconsistent across groups).
7-choice task (Johnston et al., 2006) Johnston et al. (2006) 18 SZ (outpatient) (-) correlation with Negative Symptoms (items Alogia
and Affective Flattening).
6-choice FEIT (Kerr & Neale, 1993) Bellack et al. (1996) 35 SZ/SZA (inpatient) N. S. (SANS only)
Silver & Shlomo (2001) 36 chronic SZ (inpatient) N. S.
Leitman et al. (2005) 43 SZ* N. S.
Sergi et al. (2007) 100 SZ/SZA (outpatient) (-) correlation with Affective Flattening.
6-choice task (Chambon et al., 2006) Chambon et al. (2006) 26 SZ (inpatient) (-) correlation with Positive Symptoms (items
Hallucinations, Delusions, and Bizarre Behaviour).
(-) correlation with Attention.
5-choice task (Bediou, Franck et al.
2005)
Bediou, Franck et al. (2005) 30 SZ* N. S.
BPRS 7-choice, Ekman faces (Ekman &
Friesen, 1976)
Lewis & Garver (1995) 18 SZ* N. S.
Wölwer et al. (1996) 36 remitted SZ, 32 acute SZ
(inpatient)
No consistent relationship found.
Streit et al. (1997) 16 SZ (tested as inpatient and
4 weeks later)
N. S.
6-choice FEIT (Kerr & Neale, 1993) Bellack et al. (1996) 35 SZ/SZA (inpatient) N. S.
Mueser et al. (1996) 28 SZ (inpatient) (-) correlation with Anergia.
Salem et al. (1996) 23 SZ (inpatient) N. S.
140
Kee et al. (2003) 81 SZ/ SZA
/schizophreniform (outpatient)
(-) correlation with Conceptual Disorganisation
Note: Used summed scores combining face, voice, and
video tasks.
Leitman et al. (2005) 43 SZ* N. S.
K-MS 7-choice, Ekman faces (Ekman &
Friesen, 1976)
Evangeli & Broks (2000) 12 SZ (inpatient) N. S.
RPMIP 7-choice, Ekman faces (Ekman &
Friesen, 1976)
Edwards et al. (2001) 29 SZ, 23 affective psychoses,
28 other psychosis (inc. SZA
and schizophreniform)
(first episode, inpatient).
(-) correlation with Bleulerian factor (negative,
disorganised, and catatonic features) for sadness/fear
recognition sum scores.
Note: scores summed across multiple affect tasks.
N/A – compared
paranoid and non-
paranoid groups.
7-choice, Ekman faces (Ekman &
Friesen, 1976)
Kline et al. (1992) 14 paranoid SZ, 13 non-
paranoid SZ (outpatient)
Paranoid SZ group distinguished negative emotions
better than non-paranoid group, and equivalent to HCs.
Lewis & Garver (1995) 18 SZ* Paranoid SZ group more accurate than non-paranoid
group.
N/A – compared
deficit syndrome and
non-deficit groups.
7-choice, Ekman faces (Ekman &
Friesen, 1976)
Strauss et al. (2010) 15 deficit-syndrome SZ, 15
non-deficit syndrome SZ.
Deficit syndrome (high negative symptoms) group less
accurate than non-deficit (low negative symptoms)
group.
Emotion Labelling Tasks – Restricted choice (2-3 emotions only)
PANSS 3-choice task (Faces from Mazurski
& Bond, 1993)
(Neutral, happy or sad?)
Loughland et al. (2002) 65 SZ (outpatient) (-) correlation with Negative Symptoms and Positive
Symptoms.
(-) correlation with PCA factors ‘Psychomotor Poverty’
(blunted affect, social withdrawal, rapport) and
‘Disorganisation’ for happy faces only.
Faces Test de Achaval et al. (2010) 20 SZ (outpatient) N. S.
141
(Baron-Cohen, Wheelwright, &
Jolliffe, 1997).
(select 1 of 2 choices; varying
emotion pairs)
SAPS/SANS PEAT (Erwin et al., 1992)
bipolar scale from Happy to Sad
Schneider et al. (1995) 40 SZ (mostly outpatient) (-) correlations with Negative Symptoms and the
SANS/SAPS global factors 1 (Affect, alogia, apathy &
anhedonia) and 2 (Thought disorder, bizarre behaviour
& attention; Gur et al., 1991).
Schneider et al. (1998) 13 SZ* N. S.
Kohler et al. (2000) 35 SZ* (-) correlation with Alogia (SANS), Hallucinations and
Thought Disorder (SAPS).
Silver et al. (2002) 24 chronic SZ (inpatient) (-) correlation with Positive Symptoms for happy faces.
Sachs et al. (2004) 40 SZ (inpatient) (-) correlation with Affect for happy faces.
(-) correlation with Alogia for sad faces.
Gur et al. (2006) 162 SZ (inpatient and
outpatient)
Affect scores (SANS) predicted accuracy.
Group of patients with Flat Affect were less accurate
than group without.
Note:scores were summed across PEAT and
EMODIFF.
Turetsky et al. (2007) 16 SZ (outpatient) (-) correlation with Negative Symptoms and Alogia for
happy faces.
2-choice task
(Anger or Fear?)
Baudouin et al. (2002) 12 SZ (inpatient) (-) correlation with Negative Symptoms.
2-choice task
(Tottenham et al., 2002)
Leppänen et al. (2006) 44 SZ (Xhosa ethnicity)* N. S.
142
(Happy or Angry?)
BPRS PEAT (Erwin et al., 1992)
bipolar scale from Happy to Sad
Schneider et al. (1995) 40 SZ (mostly outpatient) (-) correlation with SZ-specific symptoms.
Schneider et al. (1998) 13 SZ* N. S.
N/A – Compared
groups based on
affective symptom
severity.
PEAT (Erwin et al., 1992)
bipolar scale from Happy to Sad
Heimberg et al. (1992) 20 SZ (inpatient) Group with high SZ-specific affective symptoms was
less accurate than groups with high non-specific
affective symptoms, and both groups with low
symptoms.
N/A – Compared
groups based on
positive or negative
symptoms.
PEAT (Erwin et al., 1992)
bipolar scale from Happy to Sad
Mandal et al. (1999) 12 SZ with predominant
positive symptoms, 12 SZ
with negative symptoms
(inpatient)
Negative symptom group impaired on all emotions.
Positive symptom group impaired on sad faces only.
Same-Or-Different Emotion Discrimination Tasks
PANSS Facial affect matching task
(adapted from Feinberg et al., 1986)
Addington & Addington (1998) 40 SZ (inpatient, then 3
months later)
(-) correlation with Negative Symptoms (inpatient
phase only).
Fakra et al. (2015) 30 SZ (unmedicated)* (-) correlation with Negative Symptoms.
SAPS/SANS Facial affect matching task
(adapted from Feinberg et al., 1986)
Edwards et al. (2001) 29 SZ, 23 affective psychoses,
28 other psychosis (inc. SZA
and schizophreniform)
(first episode, inpatient).
(-) correlation with Negative Symptoms (but
inconsistent across groups).
Facial Emotion Discrimination Task
(Kerr & Neale, 1993)
Silver & Shlomo 2001) 36 chronic SZ (inpatient) N. S.
Leitman et al. (2005) 43 SZ* N. S.
Facial Affect Recognition/
Discrimination (Bediou, Franck et
al., 2005)
Bediou, Franck et al. (2005) 30 SZ* N. S.
143
Same-or-different emotion matching
task (Martin et al., 2005)
Martin et al. (2005) 20 SZ (inpatient) (+) correlation with Delusions (SAPS).
3-choice emotion matching task
(Doop & Park, 2009)
Doop & Park (2009) 16 SZ/SZA* (-) correlation with Negative Symptoms.
BPRS Facial Emotion Discrimination Task
(Kerr & Neale, 1993)
Mueser et al. (1996) 28 SZ (inpatient) N. S.
Salem et al. (1996) 23 SZ (inpatient) N. S.
Leitman et al. (2005) 43 SZ* (-) correlation with Positive Symptoms and Conceptual
Disorganisation.
3-choice emotion matching task
(Doop & Park, 2009)
Doop & Park (2009) 16 SZ/SZA* (-) correlation with BPRS scores.
RPMIP Facial affect matching task
(adapted from Feinberg et al., 1986)
Edwards et al. (2001) 29 SZ, 23 affective psychoses,
28 other psychosis (inc. SZA
and schizophreniform)
(first episode, inpatient).
(-) correlation with Bleulerian factor (negative,
disorganised, and catatonic features) for sadness/fear
recognition sum scores.
Note: scores summed across multiple affect tasks.
N/A – compared
deficit syndrome and
non-deficit groups.
Facial affect matching task
(adapted from Feinberg et al., 1986)
Strauss et al. (2010) 15 deficit-syndrome SZ, 15
non-deficit syndrome SZ.
Deficit syndrome (high negative symptoms) group less
accurate than non-deficit (low negative symptoms)
group.
Judgements of Emotional Intensity
PANSS 2-choice morphed emotion
discrimination (Norton et al., 2009)
Which face looks more afraid/happy?
Norton et al. (2009) 32 SZ/ SZA (inpatient) (-) correlation with Negative Symptoms and General
Psychopathology for fearful faces.
(-) correlation with Negative Symptoms for happy
faces.
SAPS/SANS EMODIFF (Erwin et al., 1992) Silver et al. (2002) 24 chronic SZ (inpatient) (-) correlation with Negative Symptoms and Anhedonia.
Sachs et al. (2004) 40 SZ (inpatient) N. S.
144
Which face in pair is more
sad/happy?
Gur et al. (2006) 162 SZ (inpatient and
outpatient)
Affect scores (SANS) predicted accuracy.
Group of patients with Flat Affect were less accurate
than group without.
Note:scores were summed across PEAT and
EMODIFF.
BPRS EMODIFF (Erwin et al., 1992)
Which face in pair is more
sad/happy?
Sachs et al. (2004) 40 SZ (inpatient) N. S.
Tasks Using Digitally Generated/Modified Faces
PANSS
Computer-generated morphed
emotion labelling (Sprengelmeyer et
al, 1996)
Hall et al. (2004) 20 SZ* (-) correlation with Anxiety.
Group with positive symptoms less accurate than group
without.
Degraded, morphed emotion
labelling task (Frigero et al., 2002)
4-choice
Van’t Wout et al. (2007) 37 SZ/ SZA /
Schizophreniform (inpatient
and outpatient)
(-) correlation with Negative Symptoms for fearful
faces only.
Fett et al. (2013) 1032 non-affective
psychosis*
(-) correlation with Disorganised, Negative, and
Positive Symptoms.
Note: 5 factor PANSS model used (Van der Gaag et al.,
2006)
Tasks Using Dynamic Faces
PANSS Static and Dynamic Emotion Task
(Johnston et al., 2010)
Fear or surprise?
Johnston et al. (2010) 19 SZ/SZA (outpatient) (-) correlation with Negative Symptoms for static faces.
(-) correlation with Positive Symptoms for dynamic
faces.
EEMT (Blaire et al., 2004) Mendoza et al. (2011) 93 SZ* (-) correlations with both Positive and Negative
symptoms.
145
SAPS/SANS The Videotape Affect Perception
Test (Bellack et al. 1996)
6-choice, label emotion shown in
movie clip (not limited to face).
Bellack et al. (1996) 35 SZ/SZA (inpatient) N. S.
TRENDS (Behere, 2008) Behere et al., (2011) 63 SZ (unmedicated,
outpatient)
N. S.
BPRS The Videotape Affect Perception
Test (Bellack et al. 1996)
6-choice, label emotion shown in
movie clip (not limited to face).
Bellack et al. (1996) 35 SZ/SZA (inpatient) N. S.
Kee et al. (2003) 81 SZ/ SZA
/schizophreniform (outpatient)
(-) correlation with Conceptual Disorganisation
Note: Used summed scores combining face, voice, and
video tasks.
N/A – compared
paranoid and non-
paranoid patients.
Dynamic emotion labelling task
Davis & Gibson, 2000)
6-choice
Davis & Gibson (2000) 10 paranoid SZ, 10 non-
paranoid SZ
(inpatient)
Paranoid SZ group distinguished genuine negative
emotions better than non-paranoid group, and
equivalent to HCs.
Groups equally impaired on posed emotions.
Abbreviations: PANSS = Positive and Negative Syndrome Scale (Kay, Fizbein & Opler, 1987); SAPS = Scale for the Assessment of Positive Symptoms (Andreasen, 1984); SANS = Scale for
the Assessment of Negative Symptoms (Andreasen, 1984); BPRS = The Brief Psychiatric Rating Scale (Overall & Gorham, 1962); K-MS =. Krawiecka–Manchester Scale (Krawiecka,
Goldberg, & Vaughan, 1977); RPMIP = Royal Park Multi-Diagnostic Instrument for Psychoses (McGorry et al., 1989); FEIT = Facial Emotion Identification Test (Kerr & Neale, 1993); PEAT
= Penn Emotion Acuity Test (Erwin et al., 1992); SZ = schizophrenia; SZA = schizoaffective disorder.
* = did not specify whether sample was inpatient or outpatient.
146
Table 11.2. Studies reporting symptom correlates with reduced performance on non-emotion face tasks in schizophrenia spectrum disorders.
Task Design Symptom Measure Author (year) Participants Significant Results
Identity Recognition Tasks
Benton Test of Facial
Recognition
(Benton et al., 1978)
PANSS Addington & Addington (1998) 40 SZ (inpatient, then 3
months later)
(-) correlation with Negative Symptoms (outpatient
phase only).
Hall et al., (2004) 20 SZ* N. S.
BPRS Mueser et al. (1996) 28 SZ (long-term inpatient) (-) correlation with Anergia
Salem et al. (1996) 23 SZ (inpatient) N. S.
Sachse et al. (2014) 19 SZ (paranoid)* N. S.
K-MS Evangeli & Broks (2000) 12 SZ (inpatient) N. S.
Morphed face identity matching
task
(Match target to 1 of 2 choices)
PANSS Chen et al. (2009) 29 SZ /SZA* (-) correlation with Negative Symptoms and Positive
Symptoms.
Chen et al. (2015) 25 SZ/SZA* (-) correlation with Negative Symptoms.
Chen & Ekstrom (2016) 35 SZ/SZA (outpatient) (-) correlation with Negative Symptoms and Positive
Symptoms.
Norton et al. (2009) 32 SZ/ SZA (inpatient) (-) correlation with Negative Symptoms
Identity recognition task
(Is this Person A or B?)
SAPS/SANS
Baudouin et al. (2002) 12 SZ (inpatient) No correlations with task performance, but interference
from Identity on an Emotion task correlated (+) with
Negative Symptoms.
Identity discrimination
(same/different judgment of
serially presented faces)
PANSS Fakra et al. (2015) 30 unmedicated SZ N. S.
SAPS/SANS Martin et al. (2005) 20 SZ (inpatient) (-) correlation with Negative Symptoms (SANS).
Interference from identity on Emotion task correlated (+)
with Negative Symptoms.
Single-Feature Recognition Tasks
147
Morphed sex-labelling task PANSS Bediou et al. (2005) 29 SZ* N. S.
Bediou et al. (2007) 40 drug-naïve first-episode
psychosis (inpatient)
N. S.
Sex-labelling task PANSS Van’t Wout et al. (2007) 37 SZ/ SZA/
schizophreniform (inpatient
and outpatient)
N. S.
Age discrimination task
(over or under 30 years?)
N/A – compared
paranoid and non-
paranoid patients.
Pinkham et al. (2008) 12 paranoid SZ
12 non-paranoid SZ/SZA
N. S.
Age discrimination task
(Erwin et al., 1992)
(indicate decade:
1=teens…7=70s)
SAPS/SANS Schneider et al. (1995) 40 SZ (mostly outpatient) (-) correlations with Alogia subscale and the
SANS/SAPS global Factor 2 (Thought disorder, bizarre
behaviour and attention; Gur et al., 1991).
Kohler et al. (2000) 35 SZ* N. S.
BPRS Schneider et al. (1995) 40 SZ (mostly outpatient) N. S.
Self-Other Discrimination Tasks
Morphed self-other
discrimination task
(self or other?)
PANSS Bortolon et al. (2016) 24 SZ (inpatient and
outpatient)
N. S.
Self-other discrimination task
(self, relative or unknown face?)
PANSS Kircher et al. (2007) 20 SZ (inpatient) (-) correlation with Positive Symptoms, Negative
Symptoms, Delusions (item) and Hallucinations (item)
for self-face recognition.
(-) correlation with Positive Symptoms and PANSS sum
scores for unknown-face recognition.
Self-other discrimination task
(self or famous face?)
PANSS Yun et al. (2014) 8 SZ* (-) correlation with General Psychopathology for self-
face recognition.
148
Tests of Face Memory
Delayed face matching task
(3 and 10 s delay conditions)
PANSS Chen et al. (2009) 29 SZ/ SZA* (-) correlation with Positive Symptoms for 3 s delay
condition only.
The Warrington Recognition
Memory Test–Faces subtest
(Warrington,1984).
K-MS Evangeli & Broks (2000) 12 SZ (inpatient) N. S.
Penn Face Memory Test
(Erwin et al., 1992).
SANS Sachs et al. (2004) 40 SZ (inpatient) (-) correlation with Avolition-Apathy subscale
BPRS Sachs et al. (2004) 40 SZ (inpatient) N. S.
Famous faces task
(Calder et al., 1996).
K-MS Evangeli & Broks (2000) 12 SZ (inpatient) N. S.
Abbreviations: PANSS = Positive and Negative Syndrome Scale (Kay, Fizbein & Opler, 1987); SAPS = Scale for the Assessment of Positive Symptoms (Andreasen, 1984); SANS = Scale for
the Assessment of Negative Symptoms (Andreasen, 1984); BPRS = The Brief Psychiatric Rating Scale (Overall & Gorham, 1962); K-MS =. Krawiecka–Manchester Scale (Krawiecka,
Goldberg, & Vaughan, 1977); SZ = schizophrenia; SZA = schizoaffective disorder.
* = did not specify whether sample was inpatient or outpatient.
149
Associations between symptom ratings and identity processing in
schizophrenia
Table 11.2 summarises the results from 23 symptom correlation experiments published since
1991. This section will review these findings according to the type of task used to measure static
facial identity processing. All associations mentioned refer to negative correlations unless otherwise
specified.
Identity recognition tasks
‘True’ identity recognition tasks are those which require participants to attend to the identity
of a face shown (typically with non-face features such as hair removed) and either match it to, or
distinguish it from, another face. Of the 12 experiments reviewed here, five reported no significant
associations with symptoms. One study using the Benton Test of Facial Recognition found a
correlation with negative symptoms on the PANSS (Addington & Addington, 1998) , while another
reported a correlation with Anergia from the BPRS only (Mueser et al., 1996). Two studies using a
more subtle morphed-face matching task reported correlations with both positive and negative
symptoms on the PANSS (Chen & Ekstrom, 2016; Chen et al., 2009), while two found correlations
with negative symptoms only (Y. Chen et al., 2015; Norton et al., 2009). A single study using a
learning phase followed a recognition task (Is this face person A or person B?) found a correlation
between negative symptoms on the SANS and interference from identity information on an emotion
labelling task (Baudouin et al., 2002). This finding suggests that patients with more negative
symptoms were less able to ignore identity information during an unrelated emotion task. Finally, a
study using a same-or-different identity task found an association between performance and negative
symptoms on the SANS, as well as between negative symptoms and interference from identity
information on an emotion task (Martin et al., 2005). In sum, the majority of studies reported
correlations with negative symptoms regardless of task. However, two studies using morphed faces
(argued to be better for detecting subtle impairments in identity processing), also reported correlations
with positive symptoms.
Single-feature recognition tasks
In contrast to ‘true’ identity recognition, single-feature tasks only require participants to
attend to one aspect of identity, such as the perceived age or sex of a face. Of the six studies
identified, only one reported a significant association. Schneider and colleagues (1995) found an
association between performance on an age discrimination task and Alogia on the SANS. They also
reported a correlation with a factor derived from the SAPS and SANS which included a combination
of thought disorder, bizarre behaviour, and poor attention. The fact that most studies reported null
results likely reflects the finding that patients with schizophrenia tend to be unimpaired on simple
150
single-feature recognition tasks (e.g.: Bediou et al., 2005, 2007; Pinkham et al., 2008; with the notable
exception of Schneider et al., 1995, who reported reduced performance in SZ patients compared to
HCs).
Self-other discrimination tasks
Self-other discrimination tasks require the participant to judge whether each face shows their
own face or that of another person, which may be either familiar (e.g.: a relative) or unknown. Several
studies suggest that patients with schizophrenia and those at ultra-high risk of psychosis are impaired
on such tasks compared to healthy controls (Irani et al., 2006; Jia, Yang, Zhu, Liu, & Barnaby, 2015),
although others reported that this impairment was not specific to self-recognition (Heinisch et al.,
2013; Lee et al., 2007). Three studies were identified which compared associations between
symptoms and performance. Kircher and colleagues (2007) found correlations between self-face
recognition and both positive and negative symptoms on the PANSS, while unknown-face recognition
correlated with positive symptoms and overall PANSS scores only. In contrast, Yun and colleagues
(2014) reported a correlation between self-recognition and general psychopathology on the PANSS,
while Bortolon and colleagues (2016) found no significant correlations with PANSS scores. Clearly
further evidence is required to elucidate any potential relationship between self-recognition and
symptomatology in schizophrenia.
Tests of face memory
Unlike the other tasks described here, face memory tests require participants to draw on long-
term memory to make judgements about identity. This includes old/new paradigms, matching tasks
with a substantial delay between stimulus exposures, and famous faces tasks. The use of these tasks to
evaluate identity processing in schizophrenia has attracted criticism because they may confound face
perception ability with more general memory impairments. Only three studies were identified which
examined associations between face memory performance and symptom ratings in schizophrenia.
Chen and colleagues (2009) found a correlation with positive symptoms on the PANSS for a 3-second
delay condition, but not with a 10-second delay. Sachs and colleagues (2004) found a correlation with
the Avolition-Apathy subscale on the SANS, but no significant correlations with scores on the SAPS
or BPRS. Finally, Evangeli and Broks (2000) used two different face memory paradigms and found
no significant correlations with K-MS scores, although this is unsurprising given their small sample
size (n=12).
Conclusions and summary
While the number of correlational studies investigating identity processing are far fewer than
those reviewed in the emotion processing section, some tentative conclusions may be drawn. First, the
most frequently reported correlate of deficits seen on ‘true’ identity tasks was negative symptoms.
151
However, two studies using morphed stimuli also reported correlations with positive symptoms,
suggesting that this may be a smaller effect size requiring more sensitive techniques for assessing
deficits. Tasks that assess only one aspect of identity, such as age or sex, did not tend to correlate with
symptoms. However, this is probably because single-feature tasks are rarely able to detect impairment
in schizophrenia. Although patients with schizophrenia typically show impairment on other types of
facial identity tasks – such as self-other discrimination and tests of face memory – only a handful of
studies have examined correlations with symptoms, and findings are inconsistent.
The current study
The goal of this Results section is to clarify the relationship between face processing deficits
– both emotion and facial identity – and clinical symptoms in a large patient population (n=86).
Patients from a range of diagnostic categories other than schizophrenia were included in order to
obtain a broader spread of symptoms independent of diagnosis. The research question is therefore: do
specific symptoms correlate with either facial emotion or identity performance? This will be
addressed by examining correlations between PANSS symptom ratings and performance on a range of
dynamic face tasks. This section will also describe the results of a cluster analysis conducted to
identify possible subgroups of patients based on task performance.
Results: Symptom correlates of face-processing performance
Correlations between PANSS scores and task performance
Table 11.3 shows Pearson product-moment correlations between PANSS subscale scores and
task performance. Bootstrapping was used to calculate bias corrected and accelerated (BCa)
confidence intervals using 1000 resamples (DiCiccio & Efron, 1996). Positive Symptoms were
negatively correlated with all tasks except Sex Labelling (rs= -.31-.41). General Psychopathology
scores correlated negatively with Car Discrimination only (r=.37). No correlations with Negative
Symptoms approached significance for any of the tasks.
Correlations between individual PANSS items and task performance
Spearman-rank correlations between individual items on the PANSS and task performance
are shown in Table 11.4. BCa confidence intervals were calculated using 1000 bootstrapping samples.
When considered overall, it appears that classically positive symptoms - such as Delusions,
Grandiosity, Suspiciousness, and Unusual Thought Content – correlated negatively with performance
on the two Emotion tasks, and also with Car Discrimination. In contrast, no positive symptoms
correlated with the two Identity tasks (with the exception of Unusual Thought Content, which was
positively correlated with Sex Labelling). Note that the association between PANSS Positive subscale
152
scores and Identity Discrimination described in the previous section appears to have been driven
primarily by Cognitive Disorganisation, which the PANSS classifies as a positive item, whereas
proponents of the prevailing three-factor model would categorise this item separately as a
disorganised symptom (e.g.: Tandon et al., 2009) .
The other trend to note is that performance on all tasks (except Sex Labelling) tended to
correlate negatively with cognitive symptoms, such as Conceptual Disorganisation, Difficulty in
Abstract Thinking, Stereotyped Thinking, Poor Attention, Disorientation, and Lack of Judgement and
Insight. These items also appear to drive the association reported between General Psychopathology
subscale scores and Car Discrimination performance. Interestingly, almost no correlations were found
with affective symptoms, such as Depression, Emotional withdrawal, and Blunted Affect. The only
exception was that higher Depression and Guilt Feelings was associated with better performance on
Emotion Discrimination, but not any other tasks.
Hierarchical clustering
An exploratory hierarchical cluster analysis was conducted using Ward’s method (Ward,
1963) to establish clusters based on performance (d prime) across the five tasks. A four cluster
solution was selected based on the ‘elbow’ method (Zambelli, 2016). Characteristics of the four
clusters are shown in Table 11.5.
Demographics
One-way ANOVAs revealed that the clusters differed significantly in duration of illness,
F(3,80)=2.99, p=.04, and daily benzodiazepine dose, F(3,12)=4.02, p=.03. A difference in Age also
trended towards significance, F(3,80)=2.62, p=.06. The clusters did not differ significantly in years of
education, F(3, 80)=.82, p=.49, estimated IQ, F(3, 77)=1.31, p=.28, or average daily antipsychotic
dose, F(3, 59)=.46, p=.72. Post-hoc Bonferroni-corrected comparisons revealed that Cluster 2 had a
significantly shorter average duration of illness than Cluster 4 (Mean difference= 9.06 years, p=.048).
Cluster 2 was also significantly younger compared to Cluster 3 (Mean difference = 10.57 years,
p=.04). No other comparisons approached significance.
PANSS symptom scores
One-way ANOVAs showed significant differences between clusters for Positive Symptoms
on the PANSS, F(3, 80)=4.86, p=.004. Differences between clusters for General Psychopathology,
F(3, 80)=2.44, p=.07, were not significant, however differences for Negative Symptoms, F(3,
80)=2.70, p=.051, approached significance. Bonferroni-corrected post-hoc comparisons revealed that
Cluster 2 had significantly fewer positive symptoms than Cluster 1 (Mean difference = 5.79, p=.02)
153
and Cluster 4 (Mean difference = 7.19, p=.01). Cluster 3 had significantly fewer negative symptoms
than Cluster 4 (Mean difference = 3.30, p=.04). No other comparisons approached significance.
Task performance
Mean performance (d’) across the five dynamic tasks are shown in Figure 11. A series of one-
way ANOVAs were conducted to compare performances of the four clusters for each task. Significant
group effects were found for all tasks (p=.003 to .001). Bonferroni-corrected post-hoc comparisons
were conducted for each task.
Emotion discrimination. Cluster 2 significantly outperformed all other groups (p
values<.001), while Clusters 1 and 3 both outperformed Cluster 4 (p values=.01 to <.001).
Emotion labelling. Cluster 2 significantly outperformed all other clusters (p values<.001).
Cluster 3 outperformed Clusters 1 and 4 (p values<.001), which did not differ from one another.
Identity discrimination. Clusters 2 and 3 performed significantly better than Clusters 1 (p
values=.002 to <.001) and 4 (p values<.001). Cluster 1 also outperformed Cluster 4 (p=.02).
Sex labelling. Cluster 1 and 3 outperformed Cluster 4 (p values=.01-.003). No other
comparisons were significant.
Car discrimination. Cluster 2 performed significantly better than Cluster 1 (p=.02) and
Cluster 4 (p<.001). Clusters 1 and 3 also outperformed Cluster 4 (p values<.001) but were not
different from one another.
154
Table 11.3. Pearson correlations between PANSS scores and performance (d’) on the five dynamic tasks.
Emotion Discrimination Emotion Labelling Identity Discrimination Sex Labelling Car Discrimination
r [95% CI] r [95% CI] r [95% CI] r [95% CI] r [95% CI]
Positive Symptoms -.41* [-.55, -.24] -.33* [-.50, -.13] -.31* [-.50, -.08] .06 [-.10, .22] -.35* [.55, .11]
Negative Symptoms -.06 [-.33, .21] -.20 [-.39, .03] -.21 [-.41, .00] -.09 [-.31, .13] -.24 [-.48, .03]
General Psychopathology -.11 [-.34, .14] -.16 [-.35, .05] -.22 [-.44, .05] -.12 [-.33, .09] -.37* [-.53, -.17]
Note: * indicates a significant correlation, where the 95% confidence interval for r does not include 0.
Table 11.4. Spearman rank correlations between individual PANSS items and performance (d’) on the five dynamic tasks.
Emotion Discrimination Emotion Labelling Identity Discrimination Sex Labelling Car Discrimination
rs [95% CI] rs [95% CI] rs [95% CI] rs [95% CI] rs [95% CI]
Delusions -.42* [-.60, -.23] -.24* [-.45. -.01] -.18 [-.39, .06] .09 [-.14, .30] -.27* [-.47, -.07]
Conceptual disorganisation -.50* [-.65, -.32] -.29* [-.49, -.06] -.26* [-.46, -.03] .15 [-.09, .37] -.34* [-.55, -.10]
Hallucinations -.08 [-.27, .12] -.09 [-.30, .13] -.15 [-.36., .07] -.07 [-.30, .17] -.12 [-.32, .09]
Excitement -.16 [-.38, .07] -.20 [-.41, .03] -.22 [-.42, .04] .11 [-.13, .34] -.31* [-.51, -.06]
Grandiosity -.32* [-.52, -.09] -.29* [-.49, .08] -.17 [-.38, .07] .23* [.02, .45] -.24* [-.43, -.02]
Suspiciousness -.28* [-.48, -.06] -.25* [-.46, -.03] -.14 [-.36, .09] .02 [-.20, .22] -.25* [-.46, -.02]
Hostility -.06 [-.28, .15] -.10 [-.30, .10] -.12 [-.33, .12] -.06 [-.28, .19] -.11 [-.33, .11]
Blunted affect .02 [-.20. .25] -.10 [-.31, .12] -.20 [-.40, .02] -.05 [-.26, .17] -.14 [-.36, .13]
Emotional withdrawal .00 [-.27, .27] -.02 [-.23, .20] -.03 [-.22, .18] .02 [-.14, .16] .001 [-.17, .16]
Poor rapport .07 [-.16, .30] .04 [-.17, .23] -.03 [-.24, .19] -.08 [-.27, .12] -.05 [-.29, .19]
Passive apathetic withdrawal .21 [-.03, .43] .09 [-.12, .30] .11 [-.09, .30] -.06 [-.28, .15] .12 [-.08, .31]
155
Difficulty in abstract thinking -.24* [-.44, -.04] -.31* [-.47, .13] -.32* [-.50, -.12] -.04 [-.26, .18] -.26* [-.45, -.05]
Lack of spontaneity .05 [-.18, .30] .06 [-.18, .27] -.03 [-.24, .18] -.22 [-.40, .002] -.13 [-.36, .09]
Stereotyped thinking -.29* [-.49, -.06] -.34* [-.51, -.12] -.29* [-.47, -.07] .03 [-.21, .27] -.28* [-.47, -.04]
Somatic concern .17 [-.07, .36] .17 [-.04, .37] .18 [-.05, .40] .01 [-.23, .23] .01 [-.22, .20]
Anxiety .18 [-.02, .38] .18 [-.03, .38] .04 [-.18, .24] -.09 [-.31, .13] .03 [-.20, .26]
Guilt feelings .22* [.03, .40] .12 [-.09, .31] .05 [-.16, .27] -.10 [-.31, .12] .18 [-.04, .39]
Tension .004 [-.21, .22] .05 [-.17, .26] -.01 [-.24, .24] .004 [-.24, .26] -.16 [-.38, .07]
Mannerisms and posturing .002 [-.20, .22] .04 [-.19, .27] .07 [-.12, .25] .04 [-.19, .28] -.04 [-.20, .13]
Depression .34* [.15, .52] .13 [-.08, .32] .10 [-.12, .29] -.13 [-.35, .09] .11 [-.12, .34]
Motor retardation .09 [.-.13, .29] -.02 [-.24, .17] -.03 [-.21, .15] -.16 [-.34, .04] -.11 [-.29, .12]
Uncooperativeness -.17 [-.36, .04] -.10 [-.30, .11] -.10 [-.34, .16] -.14 [-.35, .10] -.17 [-.40, .05]
Unusual thought content -.46* [-.65, -.23] -.33* [-.52, -.12] -.15 [-.38, .09] .28* [.06, .48] -.27* [-.48, -.02]
Disorientation .02 [-.16, .20] -.12 [-.36, .13] -.26* [-.41, -.08] -.17 [-.38, .07] -.24* [-.42, -.03]
Poor attention -.28* [-.47, -.07] -.29* [-.47, -.06] -.39* [-.57, -.14] -.23 [-.44, .04] -.49* [-.63, .29]
Lack of judgement and insight -.39* [-.57, -.18] -.40* [-.57, -.19] -.19 [-.40, .04] .04 [-.20, .28] -.25* [-.48, -.01]
Disturbance of volition -.02 [-.22, .18] -.08 [-.24, .07] -.29* [-.45, -.10] -.16 [-.39, .12] -.14 [-.32, .06]
Poor impulse control -.07 [-.27, .15] -.08 [-.28, .13] -.11 [-.31, .13] -.11 [-.36, .15] -.17 [-.39, .05]
Preoccupation -.14 [-.33, .07] -.06 [-.28, .17] -.14 [-.35, .07] .09 [-.12, .31] -.30* [-.47, -.10]
Active social avoidance -.15 [-.34, .06] -.14 [-.36, .07] -.14 [-.36, .10] -.03 [-.24, .18] -.22 [-.44, .02]
Note: * indicates a significant correlation, where the 95% confidence interval for rs does not include 0.
156
Table 11.5. Characteristics of four clusters based on patient performance.
Cluster 1 Cluster 2 Cluster 3 Cluster 4
n=27 n=11 n=31 n=15
M (SD) M (SD) M (SD) M (SD)
Age 35.19 (11.70) 26.36# (9.06) 36.94# (10.62) 35.13 (10.99)
Years of
Education
11.19 (1.86) 12.18 (2.480 11.55 (3.23) 10.67 (2.29)
IQ estimate 103.80 (7.34) 102.27 (8.17) 105.35 (8.29) 100.55 (6.84)
Illness duration* 6.75 (8.33) 4.35# (4.60) 7.61 (7.96) 13.41# (11.09)
Antipsychotic
daily dose
341.15
n=24
(216.90) 238.89
n=3
(315.49) 299.09
n=22
(221.83) 379.31
n=14
(312.83)
Benzodiazepine
daily dose*
20.00
n=4
(14.72)
7.50
n=1
- 24.00
n=5
(17.10) 58.33
n=6
(27.14)
Diagnostic group
(n)
Schizophrenia 11 (41%) 0 (0%) 13 (42%) 10 (67%)
Bipolar I disorder 7 (26%) 1 (9%) 4 (13%) 3 (20%)
Other psychosis 6 (22%) 3 (27%) 7 (23%) 1 (7%)
Non-psychosis 3 (11%) 7 (64%) 7 (23%) 1 (7%)
Mean task
performance (d’)
Emotion
discrimination
1.22 (.35) 2.24 (.30) 1.43 (.59) .71 (.53)
Emotion labelling .69 (.45) 3.01 (.70) 1.77 (.49) .56 (.54)
Identity
discrimination
2.49 (.37) 3.04 (.53) 2.94 (.34) 2.08 (.50)
Sex labelling 2.12 (.47) 2.07 (.33) 2.04 (.68) 1.46 (.57)
Car discrimination 2.37 (.52) 2.88 (.63) 2.49 (.35) 1.19 (.52)
PANSS
Positive total* 16.33# (5.93) 10.55#⸸ (3.83) 13.94 (5.18) 17.73⸸ (5.37)
Negative totala 10.70 (3.18) 10.82 (4.73) 10.03# (2.99) 13.33# (5.09)
General total 30.89 (4.96) 28.64 (4.18) 31.81 (5.75) 34.60 (7.95)
*p<.05; ap=.051; #,⸸Significant difference between clusters.
157
Figure 11. Mean performance across five dynamic tasks for the four patient clusters. Error bars indicate 95% confidence intervals. *p>.05; **p>.001.
158
Discussion
The primary aim of this section was to determine whether clinical symptoms are associated
with face processing deficits in patients with a range of psychiatric disorders. It was found that both
positive symptoms and symptoms of cognitive disorganisation correlated with emotion-processing
impairments, and also with performance on a non-face control task. In contrast, mostly cognitive
symptoms – as well as cognitive disorganisation – correlated with performance on identity-processing
tasks. An exploratory hierarchical cluster analysis was also conducted to cluster patients based on task
performance. The solution produced hints at four possible subgroups with differing patterns of
psychopathology.
Associations with positive and negative symptoms
The results of this experiment suggest that performance on two dynamic emotion-processing
tasks is associated with the positive symptom items on the PANSS. This implies that regardless of
diagnosis, patients with positive symptoms such as delusions, suspiciousness, and unusual beliefs are
most likely to show emotion processing difficulties compared to those with other symptoms. This
result is at odds with the majority of past research using the PANSS and other similar scales, which
predominantly report associations with negative symptoms. These include studies using same-or-
different emotion discrimination tasks (Addington & Addington, 1998; Doop & Park, 2009; Edwards
et al., 2001; Fakra et al., 2015) as well as restricted-choice emotion labelling tasks (Baudouin, Martin,
Tiberghien, Verlut, & Franck, 2002; Gur et al., 2006; Kohler et al., 2000; Sachs, Steger-Wuchse,
Kryspin-Exner, Gur, & Katschnig, 2004; Schneider, Gur, Gur, & Shtasel, 1995; Turetsky et al., 2007),
although a smaller number of studies also reported correlations with positive symptoms (Kohler et al.,
2000; Leitman et al., 2005; Loughland et al., 2002b; Silver et al., 2002). One possible explanation for
this discrepancy is that the current study used dynamic stimuli rather than static. Johnston and
colleagues (2010) found that positive symptoms correlated with dynamic emotion processing, while
negative symptoms correlated with static emotion processing. Given that dynamic faces are shown to
elicit different patterns of brain activation compared to static faces, the authors posited that positive
and negative symptoms may have differential effects on these brain networks. However, it is worth
noting that subsequent studies using dynamic emotion tasks have not replicated this dissociation
between positive and negative symptoms (Behere et al., 2011; Mendoza et al., 2011).
Interestingly, the current study found that performance on two dynamic identity-processing
tasks was not associated with positive symptoms or negative symptoms. As no previous study has
employed dynamic stimuli to examine identity processing in patients, this is a novel finding.
However, this is inconsistent with previous studies using static stimuli, which typically did report
159
associations with negative symptoms (Martin et al., 2005; Norton et al., 2009), or with both positive
and negative symptoms (Chen & Ekstrom, 2016; Chen et al., 2009).
On the surface, the finding that positive symptoms were associated with impaired processing
of expressions, but not facial identity, fits with the idea that emotional information is selectively
impaired in patients exhibiting the positive symptoms of psychosis. However, the current study also
found that positive symptoms correlated with performance on a non-face control task. Therefore, it is
unlikely that the association with positive symptoms can be meaningfully interpreted as evidence of
any kind of emotion-selective process. Furthermore, as the four clusters differed significantly in mean
benzodiazepine dose, it is possible that performance differences were at least partly affected the
sedating side-effects of this medication.
Associations with ‘cognitive’ symptoms
The results of the current experiment showed that patient performance was associated with a
range of cognitive symptoms such as cognitive disorganisation, difficulty in abstract thinking,
stereotyped thinking, and poor attention. This pattern was seen across tasks of emotion-processing,
identity-processing, and the non-face control task. The only exception was the Sex Labelling task,
which is unsurprising given that a) this task did not distinguish between patients and healthy controls
in previous analyses (see Chapter 8), and b) previous studies using static versions of this task reported
no correlations with symptoms in schizophrenia (Bediou et al., 2005, 2007; Pinkham et al., 2008).
The observation that cognitive-related symptoms correlated negatively with task performance
suggests that patients showing more generalised cognitive difficulties tended to be less accurate
overall, regardless of stimulus type. This finding is consistent with previous studies indicating
associations between cognitive factors and emotion processing in schizophrenia (Bozikas, Kosmidis,
et al., 2006; Sachs et al., 2004; Silver et al., 2002). This finding is not unexpected, particularly given
the predominance of attentional difficulties in psychiatric disorders such as schizophrenia (Keefe &
Harvey, 2012). However, it does highlight the pervasive impact of generalised cognitive deficits, even
on tasks that are intended to tap into specialised areas of perceptual deficit, such as emotion
processing. Although some authors argue that emotion-processing and generalised cognitive
impairment are two overlapping but separable areas of deficit in schizophrenia (Barkhof, de
Sonneville, Meijer, & de Haan, 2015; Megreya, 2016), the current results align with the opposing
view that emotion-processing deficits may be fully accounted for by more general cognitive
disturbance (Pomarol-Clotet et al., 2010).
Exploratory cluster analysis
160
The present study employed hierarchical clustering analysis to identify four potential
subgroups of patients based on their performance across the five tasks. Cluster 2 tended to be
younger, have shorter illness duration, fewer positive symptoms, and had the highest accuracy across
all tasks. This group was primarily composed of patients with non-psychotic disorders, or ‘other’
forms of psychosis, such as drug-induced psychosis. In contrast, Cluster 4 was the lowest performing
across all tasks, tended to be older, have a longer duration of illness, more positive and negative
symptoms, and was primarily comprised of patients with a formal schizophrenia spectrum disorder.
The two largest groups, Clusters 1 and 3, were similar in composition, with intermediate task
performance, moderate age and illness duration, and both comprised a mix of diagnoses including
bipolar I disorder, schizophrenia spectrum, and non-psychotic and other-psychotic disorders. The only
notable difference was that, when compared to the least-impaired group (Cluster 2), Cluster 3 showed
selective impairments on emotion tasks, while Cluster 1 showed impairments across tasks.
Overall, the exploratory cluster analysis only partly approximated diagnostic groups. Patients
with schizophrenia spectrum disorders were overrepresented in the most impaired cluster, while non-
psychosis patients predominantly made up the least-impaired cluster. However, the majority of all
patients, irrespective of diagnosis, fell into one of two intermediate clusters: one with
disproportionately poor emotion-processing, and one with generally impaired performance. Therefore,
there is no evidence to suggest that performance on these tasks could reliably distinguish between
different patient groups.
Limitations
Several important caveats must be made when interpreting these results. The first is that these
analyses focused on the interpretation of individual symptom items, rather than aggregated subscale
scores, which are more statistically robust (Kay et al., 1987). While previous studies typically report
subscale scores such as ‘Positive Symptoms’ on the PANSS, these were considered less interpretable
because they do not directly correspond to the prevailing three-factor model of positive, negative and
disorganised symptoms (Mortimer, 2007). The findings cited in previous studies may be further
complicated by the fact some symptom scales conflate generalised cognitive symptoms with non-
cognitive positive and negative symptoms. For example, the PANSS includes ‘Cognitive
Disorganisation’ in the calculation of positive symptom scores, and ‘Difficulty in Abstract Thinking’
and ‘Stereotyped Thinking’ in the calculation of overall negative symptoms. A further limitation is
that the cluster analysis described was only exploratory in nature, lacked theoretical context, and
produced small, uneven-sized groups. Therefore, these results should be interpreted with appropriate
caution.
Conclusions
161
This section explored associations between clinical symptom ratings and performance on a
series of tasks evaluating emotion-processing, identity-processing, and non-face discrimination tasks.
It was found that non-cognitive positive symptoms – such as delusions and suspiciousness – were
negatively correlated with both emotion-processing and non-face discrimination, but not identity-
processing ability. Although inconsistent with the majority of previous literature using static tasks,
this finding agrees with one study using dynamic stimuli (Johnston et al., 2010). The finding that
positive symptoms correlated with performance on a non-face task suggests that these symptoms are
unlikely to indicate an emotion-specific processing impairment. Furthermore, more severe cognitive
symptoms were associated with generally reduced performance across tasks. It is suggested that
previous research using aggregate scores from instruments such as the PANSS may conflate the
presence of cognitive and non-cognitive symptoms in schizophrenia. Therefore, future studies using
these instruments should consider individual items, or alternative groupings of items, to evaluate the
role of specific symptom clusters.
162
Chapter 12: General Discussion
The overarching aim of this thesis was to better characterise the nature of face-processing
deficits in schizophrenia and other psychiatric disorders. This involved addressing the following
research questions:
1. Defining the deficit: Is the emotion-processing deficit specific to emotion, or is it better
explained by broader deficits in processing faces, or a generalised cognitive impairment?
2. Are these deficits specific to schizophrenia, or shared across similar disorders?
3. What symptoms correlate with these deficits?
While emotional face-processing in schizophrenia has attracted a substantial body of research,
inconsistent findings across studies mean that the questions above remain unanswered. This
inconsistency is likely due to differences in patient populations studied, small sample sizes,
differences in methods used to assess face-processing, and a tendency to not include control tasks. A
major point raised in Chapter 5 of this thesis is that the majority of past studies used only static
images to assess face-processing, which lack the ecological validity and sensitivity of dynamic
(video-based) stimuli. In recognition of this issue, I created a novel set of dynamic morphed stimuli
that were designed to assess not only facial emotion-processing, but also non-emotional face-
processing (i.e.: facial identity) and non-face processing (i.e.: cars). To the best of my knowledge, this
is the first stimulus set to combine morphing with dynamic video to test non-emotional and non-face
processing ability. These stimuli were used in a large study of healthy controls (n=82) and two studies
of psychiatric inpatients (n=106 and n=45), and are available to download from
http://go.unimelb.edu.au/e3t6.
The healthy control study revealed that, as predicted, the newly-developed set of dynamic
stimuli were recognised more accurately than static versions of the stimuli. Furthermore, correlations
with schizotypy scores showed that healthy controls who were high in schizophrenia-like traits were
generally less accurate at labelling dynamic emotions compared to individuals who were low in
schizotypy. This suggests that these new dynamic tasks are sensitive to the subtle emotion-processing
deficits typically seen in individuals with schizophrenia-like traits.
The main inpatient study revealed that, as expected, patients with schizophrenia showed
impairments in both labelling and discriminating between emotions compared to healthy controls.
However, performance on a non-face task was also impaired, and performance on a non-emotional
(identity) face task was marginally impaired. An additional inpatient study designed to assess
visuospatial attention revealed that although patients with schizophrenia showed marked impairment
in attending to the ‘global’ level of a stimulus, this deficit does not appear to be associated with
163
emotion-processing ability. Overall, these findings suggest that the “emotion-processing deficit” seen
in schizophrenia likely extends to non-face and non-emotional tasks as well. However, these deficits
cannot be explained by broader impairments in the allocation of visual attention.
Considering all psychiatric disorders, the main inpatient study revealed that patients with
schizophrenia and those with bipolar disorder both showed comparable deficits in emotion processing,
accompanied by intact non-emotional face-processing (identity processing) compared to age-matched
healthy controls. These groups also showed similar deficits in discriminating non-face objects (cars).
Patients with non-schizophrenia psychosis showed mild deficits in emotion-processing, but no other
significant deficits. In contrast, patients with non-psychotic disorders performed no differently to
healthy controls. Correlational analyses showed that deficits in emotion-processing and non-face
processing were both associated with positive symptoms of psychosis as well as cognitive
disorganisation. In contrast, non-emotional face-processing was correlated with cognitive symptoms
only.
This chapter will begin by discussing the findings of these studies with reference to the
various models proposed to explain emotion-processing deficits in schizophrenia. It will then discuss
these findings in the context of the continuum model of psychosis, with particular attention to the role
of symptomatology. Finally, this chapter will discuss the broader implications of these studies,
limitations, and directions for future research.
Defining the deficit: Different models to explain emotion-processing
deficits in schizophrenia
The first aim of this thesis was to better characterise emotion-processing deficits in
schizophrenia. Specifically, to establish whether the facial emotion processing impairments
demonstrated in previous studies represent a true deficit in emotion-processing, or whether they are
better explained by other, overlapping areas of cognitive or visuoperceptual impairment. This section
will recap the different models proposed to explain emotion-processing in schizophrenia, then discuss
which of these models are best supported by the results of the current study.
A generalised cognitive deficit
The idea that a specific deficit in schizophrenia may actually be due to a more general deficit
is nothing new. A recurring controversy in schizophrenia research is the “generalised deficit
problem”: the concern that evidence of specific impairment on a certain task (such as face
recognition) may be better accounted for by a generalised cognitive deficit (such as poor attention;
Green, Horan & Sugar, 2013) . Many studies claim to capture specific deficits by demonstrating
impaired performance on one task (such as facial emotion recognition) but intact performance on
164
another task. However, it has been argued that the differential performance across two different tasks
may be the result of task parameters (such as difficulty or perceptual constraints) rather than a true
differential deficit (Green et al., 2013). In their 2013 commentary, Green and colleagues argue that the
literature does not support the presence of a generalised cognitive deficit in schizophrenia. Rather,
there is consistent evidence for sparing of certain cognitive domains such as speed of attentional
shifting, forms of implicit learning, emotional experience, and aspects of basic perceptual processing
(Gold, Hahn, Strauss, & Waltz, 2009; Horan, Foti, Hajcak, Wynn, & Green, 2012; Lee et al., 2013).
Furthermore, they point out that a truly generalised deficit would assume high levels of shared
variance across tasks sensitive to deficits, whereas many studies report negligible correlations
between deficits in different domains in schizophrenia (Gold et al., 2012). An opposing view is
presented by Gold and Dickinson (2013) who argue that, while there is some evidence of differential
deficits in schizophrenia, these studies are vastly outweighed by studies that report consistent, broad
deficit across all domains in schizophrenia (Green & Harvey, 2014; Mesholam-Gately, Giuliano,
Goff, Faraone, & Seidman, 2009). That is, while there is reason to believe that there are certain areas
of relatively spared performance in schizophrenia – such as selective attention in suppressing
irrelevant items on a working memory task (Gold et al., 2009) – this does not negate the finding that
patients consistently show broad, overall performance deficits compared to healthy controls (Gold &
Dickinson, 2013). A more appropriate approach is to examine whether emotion processing
impairments in schizophrenia can be entirely accounted for by the generalised performance deficit,
and in this respect the evidence is inconsistent. Two recent studies indicate that, although patients
with schizophrenia show reduced performance across tasks (indicating some degree of generalised
deficit) they nevertheless show disproportionately poor performance on facial emotion tasks
compared to matched tasks using non-face stimuli such as abstract patterns and line drawings
(Barkhof et al., 2015; Megreya, 2016). In contrast, other studies have reported comparable deficits on
both facial emotion tasks and tasks using non-face stimuli (Chen et al., 2009; Laprevote, Oliva,
Delerue, Thomas, & Boucart, 2010; Laprevote et al., 2013; Soria Bauser et al., 2012), supporting the
idea that emotion impairments in schizophrenia may be due to more general cognitive (or perceptual)
impairments.
A deficit in processing information from faces (a face -specific deficit)
Another possible explanation is that emotion-processing deficits in schizophrenia are actually
due to a more general perceptual impairment in processing faces that is unrelated to emotional
content. Chapter 3 presented a more in-depth review of studies demonstrating impairments in non-
emotional face processing in schizophrenia. To briefly recap, individuals with schizophrenia show
impairments in different aspects of basic face processing including face detection (McBain, Norton, &
Chen, 2010), extracting configural information from faces (Kim et al., 2010), discriminating age and
sex (Delerue et al., 2010; Schneider et al., 2006), and recognising the identity of a face (Bortolon,
165
Capdevielle, & Raffard, 2015). Many researchers investigating emotion-processing assume that non-
emotional face processing is unimpaired in schizophrenia, and therefore either exclude non-emotional
face tasks altogether, or include tasks that are poorly matched to the emotion task of interest (eg.: the
Benton Test of Facial Recognition). Of the small number of studies that have directly compared facial
identity processing and emotion processing in schizophrenia, results were predictably mixed. Several
studies demonstrated a disproportionate deficit in distinguishing facial emotions compared to
distinguishing facial identity (Kosmidis et al., 2007; Kucharska-Pietura et al., 2005; Penn et al., 2000;
Poole, Tobias, & Vinogradov, 2000; Schneider et al., 2006), while others showed that patients were
equally impaired on both domains (Addington & Addington, 1998; Pomarol-Clotet et al., 2010; Sachs
et al., 2004). One meta-analysis examined 28 studies comparing performance on an emotion task with
a non-emotional face control task in schizophrenia patients (Chan et al., 2010). It was found that,
while patients were significantly impaired on the control tasks with an overall effect size of -.70 (a
moderate-to-large effect), a greater impairment was seen on the facial emotion tasks with an effect
size of -.85 (a large effect). The authors concluded that, while there is evidence of a generalised
impairment in face processing, the processing of facial emotions is disproportionately affected in
schizophrenia. However, whether this indicates specific impairment in the perceptual or cognitive
processes required for recognising emotion, or simply reflects differences in task parameters (such as
greater difficulty) remains unclear.
A multi-modal deficit in processing emotional information (an emotion -specific
deficit)
Finally, a third explanation for facial emotion deficits in schizophrenia is that they represent a
multi-modal deficit in emotion processing which is not specific to faces. Although not as thoroughly
studied as the face literature, researchers have found evidence for impairments in recognising emotion
in other domains, such as affective prosody (i.e. emotion conveyed in the melody of speech).
Compared to healthy controls, individuals with schizophrenia were found to have greater difficulty
identifying the emotion portrayed in vocal recordings, regardless of the emotion being expressed
(Bozikas, Kosmidis, et al., 2006; Edwards et al., 2002; Edwards et al., 2001; Kucharska-Pietura et al.,
2005; Leitman et al., 2005; Matsumoto et al., 2006). Functional MRI investigations also revealed
abnormal activations in schizophrenia when listening to emotional prosody, including reduced
activations in the insula, superior temporal gyrus, and inferior frontal gyrus (Leitman et al., 2011;
Mitchell, Elliott, Barry, Cruttenden, & Woodruff, 2004). Similarly, two ERP studies have shown
atypical responses of the P50, N100 and P200 and waveforms in patients with schizophrenia during a
vocal emotion discrimination task, which were congruent with reduced accuracy (Pinheiro et al.,
2013; Pinheiro et al., 2014). Given that schizophrenia is associated with anatomical and functional
abnormalities of structures of the limbic system (see Chapter 2), it is possible that the
disproportionately large impairment in facial emotion recognition reported in the literature is driven
166
(at least in part) by dysfunction in higher-order emotion processing, rather than visuoperceptual
mechanisms such as face perception.
Results of the current study
Results from the inpatient studies (presented in Chapters 8 and 9) indicated that patients with
schizophrenia showed prominent impairment in both labelling and discriminating between dynamic
emotions but showed no (or negligible) impairment on two non-emotional face tasks which were
matched on task demands. Therefore, the results are consistent with previous studies which reported
that emotion-processing is disproportionately impaired compared to non-emotional face processing
(e.g.: Kosmidis et al., 2007; Kucharska-Pietura, David, Masiak, & Phillips, 2005; Penn et al., 2000;
Poole, Tobias, & Vinogradov, 2000; Schneider et al., 2006). However, inpatients with schizophrenia
also showed substantial deficits on a car discrimination task. This suggests that, while non-emotional
face processing may be (mostly) spared, there is clear evidence that the very task of discriminating
between two dynamic objects is reduced compared to healthy controls. Again, the finding that non-
face object recognition is just as impaired as emotion-processing is consistent with several previous
studies (e.g.: Laprevote et al., 2013; Soria Bauser et al., 2012), but to my knowledge this is the first
time this has been investigated using dynamic stimuli.
Taken together, these results are consistent with the idea that emotion-processing deficits are
simply the result of a general cognitive impairment, such as inattention. Performance on the non-
emotional face tasks may have been spared due to the fact that non-emotional face processing is
simply not equivalent to other visual tasks. Structural or non-emotional face information is extracted
from the environment through ‘holistic’ face processing (McKone & Robbins, 2011). Holistic face-
processing refers to a rapid, involuntary face-specific perceptual process that integrates information
across the face as a whole. It includes such information as the shapes of individual features, the
relative distances between them, and the contour of the cheeks and jaw (Maurer et al., 2002; McKone
& Yovel, 2009). Holistic processing operates for upright faces only, and is absent for objects and
inverted faces (Rossion, 2008). Importantly, this process is specific to invariant information, and is
therefore critical for perceiving identity (McKone & Robbins, 2011; also see Chapter 2). In contrast,
recognition of emotions depends on changeable face information (i.e.: transient facial movements)
which is not extracted through holistic processing. Studies suggest that holistic face processing is
intact in schizophrenia (Butler et al., 2008; Chambon et al., 2006; Schwartz et al., 2002), therefore it
is entirely possible that performance on the two non-emotional face tasks in the current studies was
facilitated through this rapid, specialised process. In contrast, performance on the two emotion tasks
and the non-face (car) task likely relied on more directed forms of voluntary attention, and as a result
may be more vulnerable to generalised cognitive impairment. Further research is necessary to explore
this hypothesis. Nevertheless, it can be concluded from the experiments described in this thesis that
167
the emotion-processing deficits typically associated with schizophrenia cannot be accounted for by
face-specific perceptual impairments. Rather, the evidence supports a generalised cognitive
impairment which is unrelated to either emotions or faces.
Results from the symptom correlate analyses also fit with the idea that impaired performance
on the emotion tasks and car task were due to a shared deficit. Both positive symptoms of psychosis
(e.g.: delusions) and cognitive symptoms (e.g.: poor attention) were associated with deficits in both
emotion-processing and non-face discrimination, as shown by the respective dynamic tasks. In
contrast, performance on the two non-emotional face tasks were unimpaired in the schizophrenia
group, and performance on these tasks was associated only with cognitive symptoms. Therefore, it
can be tentatively concluded that the “emotion-processing deficits” reported in this thesis, and in
previous literature, actually represent a generalised cognitive impairment which is particularly
associated with positive symptoms of psychosis.
Are these deficits specific to schizophrenia? Implications for
transdiagnostic models of psychiatric illness
The second aim of this thesis was to determine whether face-processing impairments are
specific to schizophrenia, and to what extent they are shared by other psychiatric disorders. The third
aim of this thesis was to examine associations between these deficits and specific symptoms in
inpatients with a range of psychiatric diagnoses. By characterising patterns of deficits across
psychiatric disorders, researchers can draw inferences about the nosological boundaries – or lack
thereof – between these disorders. This section will briefly review spectrum, dimensional, or
‘transdiagnostic’ theories of psychosis. It will then discuss the findings of the current studies with
relation to these theories.
Why should we consider dimensional models of illness?
Traditional conceptualisations of psychopathology, such as that found in the DSM-V, focuses
on defining distinct categories of illness in order to facilitate effective diagnosis, treatment planning,
and estimates of prognosis (Lawrie et al., 2010). Although this categorical approach dominates in both
clinical and research settings, evidence suggests that psychopathology does not naturally manifest in
such clearly defined categories. For instance, it has been shown that individual symptoms, such as
depression, mania, and psychosis, are better explained as dimensional phenomena that occur
frequently across different diagnoses as well as in subclinical forms (Wigman, de Vos, Wichers, van
Os, & Bartels-Velthuis, 2017). Even seemingly rare symptoms, such as the presence of delusions, are
commonly found across a wide range of disorders such as major depression (estimated 20% of
patients), bipolar disorder (up to 50%), borderline personality disorder (33%), obsessive-compulsive
disorder (14%) and even Alzheimer's disease (30%) (Bebbington & Freeman, 2017). As a result of
168
this heterogeneity, comorbidity rates in psychiatric disorders are understandably high, with almost
half of patients meeting criteria for two or more different disorders (Kessler et al., 2005). Overall, the
evidence suggests that conventional categories are not adequately capturing the expression of
different disorders. Could a dimensional approach to diagnosis overcome these problems? Some
authors argue that dimensional models do not add clinical utility (e.g. see Lawrie et al., 2010). Others
argue, however, that improving identification and treatment of comorbid symptoms could improve
patient outcomes (Kelleher & Cannon, 2016). For example, the presence of psychotic symptoms in
individuals with primary non-psychotic disorders is associated with greater suicidality and poorer
social, cognitive, and occupational outcomes (Kelleher et al., 2013; Kelleher et al., 2014; Wigman et
al., 2012). By the same token, specific treatment of non-psychotic symptoms in psychotic disorders
has been shown to improve patient outcome (de Bont et al., 2016), although more research is needed
to establish evidence-based treatment guidelines for common comorbidities, such as depression
(Upthegrove, Marwaha, & Birchwood, 2017). Consideration of spectrum models of symptoms may
well be the approach needed to develop transdiagnostic methods of treatment that focus on symptoms
in concert, rather than the symptoms of the primary diagnosis alone.
A spectrum of psychosis in the general population
As discussed in Chapter 6, psychosis-like experiences are reported not only in psychiatric
patients, but also in otherwise healthy individuals. Termed ‘schizotypy’, these experiences include a
range of milder or subclinical versions of symptoms such as delusions, hallucinations and social
anhedonia (Mason & Claridge, 2006). These psychosis-like experiences are shown to vary in intensity
across the general population, affecting an estimated 7% of adults (van Os, Linscott, Myin-Germeys,
Delespaul, & Krabbendam, 2009). It has been argued, therefore, that psychosis is not a discrete
phenomenon, but a continuum with formal schizophrenia representing the most extreme pole of a
spectrum of normal experience (Claridge, 1997; DeRosse & Karlsgodt, 2015). While approximately
8% of high-schizotypy individuals will convert to a formal psychotic diagnosis in the following years,
the majority of individuals will not go on to meet diagnostic criteria (Hanssen et al., 2005). Therefore,
although schizotypy serves as a prodrome to formal diagnosis for some, this is not accurate in the
majority of cases (Kelleher & Cannon, 2011). Importantly, high levels of schizotypy are associated
with other areas of deficit commonly seen in schizophrenia, such as mildly impaired emotion-
processing, neurophysiological abnormalities, and cognitive deficit (Batty et al., 2014; Germine &
Hooker, 2011; Korponay, Nitzburg, Malhotra, & DeRosse, 2014). This suggests that formal psychotic
disorders and trait schizotypy may share a common biological basis, further supporting the continuum
model.
The experiment described in Chapter 6 found that schizotypy scores in healthy controls were
negatively correlated with the ability to correctly label emotions on a dynamic task. Interestingly, the
169
‘Unusual Experiences’ subscale was associated with impaired performance, but not the other
subscales (‘Cognitive Disorganisation’, ‘Introvertive Anhedonia’ and ‘Impulsive Nonconformity’).
This finding is in line with previous research indicating associations between emotion-processing
deficits and ‘positive’ schizotypy (Abbott & Byrne, 2013; Kerns, 2005; Shean et al., 2007; van 't
Wout et al., 2004). However, to my knowledge this is the first study to report this result using
dynamic tasks. These findings support the notion that the same pattern of deficits seen in clinical
psychosis are also expressed to a milder degree in schizotypy, suggesting that they may represent
different severities of the same construct. Broadly speaking, these findings are consistent with the idea
that psychosis is best represented as a continuum of experience.
The bipolar-schizophrenia spectrum model
Traditionally, schizophrenia spectrum disorders and bipolar disorder have been
conceptualised as discrete nosological entities; a distinction that is sometimes referred to as the
‘Kraepelinian dichotomy’ (Craddock & Owen, 2010). However, many patients present with a
combination of prominent affective and psychotic symptoms, making it difficult to differentiate
between schizophrenia and bipolar disorder in practice (Dacquino et al., 2015; Malaspina et al., 2013).
These disorders overlap not only in the expression of symptoms, but aetiologically as well. For
instance, a substantial body of evidence point towards shared genetic risk factors for both bipolar
disorder and schizophrenia (Cardno, Rijsdijk, Sham, Murray, & McGuffin, 2002; Houenou et al.,
2017; Lichtenstein et al., 2009; The International Schizophrenia Consortium, 2009). Both disorders
respond similarly to antipsychotic treatments (Hill et al., 2013), and share environmental risk factors
such as urbanicity, childhood adversity, substance abuse, and prenatal insult (Clarke, Harley, &
Cannon, 2006; Heinz, Deserno, & Reininghaus, 2013; Lichtenstein et al., 2009; Matheson, Shepherd,
Pinchbeck, Laurens, & Carr, 2013).
A large study by the Bipolar and Schizophrenia Network for Intermediate Phenotypes
(BSNIP) consortium were conducted to identify phenotypic markers in bipolar disorder and
schizophrenia. This included over 2000 patients, their first-degree relatives, and healthy controls who
completed a battery of tests investigating social functioning, cognition, brain volumetry, and a variety
of EEG and fMRI assessments. Surprisingly, almost every one of the biomarkers tested overlapped so
strongly that they were unable to discriminate between bipolar and schizophrenia groups (Tamminga
et al., 2014). The only biomarker that differed substantially between diagnoses was grey matter
volume. While schizophrenia and schizoaffective patients showed widespread grey matter reduction
across the cortices, bipolar patients showed only minor reductions in the prefrontal and limbic
regions. Taken together, the biomarker evidence fits with genetic and phenomenological studies
which indicate that schizophrenia and bipolar disorder overlap substantially and may even fall upon a
spectrum of disease without clearly delineated boundaries.
170
Studies examining cognitive deficits across different diagnoses also appear to support a
spectrum model. Cognitive deficits in schizophrenia are present at or prior to the first episode of
psychosis, are largely unaffected by treatment, and are predictive of poorer functional outcome (Hill
et al., 2013; Keefe & Harvey, 2012). Although bipolar disorder is typically associated with less
profound cognitive deficits, these deficits are seemingly stable over time and are worse in patients
who have experienced symptoms of psychosis (Bora, Yucel, & Pantelis, 2010; Burdick, Goldberg,
Harrow, Faull, & Malhotra, 2006; Glahn et al., 2007; Hill et al., 2009). Schizoaffective disorder,
which some consider an intermediary diagnosis between bipolar disorder and schizophrenia, is
similarly associated with cognitive deficits which appear to fall in between these two groups (Hill et
al., 2013; Hooper et al., 2010). One large study investigating cognition across a range of psychotic
disorders (n>650) found that deficits were largest in the schizophrenia group, intermediate in
schizoaffective disorder, and smallest (but still significant) in psychotic bipolar disorder (Hill et al.,
2013). These findings fit with the idea that bipolar disorder and schizophrenia may be conceptualised
as opposing poles on a spectrum, with schizoaffective disorder in the middle. Rather than distinct
entities, these disorders vary along two factors: severity of affective symptoms, and severity of
psychotic symptoms.
The schizophrenia-bipolar spectrum model is supported by evidence indicating substantial
overlap in symptomatology, genetic and environmental risks, biomarkers and neurocognitive deficits
between these disorders. However, this model does not take into account the substantial overlap
between psychotic disorders and other types of disorders, such as unipolar depression, anxiety, autism
and even intellectual disability (van Os & Reininghaus, 2016). For instance, evidence suggests that
symptoms of psychosis or psychosis-like experiences are not randomly distributed among the general
population, but are twice as likely to occur in individuals with a diagnosis of anxiety or depression
(Wigman et al., 2012). In fact, symptoms of psychosis are predicted not only by psychotic diagnoses,
but also by all mood and anxiety disorders, eating disorders, ADHD, and alcohol abuse (McGrath et
al., 2016). There is also evidence of similar biomarkers in schizophrenia and depression, including
white matter changes, circulating inflammatory markers, and limbic system abnormalities
(Upthegrove et al., 2017). In recognition of these shared factors, Reininghaus and colleagues (2016)
have suggested a bi-factor model consisting of a general psychosis dimension, which underlies both
affective and non-affective psychotic symptoms across disorders, and five specific psychosis
dimensions (see Figure 12.1). The specific factors include positive symptoms, negative symptoms,
disorganisation, mania and depression. Unlike the simpler schizophrenia-bipolar spectrum, this six-
dimensional model accounts for both shared and specific features of different disorders.
171
Figure 12.1. Pentagonal bifactor model of psychosis. Standardised mean factor scores on each
dimension are shown for four different diagnostic categories: bipolar disorder with mania, bipolar
disorder with hypomania, schizoaffective disorder and schizophrenia. Standardised factor scores have
a mean of 0 and standard deviation of 1. Image from Reininghaus and colleagues (2016;
supplementary materials).
The results of the experiments outlined in this thesis provide further support for the idea that
schizophrenia and bipolar disorder may exist on a spectrum. The schizophrenia and bipolar groups
performed similarly across all tasks, and were significantly different from controls on two emotion-
processing tasks and a car discrimination task. This suggests that, in line with previous studies
investigating face processing using static stimuli (Bellack et al., 1996; Derntl et al., 2012; Edwards et
al., 2001), both schizophrenia and bipolar patients have similar deficits in both emotion-processing
and non-face discrimination. Correlational analyses also indicated that performance on these tasks
was positively associated with positive symptoms and cognitive symptoms. As discussed in the
172
previous section, the performance on these tasks may represent a generalised cognitive deficit which
is most severe in patients with psychosis but negligible in patients with non-psychotic disorders.
Where does substance-induced psychosis fit in?
The previous sections have explored the concept of psychosis as a dimensional phenomenon
that varies in intensity and frequency across the general population and transcends diagnostic
categories. However, our understanding of psychosis is further complicated by a comparatively under-
studied form of psychotic disorder: substance-induced psychosis. The incidence rate of substance-
induced psychosis is estimated at 5 per 100,000 people per year (Chen, Hsieh, Chang, Hung, & Chan,
2015). A percentage of these individuals will progress to a formal diagnosis of schizophrenia in the
years following, however this number varies according to the substance implicated. For instance, 5-
11% of patients with alcohol-induced psychosis will develop schizophrenia within 8 years, compared
to approximately 20-30% of patients with amphetamine-induced psychosis and 46% of patients
cannabis-induced psychosis (W. L. Chen et al., 2015; Niemi-Pynttari et al., 2013; Starzer et al., 2018).
This raises the question: is substance-induced psychosis a discrete disorder, or merely a prodrome to
schizophrenia?
The relationship between drug use and psychosis is complex. Studies suggest that individuals
with psychotic disorders are at a greatly increased risk of developing a substance use disorder, with a
lifetime prevalence of 48-60% (Cantor-Graae, Nordstrom, & McNeil, 2001; Swartz et al., 2006), and
comorbid substance abuse is associated with poorer remission rates, more severe symptoms and
treatment non-compliance (Lambert et al., 2008; Malla et al., 2008; Miller, 2008). Animal and human
studies suggest that schizophrenia and substance use disorders share a neurobiological basis, possibly
relating to abnormalities in striatal dopamine levels (Chambers, Sentir, & Engleman, 2010; Thompson
et al., 2013).
Clinically, the acute presentation of substance-induced psychosis is indistinguishable from
primary psychosis (Bramness et al., 2012). Generally speaking, substance-induced psychosis is
thought to involve similar neurochemical pathways to schizophrenia (Paparelli et al., 2011), however
the exact mechanism is unclear. For example, repeated amphetamine abuse may precipitate psychosis
through acute effects on CNS dopamine levels, but indirect effects such as sleep deprivation may also
play a role (Murray, Paparelli, Morrison, Marconi, & Di Forti, 2013). Additionally, the long-term
neurotoxic effects of amphetamines may reduce the threshold for developing psychosis over time
(Bramness et al., 2012). Cannabis-induced psychosis may also precipitate psychosis by altering
dopamine levels. Studies suggest that cannabis use affects striatal dopamine, particularly in
individuals with a familial history of schizophrenia, likely indicating a shared genetic vulnerability to
substance-induced psychosis (Kuepper et al., 2013).
173
Should substance-induced psychosis be considered a separate diagnostic entity to
schizophrenia? Despite high rates of transformation to schizophrenia later in life, the majority of
patients with substance-induced psychosis will never meet criteria for a primary psychosis (W. L.
Chen et al., 2015; Niemi-Pynttari et al., 2013). Clearly there are other factors involved that determine
whether an individual converts to primary psychosis or remits altogether. Bramness and colleagues
(2012) argue that substance-induced psychosis and schizophrenia can be reconciled using the
traditional stress-diathesis model (See Figure 12.2). In this model, the amount of exposure to a
substance (such as amphetamines) sufficient to precipitate psychosis depends on an individual’s level
of genetic vulnerability. That is, an individual with low vulnerability may use large amounts of
amphetamines without becoming psychotic (substance use disorder without psychosis), while
someone with high vulnerability develops primary psychosis without any exposure to drugs
(schizophrenia). Those who develop substance-induced psychosis therefore have a moderate genetic
vulnerability to psychosis which is precipitated by drug use, and whether they transition to
schizophrenia may depend on both genetic vulnerability and severity of repeated amphetamine use.
This model of schizophrenia is also consistent with the ‘continuum model of psychosis’ discussed in a
previous section, which posits that formal psychotic disorders represent the most extreme end of a
continuum of psychotic-like traits distributed throughout the general population (Mason & Claridge,
2006).
Figure 12.2. A stress-diathesis model demonstrating the interaction between exposure and genetic
vulnerability to psychosis. From Bramness and colleagues (2012, p.4).
The results of the experiments presented in this thesis also provide some indirect support for
the idea that substance-induced psychosis sits in an intermediate position between psychotic disorders
and unaffected individuals. Although substance-induced psychosis was not examined in isolation, the
174
‘non-schizophrenia psychosis’ group was comprised of 53% substance-induced psychosis patients,
with the remainder mostly depression with psychosis. Results showed that performance on the
emotion tasks was mildly impaired for this group, with accuracy falling in between the healthy
controls and the more substantially impaired schizophrenia and bipolar groups. This suggests that this
group of non-schizophrenia psychotic disorders may experience a less severe form of emotion-
processing deficit which is positively associated with the severity of psychotic symptoms.
Methodological and practical implications of this research
Teasing apart the role of face and emotion-processing deficits in schizophrenia may seem, to
some, to be an exercise in academic pedantry. However, this research has important practical
implications. Deficits in facial emotion-processing have real-world consequences, as they are strongly
associated with poorer social, occupational, and community functioning in patients with schizophrenia
(Irani et al., 2012; Meyer & Kurtz, 2009; Tsunoda et al., 2012). Furthermore, these deficits may
predict conversion to schizophrenia in healthy individuals who are at clinically-high risk of psychotic
illness (Corcoran et al., 2015). Tasks that effectively assess these deficits, therefore, may have clinical
utility in identifying patients who are at greater risk of poorer functioning, and may even play a role in
predicting who will develop schizophrenia. Tools that can increase timely diagnosis of psychosis are
especially useful, as the duration of untreated psychosis prior to diagnosis is associated with poorer
outcome, and more severe cognitive and negative symptoms following treatment (Chang et al., 2013).
Another way in which emotion-processing tasks may benefit patients is through remediation.
Several studies have trialled programs designed to improve emotion recognition ability in patients
with schizophrenia. Although some authors reported no improvement with training (Bechi et al.,
2012) other studies have demonstrated emotion-processing improvements following training which
are accompanied by increased activation in brain regions relevant to recognising emotions (Habel et
al., 2010; Hooker et al., 2013; Hooker et al., 2012; Ramsay & MacDonald, 2015). Whether these
improvements translate into improved real-life functioning, however, remains to be seen.
If emotion-processing tasks are to be useful for either assessment or remediation, it is
important that they be sensitive to the deficit of interest. This thesis uncovered several methodological
issues that may have implications for future research. First, it was found that dynamic stimuli are
more accurately recognised and are also more sensitive to positive schizotypy (psychosis-like
experiences) in healthy participants compared to traditional static stimuli. This finding adds further
weight to a growing body of evidence that dynamic emotion tasks are more ecologically valid, less
prone to ceiling effects, and more sensitive to subtle impairments than static tasks (Davis & Gibson,
2000; Harms et al., 2010). Taken together, there are clear incentives for researchers to choose to
175
employ dynamic tasks over traditional static measures, especially considering the increasing
availability of video-based emotion stimuli online (e.g.: https://mmifacedb.eu/).
An additional finding was that the lesser-utilised emotion discrimination paradigm (i.e.:same-
or-different?) may in fact be a better measure of motion-sensitive emotional information than the
more popular emotion labelling paradigm. Although this incidental discovery has not yet been
replicated, it would be interesting to explore whether these paradigms also produce differing patterns
of neural activity, and whether this has further implications for the use of one paradigm over another.
A third finding from this thesis was that sex-labelling tasks are likely not an appropriate
measure of facial identity processing. Deficits in identity-processing are less well-studied than
emotion-processing deficits in schizophrenia, however many authors assume that non-emotional face
processing is intact in this disorder (see Chapter 3, however, for a review of studies that suggest the
opposite). One problem with this area of research is that the varying paradigms used to assess
identity-processing deficits may not be measuring the same construct. This thesis revealed that a
dynamic sex-labelling task (i.e., is this morphed person male or female?) was entirely insensitive to
patient symptomatology, whereas a dynamic identity discrimination task (same-or-different?) was
associated with a variety of patient symptoms. This finding is in line with three previous studies
which reported that sex-labelling tasks were not correlated with deficits in patients with schizophrenia
(Bediou et al., 2007; Bediou, Krolak-Salmon, et al., 2005; van 't Wout et al., 2007). It is possible that
making decisions about the sex of a face does not involve the same neural processes as recognising
the entire identity of a face. Alternatively, the same-or-different discrimination paradigm may simply
be more sensitive to cognitive deficits in schizophrenia. Regardless, it is clear that sex-labelling tasks
are not appropriate for investigating perceptual deficits in schizophrenia, regardless of whether stimuli
are dynamic or static.
Study limitations
As was touched upon in previous chapters, the research presented in this thesis has several
limitations. First, despite attempts to match participant demographics in the main inpatient study, the
schizophrenia group had significantly fewer years of education, and lower IQ estimates compared to
healthy controls. Moreover, almost all patients were receiving psychotropic medication, which may
have impacted on their performance. Although our analyses (see Chapter 7) suggest that these factors
alone cannot account for the group differences reported in this study, their influence cannot be
definitively excluded.
An additional limitation is group size. Due to time constraint and the challenges surrounding
the recruitment of acutely unwell psychiatric inpatients, group sizes were uneven and relatively small
176
(n<20 for the three non-schizophrenia patient groups). This reduced the power to detect significant
differences in performance between these comparison groups. Given that significant differences were
still found, however, this limitation is unlikely to have had a major impact on the interpretation of
these results.
Furthermore, while our results support the idea of a generalised cognitive deficit, the inpatient
studies did not include any direct measures of cognition. These were excluded from the dynamic
battery in order to keep testing sessions within a 2-hour time frame. However, future studies would
benefit from including standardised cognitive measures, such as tests of basic attention or processing
speed, to evaluate whether specific domains of cognition correlate with performance on these
dynamic tasks.
The inpatient experiments reaffirmed the reports of deficits in face processing previously
demonstrated using static images using more ecologically-valid dynamic stimuli (Johnston et al.,
2010). However, it is important to acknowledge that these stimuli are still a distant approximation of
everyday social interactions. In future, tasks with even greater ecological validity, such as immersive
virtual environments, may be an advance in examining these mechanisms further.
Conclusions and future directions
The findings of this thesis suggest that, contrary to widespread belief, there is no evidence for
an emotion-specific face processing impairment in schizophrenia. Rather, the results point towards a
generalised cognitive deficit that is found across tasks assessing both emotion-processing and non-
face object processing. Correlational analyses revealed that both positive symptoms (such as
hallucinations and delusions) and cognitive symptoms (such as inattention) were associated with
deficits on these tasks. Moreover, these impairments were not specific to schizophrenia patients, but
also found in bipolar patients and, to a lesser degree, in patients with non-schizophrenia psychosis.
Patients with non-psychotic disorders, in contrast, were unimpaired. These findings support disease
models which conceptualise psychosis as a transdiagnostic spectrum rather than a series of categorical
disorders. Further support for dimensional models comes from a study of healthy controls, which
found that the positive dimension of schizotypy (e.g.: unusual perceptual experiences) was associated
with poorer performance on emotion-processing tasks. This suggest that the deficits associated with
clinical psychosis are present in an attenuated form in healthy individuals who report psychosis-like
experiences.
This was the first time that emotion-processing, identity-processing, and non-face processing
were examined in inpatients using a newly developed set of matched dynamic tasks. These studies
added to literature demonstrating the increased effectiveness of using dynamic stimuli over traditional
177
static measures. However, future research is needed to explore the relationship between cognitive
deficits – such as those seen in schizophrenia – and impairments in processing dynamic emotional and
non-emotional stimuli. It would also be useful to explore the neural correlates of these deficits in
patients, particularly as the correlates of dynamic emotion-processing are shown to differ from those
shown during static emotion-processing (Arsalidou et al., 2011; Sato et al., 2004). This thesis also
highlights the need for researchers to focus on populations with psychotic symptoms outside of
schizophrenia, such as patients with substance-induced psychosis or depression with psychotic
features. This would shed additional light on the specific cognitive domains and brain pathways
affected by psychosis, allowing researchers to separate the impact of psychotic features from the more
severe, global sequelae seen in those diagnosed with schizophrenia. Furthermore, characterisation of
these patterns of deficits, including the identification of accompanying genetic and neuroimaging
markers, is necessary for creating a comprehensive model of psychotic disorders. This could
eventually inform the development of better interventions to ameliorate the devastating functional
disability experienced by sufferers of these disorders.
178
References
Abbott, G. R., & Byrne, L. K. (2013). Schizotypal traits are associated with poorer identification of
emotions from dynamic stimuli. Psychiatry Res, 207(1-2), 40-44.
doi:10.1016/j.psychres.2013.03.005
Abbott, G. R., & Green, M. J. (2013). Facial affect recognition and schizotypal personality
characteristics. Early Interv Psychiatry, 7(1), 58-63. doi:10.1111/j.1751-7893.2012.00346.x
Abrosoft. (2012). Fantamorph 5 [Computer software]. Retrieved from http://www.fantamorph.com/.
Adams, R. B., Jr., Gordon, H. L., Baird, A. A., Ambady, N., & Kleck, R. E. (2003). Effects of gaze on
amygdala sensitivity to anger and fear faces. Science, 300(5625), 1536.
doi:10.1126/science.1082244
Addington, J., & Addington, D. (1998). Facial affect recognition and information processing in
schizophrenia and bipolar disorder. Schizophr Res, 32(3), 171-181.
Addington, J., Saeedi, H., & Addington, D. (2006). Facial affect recognition: a mediator between
cognitive and social functioning in psychosis? Schizophr Res, 85(1-3), 142-150.
doi:10.1016/j.schres.2006.03.028
Adolphs, R., Gosselin, F., Buchanan, T. W., Tranel, D., Schyns, P., & Damasio, A. R. (2005). A
mechanism for impaired fear recognition after amygdala damage. Nature, 433(7021), 68-72.
doi:10.1038/nature03086
Adolphs, R., Tranel, D., & Damasio, A. R. (2003). Dissociable neural systems for recognizing
emotions. Brain Cogn, 52(1), 61-69.
Aftanas, L., Varlamov, A., Pavlov, S., Makhnev, V., & Reva, N. (2001). Event-related
synchronization and desynchronization during affective processing: emergence of valence-
related time-dependent hemispheric asymmetries in theta and upper alpha band. Int J
Neurosci, 110(3-4), 197-219.
Almeida, J. R., Versace, A., Mechelli, A., Hassel, S., Quevedo, K., Kupfer, D. J., & Phillips, M. L.
(2009). Abnormal amygdala-prefrontal effective connectivity to happy faces differentiates
bipolar from major depression. Biol Psychiatry, 66(5), 451-459.
doi:10.1016/j.biopsych.2009.03.024
Alves, N. T. (2013). Recognition of static and dynamic facial expressions: a study review. Estudos de
Psicologia (Natal), 18(1), 125-130.
Ambadar, Z., Schooler, J. W., & Cohn, J. F. (2005). Deciphering the enigmatic face: the importance
of facial dynamics in interpreting subtle facial expressions. Psychol Sci, 16(5), 403-410.
doi:10.1111/j.0956-7976.2005.01548.x
Amminger, G. P., Henry, L. P., Harrigan, S. M., Harris, M. G., Alvarez-Jimenez, M., Herrman, H., . .
. McGorry, P. D. (2011). Outcome in early-onset schizophrenia revisited: findings from the
Early Psychosis Prevention and Intervention Centre long-term follow-up study. Schizophr
Res, 131(1-3), 112-119. doi:10.1016/j.schres.2011.06.009
Anderson, I. M., Shippen, C., Juhasz, G., Chase, D., Thomas, E., Downey, D., . . . Deakin, J. F.
(2011). State-dependent alteration in face emotion recognition in depression. Br J Psychiatry,
198(4), 302-308. doi:10.1192/bjp.bp.110.078139
Andreasen, N. C. (1984a). Scale for the Assessment of Negative Symptoms: SANS: Department of
Psychiatry, College of Medicine, the University of Iowa.
Andreasen, N. C. (1984b). Scale for the Assessment of Positive Symptoms (SAPS): University of Iowa.
Andrews, S., Jenkins, R., Cursiter, H., & Burton, A. M. (2015). Telling faces together: Learning new
faces through exposure to multiple instances. Q J Exp Psychol (Hove), 68(10), 2041-2050.
doi:10.1080/17470218.2014.1003949
Anticevic, A., Van Snellenberg, J. X., Cohen, R. E., Repovs, G., Dowd, E. C., & Barch, D. M. (2012).
Amygdala recruitment in schizophrenia in response to aversive emotional material: a meta-
analysis of neuroimaging studies. Schizophr Bull, 38(3), 608-621. doi:10.1093/schbul/sbq131
Archer, J., Hay, D. C., & Young, A. W. (1994). Movement, face processing and schizophrenia:
evidence of a differential deficit in expression analysis. Br J Clin Psychol, 33 ( Pt 4), 517-
528.
179
Arsalidou, M., Morris, D., & Taylor, M. J. (2011). Converging evidence for the advantage of dynamic
facial expressions. Brain Topogr, 24(2), 149-163. doi:10.1007/s10548-011-0171-4
Asthana, H. S., Mandal, M. K., Khurana, H., & Haque-Nizamie, S. (1998). Visuospatial and affect
recognition deficit in depression. J Affect Disord, 48(1), 57-62.
Atkinson, A. P., Dittrich, W. H., Gemmell, A. J., & Young, A. W. (2004). Emotion perception from
dynamic and static body expressions in point-light and full-light displays. Perception, 33(6),
717-746. doi:10.1068/p5096
Australian Institute of Health and Welfare. (2015). Health expenditure Australia 2013–14. Canberra:
AIHW.
Autodesk Inc. (2012). 3DS Max Design [Computer software]. Retrieved from
http://www.autodesk.com.au/products/3ds-max/overview.
Avons, S. E., Nunn, J. A., Chan, L., & Armstrong, H. (2003). Executive function assessed by memory
updating and random generation in schizotypal individuals. Psychiatry Res, 120(2), 145-154.
Baiano, M., David, A., Versace, A., Churchill, R., Balestrieri, M., & Brambilla, P. (2007). Anterior
cingulate volumes in schizophrenia: a systematic review and a meta-analysis of MRI studies.
Schizophr Res, 93(1-3), 1-12. doi:10.1016/j.schres.2007.02.012
Balconi, M., & Pozzoli, U. (2008). Event-related oscillations (ERO) and event-related potentials
(ERP) in emotional face recognition. Int J Neurosci, 118(10), 1412-1424.
doi:10.1080/00207450601047119
Barkhof, E., de Sonneville, L. M., Meijer, C. J., & de Haan, L. (2015). Processing of facial and
nonsocial information is differentially associated with severity of symptoms in patients with
multiepisode schizophrenia. J Nerv Ment Dis, 203(2), 112-119.
doi:10.1097/NMD.0000000000000246
Basso, M. R., Schefft, B. K., Ris, M. D., & Dember, W. N. (1996). Mood and global-local visual
processing. J Int Neuropsychol Soc, 2(3), 249-255.
Batty, R. A., Francis, A. J., Innes-Brown, H., Joshua, N. R., & Rossell, S. L. (2014).
Neurophysiological correlates of configural face processing in schizotypy. Front Psychiatry,
5, 101. doi:10.3389/fpsyt.2014.00101
Baudouin, J. Y., & Humphreys, G. W. (2006). Compensatory strategies in processing facial emotions:
evidence from prosopagnosia. Neuropsychologia, 44(8), 1361-1369.
doi:10.1016/j.neuropsychologia.2006.01.006
Baudouin, J. Y., Martin, F., Tiberghien, G., Verlut, I., & Franck, N. (2002). Selective attention to
facial emotion and identity in schizophrenia. Neuropsychologia, 40(5), 503-511.
Beaupré, M. G., & Hess, U. (2005). Cross-cultural emotion recognition among Canadian ethnic
groups. Journal of cross-cultural psychology, 36(3), 355-370.
Bebbington, P., & Freeman, D. (2017). Transdiagnostic Extension of Delusions: Schizophrenia and
Beyond. Schizophr Bull, 43(2), 273-282. doi:10.1093/schbul/sbw191
Becerril, K., & Barch, D. (2011). Influence of emotional processing on working memory in
schizophrenia. Schizophr Bull, 37(5), 1027-1038. doi:10.1093/schbul/sbq009
Bechara, A., Damasio, H., & Damasio, A. R. (2000). Emotion, decision making and the orbitofrontal
cortex. Cereb Cortex, 10(3), 295-307.
Bechi, M., Riccaboni, R., Ali, S., Fresi, F., Buonocore, M., Bosia, M., . . . Cavallaro, R. (2012).
Theory of mind and emotion processing training for patients with schizophrenia: preliminary
findings. Psychiatry Res, 198(3), 371-377. doi:10.1016/j.psychres.2012.02.004
Bediou, B., Asri, F., Brunelin, J., Krolak-Salmon, P., D'Amato, T., Saoud, M., & Tazi, I. (2007).
Emotion recognition and genetic vulnerability to schizophrenia. Br J Psychiatry, 191, 126-
130. doi:10.1192/bjp.bp.106.028829
Bediou, B., Brunelin, J., d'Amato, T., Fecteau, S., Saoud, M., Henaff, M. A., & Krolak-Salmon, P.
(2012). A comparison of facial emotion processing in neurological and psychiatric conditions.
Front Psychol, 3, 98. doi:10.3389/fpsyg.2012.00098
Bediou, B., Franck, N., Saoud, M., Baudouin, J. Y., Tiberghien, G., Dalery, J., & d'Amato, T. (2005).
Effects of emotion and identity on facial affect processing in schizophrenia. Psychiatry Res,
133(2-3), 149-157. doi:10.1016/j.psychres.2004.08.008
180
Bediou, B., Krolak-Salmon, P., Saoud, M., Henaff, M. A., Burt, M., Dalery, J., & D'Amato, T. (2005).
Facial expression and sex recognition in schizophrenia and depression. Can J Psychiatry,
50(9), 525-533.
Beedie, S. A., Benson, P. J., & St Clair, D. M. (2011). Atypical scanpaths in schizophrenia: evidence
of a trait- or state-dependent phenomenon? J Psychiatry Neurosci, 36(3), 150-164.
doi:10.1503/jpn.090169
Behere, R. V., Venkatasubramanian, G., Arasappa, R., Reddy, N. N., & Gangadhar, B. N. (2011).
First rank symptoms & facial emotion recognition deficits in antipsychotic naive
schizophrenia: Implications for social threat perception model. Prog Neuropsychopharmacol
Biol Psychiatry, 35(7), 1653-1658. doi:10.1016/j.pnpbp.2011.05.019
Bellack, A. S., Blanchard, J. J., & Mueser, K. T. (1996). Cue availability and affect perception in
schizophrenia. Schizophr Bull, 22(3), 535-544.
Benes, F. M. (1998). Model generation and testing to probe neural circuitry in the cingulate cortex of
postmortem schizophrenic brain. Schizophr Bull, 24(2), 219-230.
Benson, P. J., Leonards, U., Lothian, R. M., St Clair, D. M., & Merlo, M. C. (2007). Visual scan paths
in first-episode schizophrenia and cannabis-induced psychosis. J Psychiatry Neurosci, 32(4),
267-274.
Bentin, S., Allison, T., Puce, A., Perez, E., & McCarthy, G. (1996). Electrophysiological Studies of
Face Perception in Humans. J Cogn Neurosci, 8(6), 551-565. doi:10.1162/jocn.1996.8.6.551
Benton, A. L. (1983). Contributions to Neuropsychological Assessment: A Clinical Manual. New
York: Oxford University Press.
Bestelmeyer, P. E., Tatler, B. W., Phillips, L. H., Fraser, G., Benson, P. J., & St Clair, D. (2006).
Global visual scanning abnormalities in schizophrenia and bipolar disorder. Schizophr Res,
87(1-3), 212-222. doi:10.1016/j.schres.2006.06.015
Biehl, M., Matsumoto, D., Ekman, P., Hearn, V., Heider, K., Kudoh, T., & Ton, V. (1997).
Matsumoto and Ekman's Japanese and Caucasian Facial Expressions of Emotion (JACFEE):
Reliability data and cross-national differences. Journal of Nonverbal behavior, 21(1), 3-21.
Biele, C., & Grabowska, A. (2006). Sex differences in perception of emotion intensity in dynamic and
static facial expressions. Experimental Brain Research, 171(1), 1-6.
Bigelow, N. O., Paradiso, S., Adolphs, R., Moser, D. J., Arndt, S., Heberlein, A., . . . Andreasen, N. C.
(2006). Perception of socially relevant stimuli in schizophrenia. Schizophr Res, 83(2-3), 257-
267. doi:10.1016/j.schres.2005.12.856
Bimler, D. L., Skwarek, S. J., & Paramei, G. V. (2013). Processing Facial Expressions of Emotion:
Upright vs. Inverted Images. Front Psychol, 4, 54. doi:10.3389/fpsyg.2013.00054
Blanchard, J. J., & Cohen, A. S. (2006). The structure of negative symptoms within schizophrenia:
implications for assessment. Schizophr Bull, 32(2), 238-245. doi:10.1093/schbul/sbj013
Bora, E., Yucel, M., & Pantelis, C. (2010). Cognitive impairment in affective psychoses: a meta-
analysis. Schizophr Bull, 36(1), 112-125. doi:10.1093/schbul/sbp093
Bortolon, C., Capdevielle, D., & Raffard, S. (2015). Face recognition in schizophrenia disorder: A
comprehensive review of behavioral, neuroimaging and neurophysiological studies. Neurosci
Biobehav Rev, 53, 79-107. doi:10.1016/j.neubiorev.2015.03.006
Bortolon, C., Capdevielle, D., Salesse, R. N., & Raffard, S. (2016). Further insight into self-face
recognition in schizophrenia patients: Why ambiguity matters. J Behav Ther Exp Psychiatry,
50, 215-222. doi:10.1016/j.jbtep.2015.09.006
Bouhuys, A. L., Geerts, E., & Gordijn, M. C. (1999). Depressed patients' perceptions of facial
emotions in depressed and remitted states are associated with relapse: a longitudinal study. J
Nerv Ment Dis, 187(10), 595-602.
Bouhuys, A. L., Geerts, E., Mersch, P. P., & Jenner, J. A. (1996). Nonverbal interpersonal sensitivity
and persistence of depression: perception of emotions in schematic faces. Psychiatry Res,
64(3), 193-203.
Bourke, C., Douglas, K., & Porter, R. (2010). Processing of facial emotion expression in major
depression: a review. Aust N Z J Psychiatry, 44(8), 681-696.
doi:10.3109/00048674.2010.496359
181
Boyer, P., Phillips, J. L., Rousseau, F. L., & Ilivitsky, S. (2007). Hippocampal abnormalities and
memory deficits: new evidence of a strong pathophysiological link in schizophrenia. Brain
Res Rev, 54(1), 92-112. doi:10.1016/j.brainresrev.2006.12.008
Bozikas, V. P., Kosmidis, M. H., Anezoulaki, D., Giannakou, M., Andreou, C., & Karavatos, A.
(2006). Impaired perception of affective prosody in schizophrenia. J Neuropsychiatry Clin
Neurosci, 18(1), 81-85. doi:10.1176/jnp.18.1.81
Bozikas, V. P., Tonia, T., Fokas, K., Karavatos, A., & Kosmidis, M. H. (2006). Impaired emotion
processing in remitted patients with bipolar disorder. J Affect Disord, 91(1), 53-56.
doi:10.1016/j.jad.2005.11.013
Bramness, J. G., Gundersen, O. H., Guterstam, J., Rognli, E. B., Konstenius, M., Loberg, E. M., . . .
Franck, J. (2012). Amphetamine-induced psychosis--a separate diagnostic entity or primary
psychosis triggered in the vulnerable? BMC Psychiatry, 12, 221. doi:10.1186/1471-244X-12-
221
Brenner, C. A., Rumak, S. P., & Burns, A. M. (2016). Facial emotion memory in schizophrenia: From
encoding to maintenance-related EEG. Clin Neurophysiol, 127(2), 1366-1373.
doi:10.1016/j.clinph.2015.10.061
Brenner, C. A., Wilt, M. A., Lysaker, P. H., Koyfman, A., & O'Donnell, B. F. (2003).
Psychometrically matched visual-processing tasks in schizophrenia spectrum disorders. J
Abnorm Psychol, 112(1), 28-37.
Brotman, M. A., Skup, M., Rich, B. A., Blair, K. S., Pine, D. S., Blair, J. R., & Leibenluft, E. (2008).
Risk for bipolar disorder is associated with face-processing deficits across emotions. J Am
Acad Child Adolesc Psychiatry, 47(12), 1455-1461. doi:10.1097/CHI.0b013e318188832e
Brotman, M. A., Tseng, W. L., Olsavsky, A. K., Fromm, S. J., Muhrer, E. J., Rutenberg, J. G., . . .
Leibenluft, E. (2014). Fronto-limbic-striatal dysfunction in pediatric and adult patients with
bipolar disorder: impact of face emotion and attentional demands. Psychol Med, 44(8), 1639-
1651. doi:10.1017/S003329171300202X
Brown, L. A., & Cohen, A. S. (2010). Facial emotion recognition in schizotypy: the role of accuracy
and social cognitive bias. J Int Neuropsychol Soc, 16(3), 474-483.
doi:10.1017/S135561771000007X
Bruce, V. (1986). Influences of familiarity on the processing of faces. Perception, 15(4), 387-397.
doi:10.1068/p150387
Bruce, V., & Young, A. (1986). Understanding face recognition. Br J Psychol, 77 ( Pt 3), 305-327.
Brune, M. (2005). Emotion recognition, 'theory of mind,' and social behavior in schizophrenia.
Psychiatry Res, 133(2-3), 135-147. doi:10.1016/j.psychres.2004.10.007
Burch, G. S., Steel, C., & Hemsley, D. R. (1998). Oxford-Liverpool Inventory of Feelings and
Experiences: reliability in an experimental population. Br J Clin Psychol, 37 ( Pt 1), 107-108.
Burdick, K. E., Goldberg, J. F., Harrow, M., Faull, R. N., & Malhotra, A. K. (2006). Neurocognition
as a stable endophenotype in bipolar disorder and schizophrenia. J Nerv Ment Dis, 194(4),
255-260. doi:10.1097/01.nmd.0000207360.70337.7e
Burton, A. M. (2013). Why has research in face recognition progressed so slowly? The importance of
variability. Quarterly Journal of Experimental Psychology, 66(8), 1467-1485.
doi:10.1080/17470218.2013.800125
Busigny, T., & Rossion, B. (2010). Acquired prosopagnosia abolishes the face inversion effect.
Cortex, 46(8), 965-981. doi:10.1016/j.cortex.2009.07.004
Butler, P. D., Tambini, A., Yovel, G., Jalbrzikowski, M., Ziwich, R., Silipo, G., . . . Javitt, D. C.
(2008). What's in a face? Effects of stimulus duration and inversion on face processing in
schizophrenia. Schizophr Res, 103(1-3), 283-292. doi:10.1016/j.schres.2008.03.007
Caharel, S., Bernard, C., Thibaut, F., Haouzir, S., Di Maggio-Clozel, C., Allio, G., . . . Rebai, M.
(2007). The effects of familiarity and emotional expression on face processing examined by
ERPs in patients with schizophrenia. Schizophr Res, 95(1-3), 186-196.
doi:10.1016/j.schres.2007.06.015
Calder, A. (2011). Does facial identity and facial expression recognition involve separate visual
routes? In A. Calder, G. Rhodes, M. Johnson, & J. V. Haxby (Eds.), Oxford handbook of face
perception (pp. 427-448): Oxford University Press.
182
Calder, A. J. (1996). Facial emotion recognition after bilateral amygdala damage: Differentially
severe impairment of fear. Cognitive Neuropsychology, 13(5), 699-745.
Calder, A. J., & Young, A. W. (2005). Understanding the recognition of facial identity and facial
expression. Nat Rev Neurosci, 6(8), 641-651. doi:10.1038/nrn1724
Calder, A. J., Young, A. W., Keane, J., & Dean, M. (2000). Configural information in facial
expression perception. J Exp Psychol Hum Percept Perform, 26(2), 527-551.
Calkins, M. E., Gur, R. C., Ragland, J. D., & Gur, R. E. (2005). Face recognition memory deficits and
visual object memory performance in patients with schizophrenia and their relatives. Am J
Psychiatry, 162(10), 1963-1966. doi:10.1176/appi.ajp.162.10.1963
Campbell, R., Brooks, B., de Haan, E., & Roberts, T. (1996). Dissociating face processing skills:
decision about lip-read speech, expression, and identity. Q J Exp Psychol A, 49(2), 295-314.
doi:10.1080/713755619
Cantor-Graae, E., Nordstrom, L. G., & McNeil, T. F. (2001). Substance abuse in schizophrenia: a
review of the literature and a study of correlates in Sweden. Schizophr Res, 48(1), 69-82.
Cardno, A. G., Rijsdijk, F. V., Sham, P. C., Murray, R. M., & McGuffin, P. (2002). A twin study of
genetic relationships between psychotic symptoms. Am J Psychiatry, 159(4), 539-545.
doi:10.1176/appi.ajp.159.4.539
Cauchoix, M., Barragan-Jason, G., Serre, T., & Barbeau, E. J. (2014). The neural dynamics of face
detection in the wild revealed by MVPA. J Neurosci, 34(3), 846-854.
doi:10.1523/JNEUROSCI.3030-13.2014
Chambers, R. A., Sentir, A. M., & Engleman, E. A. (2010). Ventral and dorsal striatal dopamine
efflux and behavior in rats with simple vs. co-morbid histories of cocaine sensitization and
neonatal ventral hippocampal lesions. Psychopharmacology (Berl), 212(1), 73-83.
doi:10.1007/s00213-010-1929-1
Chambon, V., Baudouin, J. Y., & Franck, N. (2006). The role of configural information in facial
emotion recognition in schizophrenia. Neuropsychologia, 44(12), 2437-2444.
doi:10.1016/j.neuropsychologia.2006.04.008
Chan, R. C., Li, H., Cheung, E. F., & Gong, Q. Y. (2010). Impaired facial emotion perception in
schizophrenia: a meta-analysis. Psychiatry Res, 178(2), 381-390.
doi:10.1016/j.psychres.2009.03.035
Chang, W. C., Hui, C. L., Tang, J. Y., Wong, G. H., Chan, S. K., Lee, E. H., & Chen, E. Y. (2013).
Impacts of duration of untreated psychosis on cognition and negative symptoms in first-
episode schizophrenia: a 3-year prospective follow-up study. Psychol Med, 43(9), 1883-1893.
doi:10.1017/S0033291712002838
Chang, Y. C., & Huang, S. L. (2012). The influence of attention levels on psychophysiological
responses. Int J Psychophysiol, 86(1), 39-47. doi:10.1016/j.ijpsycho.2012.09.001
Chapman, L. J., Chapman, J. P., & Raulin, M. L. (1978). Body-image aberration in Schizophrenia. J
Abnorm Psychol, 87(4), 399-407.
Chen, W. L., Hsieh, C. H., Chang, H. T., Hung, C. C., & Chan, C. H. (2015). The epidemiology and
progression time from transient to permanent psychiatric disorders of substance-induced
psychosis in Taiwan. Addict Behav, 47, 1-4. doi:10.1016/j.addbeh.2015.02.013
Chen, Y. (2011). Abnormal visual motion processing in schizophrenia: a review of research progress.
Schizophr Bull, 37(4), 709-715. doi:10.1093/schbul/sbr020
Chen, Y., Cataldo, A., Norton, D. J., & Ongur, D. (2012). Distinct facial processing in schizophrenia
and schizoaffective disorders. Schizophr Res, 134(1), 95-100.
doi:10.1016/j.schres.2011.08.001
Chen, Y., & Ekstrom, T. (2016). Perception of faces in schizophrenia: Subjective (self-report) vs.
objective (psychophysics) assessments. J Psychiatr Res, 76, 136-142.
doi:10.1016/j.jpsychires.2016.02.012
Chen, Y., McBain, R., & Norton, D. (2015). Specific vulnerability of face perception to noise: a
similar effect in schizophrenia patients and healthy individuals. Psychiatry Res, 225(3), 619-
624. doi:10.1016/j.psychres.2014.11.035
Chen, Y., Nakayama, K., Levy, D., Matthysse, S., & Holzman, P. (2003). Processing of global, but
not local, motion direction is deficient in schizophrenia. Schizophr Res, 61(2-3), 215-227.
183
Chen, Y., Norton, D., McBain, R., Ongur, D., & Heckers, S. (2009). Visual and cognitive processing
of face information in schizophrenia: detection, discrimination and working memory.
Schizophr Res, 107(1), 92-98. doi:10.1016/j.schres.2008.09.010
Claridge, G. (1997). Schizotypy: Implications for Illness and Health: Oxford University Press.
Claridge, G., McCreery, C., Mason, O., Bentall, R., Boyle, G., Slade, P., & Popplewell, D. (1996).
The factor structure of "schizotypal' traits: a large replication study. Br J Clin Psychol, 35 ( Pt
1), 103-115.
Clarke, M. C., Harley, M., & Cannon, M. (2006). The role of obstetric events in schizophrenia.
Schizophr Bull, 32(1), 3-8. doi:10.1093/schbul/sbj028
Comparelli, A., De Carolis, A., Corigliano, V., Di Pietro, S., Trovini, G., Granese, C., . . . Girardi, P.
(2014). Symptom correlates of facial emotion recognition impairment in schizophrenia.
Psychopathology, 47(1), 65-70. doi:10.1159/000350453
Corcoran, A., Cropper, S. J., Wong, L. Y., Rutherford, K., & Groot, C. J. (2016). Judging judgements
about time: Adapting the temporal bisection task to investigate individual differences in
temporal metacognition. Paper presented at the 43rd Experimental Psychology Conference
Melbourne, Australia.
Corcoran, C. M., Keilp, J. G., Kayser, J., Klim, C., Butler, P. D., Bruder, G. E., . . . Javitt, D. C.
(2015). Emotion recognition deficits as predictors of transition in individuals at clinical high
risk for schizophrenia: a neurodevelopmental perspective. Psychol Med, 45(14), 2959-2973.
doi:10.1017/S0033291715000902
Corwin, J. (1994). On measuring discrimination and response bias: Unequal numbers of targets and
distractors and two classes of distractors. Neuropsychology, 8(1), 110.
Couture, S. M., Penn, D. L., & Roberts, D. L. (2006). The functional significance of social cognition
in schizophrenia: a review. Schizophr Bull, 32 Suppl 1, S44-63. doi:10.1093/schbul/sbl029
Craddock, N., & Owen, M. J. (2010). The Kraepelinian dichotomy - going, going... but still not gone.
Br J Psychiatry, 196(2), 92-95. doi:10.1192/bjp.bp.109.073429
Crawford, J. R., Besson, J. A., Bremner, M., Ebmeier, K. P., Cochrane, R. H., & Kirkwood, K.
(1992). Estimation of premorbid intelligence in schizophrenia. Br J Psychiatry, 161, 69-74.
Cropper, S. J., Groot, C. J., Corcoran, A., Bruno, A., & Johnston, A. (2017). Individual differences in
the perception of (a bigger) time. Journal of Vision, VSS Abstracts.
Cropper, S. J., Johnston, A., & Groot, C. J. (2015). Individual differences in the perception of time.
Journal of Vision, 15(12), 813-813.
Csukly, G., Stefanics, G., Komlosi, S., Czigler, I., & Czobor, P. (2014). Event-related theta
synchronization predicts deficit in facial affect recognition in schizophrenia. J Abnorm
Psychol, 123(1), 178-189. doi:10.1037/a0035793
Cutting, J. E., & Kozlowski, L. T. (1977). Recognizing friends by their walk: Gait perception without
familiarity cues. Bulletin of the Psychonomic Society, 9(5), 353-356. doi:10.3758/bf03337021
Dacquino, C., De Rossi, P., & Spalletta, G. (2015). Schizophrenia and bipolar disorder: The road from
similarities and clinical heterogeneity to neurobiological types. Clin Chim Acta, 449, 49-59.
doi:10.1016/j.cca.2015.02.029
Dalili, M. N., Penton-Voak, I. S., Harmer, C. J., & Munafo, M. R. (2015). Meta-analysis of emotion
recognition deficits in major depressive disorder. Psychol Med, 45(6), 1135-1144.
doi:10.1017/S0033291714002591
Darke, H., Peterman, J. S., Park, S., Sundram, S., & Carter, O. (2013). Are patients with
schizophrenia impaired in processing non-emotional features of human faces? Front Psychol,
4, 529. doi:10.3389/fpsyg.2013.00529
Daros, A. R., Zakzanis, K. K., & Ruocco, A. C. (2013). Facial emotion recognition in borderline
personality disorder. Psychol Med, 43(9), 1953-1963. doi:10.1017/S0033291712002607
Darwin, C., Ekman, P., & Prodger, P. (1998). The expression of the emotions in man and animals:
Oxford University Press, USA.
Davis, P. J., & Gibson, M. G. (2000). Recognition of posed and genuine facial expressions of emotion
in paranoid and nonparanoid schizophrenia. J Abnorm Psychol, 109(3), 445-450.
de Achaval, D., Costanzo, E. Y., Villarreal, M., Jauregui, I. O., Chiodi, A., Castro, M. N., . . .
Guinjoan, S. M. (2010). Emotion processing and theory of mind in schizophrenia patients and
184
their unaffected first-degree relatives. Neuropsychologia, 48(5), 1209-1215.
doi:10.1016/j.neuropsychologia.2009.12.019
de Bont, P. A., van den Berg, D. P., van der Vleugel, B. M., de Roos, C., de Jongh, A., van der Gaag,
M., & van Minnen, A. M. (2016). Prolonged exposure and EMDR for PTSD v. a PTSD
waiting-list condition: effects on symptoms of psychosis, depression and social functioning in
patients with chronic psychotic disorders. Psychol Med, 46(11), 2411-2421.
doi:10.1017/S0033291716001094
Deffke, I., Sander, T., Heidenreich, J., Sommer, W., Curio, G., Trahms, L., & Lueschow, A. (2007).
MEG/EEG sources of the 170-ms response to faces are co-localized in the fusiform gyrus.
Neuroimage, 35(4), 1495-1501. doi:10.1016/j.neuroimage.2007.01.034
Deiber, M. P., Missonnier, P., Bertrand, O., Gold, G., Fazio-Costa, L., Ibanez, V., & Giannakopoulos,
P. (2007). Distinction between perceptual and attentional processing in working memory
tasks: a study of phase-locked and induced oscillatory brain dynamics. J Cogn Neurosci,
19(1), 158-172. doi:10.1162/jocn.2007.19.1.158
Delerue, C., Laprevote, V., Verfaillie, K., & Boucart, M. (2010). Gaze control during face exploration
in schizophrenia. Neurosci Lett, 482(3), 245-249. doi:10.1016/j.neulet.2010.07.048
Delvecchio, G., Fossati, P., Boyer, P., Brambilla, P., Falkai, P., Gruber, O., . . . Frangou, S. (2012).
Common and distinct neural correlates of emotional processing in Bipolar Disorder and
Major Depressive Disorder: a voxel-based meta-analysis of functional magnetic resonance
imaging studies. Eur Neuropsychopharmacol, 22(2), 100-113.
doi:10.1016/j.euroneuro.2011.07.003
Delvecchio, G., Sugranyes, G., & Frangou, S. (2013). Evidence of diagnostic specificity in the neural
correlates of facial affect processing in bipolar disorder and schizophrenia: a meta-analysis of
functional imaging studies. Psychol Med, 43(3), 553-569. doi:10.1017/S0033291712001432
Demos, K. E., Kelley, W. M., Ryan, S. L., Davis, F. C., & Whalen, P. J. (2008). Human amygdala
sensitivity to the pupil size of others. Cereb Cortex, 18(12), 2729-2734.
doi:10.1093/cercor/bhn034
Derntl, B., Seidel, E. M., Kryspin-Exner, I., Hasmann, A., & Dobmeier, M. (2009). Facial emotion
recognition in patients with bipolar I and bipolar II disorder. Br J Clin Psychol, 48(Pt 4), 363-
375. doi:10.1348/014466509X404845
Derntl, B., Seidel, E. M., Schneider, F., & Habel, U. (2012). How specific are emotional deficits? A
comparison of empathic abilities in schizophrenia, bipolar and depressed patients. Schizophr
Res, 142(1-3), 58-64. doi:10.1016/j.schres.2012.09.020
DeRosse, P., & Karlsgodt, K. H. (2015). Examining the Psychosis Continuum. Curr Behav Neurosci
Rep, 2(2), 80-89. doi:10.1007/s40473-015-0040-7
DiCiccio, T. J., & Efron, B. (1996). Bootstrap confidence intervals. Statist. Sci., 11(3), 189-228.
doi:10.1214/ss/1032280214
Doniger, G. M., Foxe, J. J., Murray, M. M., Higgins, B. A., & Javitt, D. C. (2002). Impaired visual
object recognition and dorsal/ventral stream interaction in schizophrenia. Arch Gen
Psychiatry, 59(11), 1011-1020.
Doop, M. L., & Park, S. (2009). Facial expression and face orientation processing in schizophrenia.
Psychiatry Res, 170(2-3), 103-107. doi:10.1016/j.psychres.2009.06.009
Dorph-Petersen, K. A., Pierri, J. N., Wu, Q., Sampson, A. R., & Lewis, D. A. (2007). Primary visual
cortex volume and total neuron number are reduced in schizophrenia. J Comp Neurol, 501(2),
290-301. doi:10.1002/cne.21243
Douglas, K. M., & Porter, R. J. (2010). Recognition of disgusted facial expressions in severe
depression. Br J Psychiatry, 197(2), 156-157. doi:10.1192/bjp.bp.110.078113
Duchaine, B. C., Parker, H., & Nakayama, K. (2003). Normal recognition of emotion in a
prosopagnosic. Perception, 32(7), 827-838. doi:10.1068/p5067
Edwards, J., Jackson, H. J., & Pattison, P. E. (2002). Emotion recognition via facial expression and
affective prosody in schizophrenia: a methodological review. Clin Psychol Rev, 22(6), 789-
832.
Edwards, J., Pattison, P. E., Jackson, H. J., & Wales, R. J. (2001). Facial affect and affective prosody
recognition in first-episode schizophrenia. Schizophr Res, 48(2-3), 235-253.
185
Eimer, M. (2000). The face-specific N170 component reflects late stages in the structural encoding of
faces. Neuroreport, 11(10), 2319-2324.
Ekman, P., & Friesen, W. V. (1971). Constants across cultures in the face and emotion. J Pers Soc
Psychol, 17(2), 124-129.
Ekman, P., & Friesen, W. V. (1976). Pictures of Facial Affect: Consulting psychologists Press.
Ekman, P., Friesen, W. V., O'Sullivan, M., Chan, A., Diacoyanni-Tarlatzis, I., Heider, K., . . . et al.
(1987). Universals and cultural differences in the judgments of facial expressions of emotion.
J Pers Soc Psychol, 53(4), 712-717.
Ekman, P., Sorenson, E. R., & Friesen, W. V. (1969). Pan-cultural elements in facial displays of
emotion. Science, 164(3875), 86-88.
Etkin, A., & Wager, T. D. (2007). Functional neuroimaging of anxiety: a meta-analysis of emotional
processing in PTSD, social anxiety disorder, and specific phobia. Am J Psychiatry, 164(10),
1476-1488. doi:10.1176/appi.ajp.2007.07030504
Evangeli, M., & Broks, M. E. (2000). Face processing in schizophrenia: parallels with the effects of
amygdala damage. Cogn Neuropsychiatry, 5(2), 81-104. doi:10.1080/135468000395754
Eysenck, M. W., & Derakshan, N. (2011). New perspectives in attentional control theory. Personality
and Individual Differences, 50(7), 955-960. doi:https://doi.org/10.1016/j.paid.2010.08.019
Fakra, E., Jouve, E., Guillaume, F., Azorin, J. M., & Blin, O. (2015). Relation between facial affect
recognition and configural face processing in antipsychotic-free schizophrenia.
Neuropsychology, 29(2), 197-204. doi:10.1037/neu0000136
Fakra, E., Salgado-Pineda, P., Delaveau, P., Hariri, A. R., & Blin, O. (2008). Neural bases of different
cognitive strategies for facial affect processing in schizophrenia. Schizophr Res, 100(1-3),
191-205. doi:10.1016/j.schres.2007.11.040
Fett, A. K., Maat, A., & Investigators, G. (2013). Social cognitive impairments and psychotic
symptoms: what is the nature of their association? Schizophr Bull, 39(1), 77-85.
doi:10.1093/schbul/sbr058
Fiorentini, C., & Viviani, P. (2011). Is there a dynamic advantage for facial expressions? J Vis, 11(3).
doi:10.1167/11.3.17
Fitousi, D., & Wenger, M. J. (2013). Variants of independence in the perception of facial identity and
expression. J Exp Psychol Hum Percept Perform, 39(1), 133-155. doi:10.1037/a0028001
Forbes, N. F., Carrick, L. A., McIntosh, A. M., & Lawrie, S. M. (2009). Working memory in
schizophrenia: a meta-analysis. Psychol Med, 39(6), 889-905.
doi:10.1017/S0033291708004558
Fornito, A., Yucel, M., Patti, J., Wood, S. J., & Pantelis, C. (2009). Mapping grey matter reductions in
schizophrenia: an anatomical likelihood estimation analysis of voxel-based morphometry
studies. Schizophr Res, 108(1-3), 104-113. doi:10.1016/j.schres.2008.12.011
Fujimura, T., & Suzuki, N. (2010). [Recognition of dynamic facial expressions in peripheral and
central vision]. Shinrigaku Kenkyu, 81(4), 348-355.
Fusar-Poli, P., Placentino, A., Carletti, F., Landi, P., Allen, P., Surguladze, S., . . . Politi, P. (2009).
Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional
magnetic resonance imaging studies. J Psychiatry Neurosci, 34(6), 418-432.
Gaebel, W., & Wolwer, W. (1992). Facial expression and emotional face recognition in schizophrenia
and depression. Eur Arch Psychiatry Clin Neurosci, 242(1), 46-52.
Galati, D., Sini, B., Schmidt, S., & Tinti, C. (2003). Spontaneous facial expressions in congenitally
blind and sighted children aged 8-11. Journal of Visual Impairment and Blindness, 97(7),
418-428.
Gasper, K., & Clore, G. L. (2002). Attending to the big picture: mood and global versus local
processing of visual information. Psychol Sci, 13(1), 34-40. doi:10.1111/1467-9280.00406
Gauthier, I., Tarr, M. J., Moylan, J., Skudlarski, P., Gore, J. C., & Anderson, A. W. (2000). The
fusiform "face area" is part of a network that processes faces at the individual level. J Cogn
Neurosci, 12(3), 495-504.
George, M. S., Ketter, T. A., Gill, D. S., Haxby, J. V., Ungerleider, L. G., Herscovitch, P., & Post, R.
M. (1993). Brain regions involved in recognizing facial emotion or identity: an oxygen-15
PET study. J Neuropsychiatry Clin Neurosci, 5(4), 384-394. doi:10.1176/jnp.5.4.384
186
Germine, L. T., & Hooker, C. I. (2011). Face emotion recognition is related to individual differences
in psychosis-proneness. Psychol Med, 41(5), 937-947. doi:10.1017/S0033291710001571
Getz, G. E., Shear, P. K., & Strakowski, S. M. (2003). Facial affect recognition deficits in bipolar
disorder. J Int Neuropsychol Soc, 9(4), 623-632. doi:10.1017/S1355617703940021
Glahn, D. C., Bearden, C. E., Barguil, M., Barrett, J., Reichenberg, A., Bowden, C. L., . . . Velligan,
D. I. (2007). The neurocognitive signature of psychotic bipolar disorder. Biol Psychiatry,
62(8), 910-916. doi:10.1016/j.biopsych.2007.02.001
Goghari, V. M., Macdonald, A. W., 3rd, & Sponheim, S. R. (2011). Temporal lobe structures and
facial emotion recognition in schizophrenia patients and nonpsychotic relatives. Schizophr
Bull, 37(6), 1281-1294. doi:10.1093/schbul/sbq046
Goghari, V. M., & Sponheim, S. R. (2013). More pronounced deficits in facial emotion recognition
for schizophrenia than bipolar disorder. Compr Psychiatry, 54(4), 388-397.
doi:10.1016/j.comppsych.2012.10.012
Gold, J. M., Barch, D. M., Carter, C. S., Dakin, S., Luck, S. J., MacDonald, A. W., 3rd, . . . Strauss,
M. (2012). Clinical, functional, and intertask correlations of measures developed by the
Cognitive Neuroscience Test Reliability and Clinical Applications for Schizophrenia
Consortium. Schizophr Bull, 38(1), 144-152. doi:10.1093/schbul/sbr142
Gold, J. M., & Dickinson, D. (2013). "Generalized cognitive deficit" in schizophrenia: overused or
underappreciated? Schizophr Bull, 39(2), 263-265. doi:10.1093/schbul/sbs143
Gold, J. M., Hahn, B., Strauss, G. P., & Waltz, J. A. (2009). Turning it upside down: areas of
preserved cognitive function in schizophrenia. Neuropsychol Rev, 19(3), 294-311.
doi:10.1007/s11065-009-9098-x
Gollan, J. K., Pane, H. T., McCloskey, M. S., & Coccaro, E. F. (2008). Identifying differences in
biased affective information processing in major depression. Psychiatry Res, 159(1-2), 18-24.
doi:10.1016/j.psychres.2007.06.011
Goodarzi, M. A., Wykes, T., & Hemsley, D. R. (2000). Cerebral lateralization of global-local
processing in people with schizotypy. Schizophr Res, 45(1-2), 115-121.
Gordon, E., Coyle, S., Anderson, J., Healey, P., Cordaro, J., Latimer, C., & Meares, R. (1992). Eye
movement response to a facial stimulus in schizophrenia. Biol Psychiatry, 31(6), 626-629.
Gray, J., Venn, H., Montagne, B., Murray, L., Burt, M., Frigerio, E., . . . Young, A. H. (2006). Bipolar
patients show mood-congruent biases in sensitivity to facial expressions of emotion when
exhibiting depressed symptoms, but not when exhibiting manic symptoms. Cogn
Neuropsychiatry, 11(6), 505-520. doi:10.1080/13546800544000028
Green, M. F., & Harvey, P. D. (2014). Cognition in schizophrenia: Past, present, and future.
Schizophr Res Cogn, 1(1), e1-e9. doi:10.1016/j.scog.2014.02.001
Green, M. F., Horan, W. P., & Lee, J. (2015). Social cognition in schizophrenia. Nat Rev Neurosci,
16(10), 620-631. doi:10.1038/nrn4005
Green, M. F., Horan, W. P., & Sugar, C. A. (2013). Has the generalized deficit become the
generalized criticism? Schizophr Bull, 39(2), 257-262. doi:10.1093/schbul/sbs146
Green, M. J., & Phillips, M. L. (2004). Social threat perception and the evolution of paranoia.
Neurosci Biobehav Rev, 28(3), 333-342. doi:10.1016/j.neubiorev.2004.03.006
Green, M. J., Waldron, J. H., & Coltheart, M. (2007). Emotional context processing is impaired in
schizophrenia. Cogn Neuropsychiatry, 12(3), 259-280. doi:10.1080/13546800601051847
Green, M. J., Waldron, J. H., Simpson, I., & Coltheart, M. (2008). Visual processing of social context
during mental state perception in schizophrenia. J Psychiatry Neurosci, 33(1), 34-42.
Green, M. J., Williams, L. M., & Davidson, D. (2003). Visual scanpaths to threat-related faces in
deluded schizophrenia. Psychiatry Res, 119(3), 271-285.
Grill-Spector, K., Knouf, N., & Kanwisher, N. (2004). The fusiform face area subserves face
perception, not generic within-category identification. Nat Neurosci, 7(5), 555-562.
doi:10.1038/nn1224
Gur, R. C., Erwin, R. J., Gur, R. E., Zwil, A. S., Heimberg, C., & Kraemer, H. C. (1992). Facial
emotion discrimination: II. Behavioral findings in depression. Psychiatry Res, 42(3), 241-251.
Gur, R. C., Schroeder, L., Turner, T., McGrath, C., Chan, R. M., Turetsky, B. I., . . . Gur, R. E.
(2002). Brain activation during facial emotion processing. Neuroimage, 16(3 Pt 1), 651-662.
187
Gur, R. E., Kohler, C. G., Ragland, J. D., Siegel, S. J., Lesko, K., Bilker, W. B., & Gur, R. C. (2006).
Flat affect in schizophrenia: relation to emotion processing and neurocognitive measures.
Schizophr Bull, 32(2), 279-287. doi:10.1093/schbul/sbj041
Gur, R. E., McGrath, C., Chan, R. M., Schroeder, L., Turner, T., Turetsky, B. I., . . . Gur, R. C.
(2002). An fMRI study of facial emotion processing in patients with schizophrenia. Am J
Psychiatry, 159(12), 1992-1999. doi:10.1176/appi.ajp.159.12.1992
Habel, U., Chechko, N., Pauly, K., Koch, K., Backes, V., Seiferth, N., . . . Kellermann, T. (2010).
Neural correlates of emotion recognition in schizophrenia. Schizophr Res, 122(1-3), 113-123.
doi:10.1016/j.schres.2010.06.009
Habel, U., Gur, R. C., Mandal, M. K., Salloum, J. B., Gur, R. E., & Schneider, F. (2000). Emotional
processing in schizophrenia across cultures: standardized measures of discrimination and
experience. Schizophr Res, 42(1), 57-66.
Hale, W. W., 3rd. (1998). Judgment of facial expressions and depression persistence. Psychiatry Res,
80(3), 265-274.
Hall, J., Harris, J. M., Sprengelmeyer, R., Sprengelmeyer, A., Young, A. W., Santos, I. M., . . .
Lawrie, S. M. (2004). Social cognition and face processing in schizophrenia. Br J Psychiatry,
185, 169-170. doi:10.1192/bjp.185.2.169
Hall, J., Whalley, H. C., McKirdy, J. W., Romaniuk, L., McGonigle, D., McIntosh, A. M., . . . Lawrie,
S. M. (2008). Overactivation of fear systems to neutral faces in schizophrenia. Biol
Psychiatry, 64(1), 70-73. doi:10.1016/j.biopsych.2007.12.014
Hanley, J. R. (2011). An appreciation of Bruce and Young's (1986) serial stage model of face naming
after 25 years. Br J Psychol, 102(4), 915-930. doi:10.1111/j.2044-8295.2011.02032.x
Hanssen, M., Bak, M., Bijl, R., Vollebergh, W., & van Os, J. (2005). The incidence and outcome of
subclinical psychotic experiences in the general population. Br J Clin Psychol, 44(Pt 2), 181-
191. doi:10.1348/014466505X29611
Hargreaves, A., Mothersill, O., Anderson, M., Lawless, S., Corvin, A., & Donohoe, G. (2016).
Detecting facial emotion recognition deficits in schizophrenia using dynamic stimuli of
varying intensities. Neurosci Lett, 633, 47-54. doi:10.1016/j.neulet.2016.09.017
Harmer, C. J., Grayson, L., & Goodwin, G. M. (2002). Enhanced recognition of disgust in bipolar
illness. Biol Psychiatry, 51(4), 298-304.
Harms, M. B., Martin, A., & Wallace, G. L. (2010). Facial emotion recognition in autism spectrum
disorders: a review of behavioral and neuroimaging studies. Neuropsychol Rev, 20(3), 290-
322. doi:10.1007/s11065-010-9138-6
Harvey Jr, L. O. (2003). Detection sensitivity and response bias. Psychology of Perception.
Psychology, 4165.
Harvey, P. D., Patterson, T. L., Potter, L. S., Zhong, K., & Brecher, M. (2006). Improvement in social
competence with short-term atypical antipsychotic treatment: a randomized, double-blind
comparison of quetiapine versus risperidone for social competence, social cognition, and
neuropsychological functioning. Am J Psychiatry, 163(11), 1918-1925.
doi:10.1176/ajp.2006.163.11.1918
Hattingh, C. J., Ipser, J., Tromp, S. A., Syal, S., Lochner, C., Brooks, S. J., & Stein, D. J. (2012).
Functional magnetic resonance imaging during emotion recognition in social anxiety
disorder: an activation likelihood meta-analysis. Front Hum Neurosci, 6, 347.
doi:10.3389/fnhum.2012.00347
Haxby, J. V., & Gobbini, M. I. (2011). Distributed neural systems for face perception. In A. Calder,
G. Rhodes, M. Johnson, & J. V. Haxby (Eds.), Oxford handbook of face perception (pp. 93-
110): Oxford University Press.
Haxby, J. V., Hoffman, E. A., & Gobbini, M. I. (2000). The distributed human neural system for face
perception. Trends Cogn Sci, 4(6), 223-233.
Haxby, J. V., Ungerleider, L. G., Horwitz, B., Maisog, J. M., Rapoport, S. I., & Grady, C. L. (1996).
Face encoding and recognition in the human brain. Proc Natl Acad Sci U S A, 93(2), 922-927.
Hayes, J. P., Hayes, S. M., & Mikedis, A. M. (2012). Quantitative meta-analysis of neural activity in
posttraumatic stress disorder. Biol Mood Anxiety Disord, 2, 9. doi:10.1186/2045-5380-2-9
Heimberg, C., Gur, R. E., Erwin, R. J., Shtasel, D. L., & Gur, R. C. (1992). Facial emotion
discrimination: III. Behavioral findings in schizophrenia. Psychiatry Res, 42(3), 253-265.
188
Heinisch, C., Wiens, S., Grundl, M., Juckel, G., & Brune, M. (2013). Self-face recognition in
schizophrenia is related to insight. Eur Arch Psychiatry Clin Neurosci, 263(8), 655-662.
doi:10.1007/s00406-013-0400-9
Heinz, A., Deserno, L., & Reininghaus, U. (2013). Urbanicity, social adversity and psychosis. World
Psychiatry, 12(3), 187-197. doi:10.1002/wps.20056
Hempel, A., Hempel, E., Schonknecht, P., Stippich, C., & Schroder, J. (2003). Impairment in basal
limbic function in schizophrenia during affect recognition. Psychiatry Res, 122(2), 115-124.
Herrmann, M. J., Ehlis, A. C., Muehlberger, A., & Fallgatter, A. J. (2005). Source localization of
early stages of face processing. Brain Topogr, 18(2), 77-85. doi:10.1007/s10548-005-0277-7
Hill, S. K., Reilly, J. L., Harris, M. S., Rosen, C., Marvin, R. W., Deleon, O., & Sweeney, J. A.
(2009). A comparison of neuropsychological dysfunction in first-episode psychosis patients
with unipolar depression, bipolar disorder, and schizophrenia. Schizophr Res, 113(2-3), 167-
175. doi:10.1016/j.schres.2009.04.020
Hill, S. K., Reilly, J. L., Keefe, R. S., Gold, J. M., Bishop, J. R., Gershon, E. S., . . . Sweeney, J. A.
(2013). Neuropsychological impairments in schizophrenia and psychotic bipolar disorder:
findings from the Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP)
study. Am J Psychiatry, 170(11), 1275-1284. doi:10.1176/appi.ajp.2013.12101298
Hjorthoj, C., Sturup, A. E., McGrath, J. J., & Nordentoft, M. (2017). Years of potential life lost and
life expectancy in schizophrenia: a systematic review and meta-analysis. Lancet Psychiatry,
4(4), 295-301. doi:10.1016/S2215-0366(17)30078-0
Hoffman, E. A., & Haxby, J. V. (2000). Distinct representations of eye gaze and identity in the
distributed human neural system for face perception. Nat Neurosci, 3(1), 80-84.
doi:10.1038/71152
Holt, D. J., Kunkel, L., Weiss, A. P., Goff, D. C., Wright, C. I., Shin, L. M., . . . Heckers, S. (2006).
Increased medial temporal lobe activation during the passive viewing of emotional and
neutral facial expressions in schizophrenia. Schizophr Res, 82(2-3), 153-162.
doi:10.1016/j.schres.2005.09.021
Holt, D. J., Titone, D., Long, L. S., Goff, D. C., Cather, C., Rauch, S. L., . . . Kuperberg, G. R. (2006).
The misattribution of salience in delusional patients with schizophrenia. Schizophr Res, 83(2-
3), 247-256. doi:10.1016/j.schres.2005.12.858
Holt, D. J., Weiss, A. P., Rauch, S. L., Wright, C. I., Zalesak, M., Goff, D. C., . . . Heckers, S. (2005).
Sustained activation of the hippocampus in response to fearful faces in schizophrenia. Biol
Psychiatry, 57(9), 1011-1019. doi:10.1016/j.biopsych.2005.01.033
Honea, R., Crow, T. J., Passingham, D., & Mackay, C. E. (2005). Regional deficits in brain volume in
schizophrenia: a meta-analysis of voxel-based morphometry studies. Am J Psychiatry,
162(12), 2233-2245. doi:10.1176/appi.ajp.162.12.2233
Hooker, C., & Park, S. (2002). Emotion processing and its relationship to social functioning in
schizophrenia patients. Psychiatry Res, 112(1), 41-50.
Hooker, C. I., Bruce, L., Fisher, M., Verosky, S. C., Miyakawa, A., D'Esposito, M., & Vinogradov, S.
(2013). The influence of combined cognitive plus social-cognitive training on amygdala
response during face emotion recognition in schizophrenia. Psychiatry Res, 213(2), 99-107.
doi:10.1016/j.pscychresns.2013.04.001
Hooker, C. I., Bruce, L., Fisher, M., Verosky, S. C., Miyakawa, A., & Vinogradov, S. (2012). Neural
activity during emotion recognition after combined cognitive plus social cognitive training in
schizophrenia. Schizophr Res, 139(1-3), 53-59. doi:10.1016/j.schres.2012.05.009
Hooper, S. R., Giuliano, A. J., Youngstrom, E. A., Breiger, D., Sikich, L., Frazier, J. A., . . .
Lieberman, J. A. (2010). Neurocognition in early-onset schizophrenia and schizoaffective
disorders. J Am Acad Child Adolesc Psychiatry, 49(1), 52-60.
Horan, W. P., Foti, D., Hajcak, G., Wynn, J. K., & Green, M. F. (2012). Intact motivated attention in
schizophrenia: evidence from event-related potentials. Schizophr Res, 135(1-3), 95-99.
doi:10.1016/j.schres.2011.11.005
Hori, Y., Fukuzako, H., Sugimoto, Y., & Takigawa, M. (2002). Eye movements during the Rorschach
test in schizophrenia. Psychiatry Clin Neurosci, 56(4), 409-418. doi:10.1046/j.1440-
1819.2002.01030.x
189
Horley, K., Williams, L. M., Gonsalvez, C., & Gordon, E. (2003). Social phobics do not see eye to
eye: a visual scanpath study of emotional expression processing. J Anxiety Disord, 17(1), 33-
44.
Hornak, J., Rolls, E. T., & Wade, D. (1996). Face and voice expression identification in patients with
emotional and behavioural changes following ventral frontal lobe damage. Neuropsychologia,
34(4), 247-261.
Hoshi, R., Scoales, M., Mason, O., & Kamboj, S. K. (2011). Schizotypy and emotional memory.
Journal of Behavior Therapy and Experimental Psychiatry, 42(4), 504-510.
doi:http://dx.doi.org/10.1016/j.jbtep.2011.05.003
Houenou, J., Boisgontier, J., Henrion, A., d'Albis, M. A., Dumaine, A., Linke, J., . . . Jamain, S.
(2017). A Multilevel Functional Study of a SNAP25 At-Risk Variant for Bipolar Disorder
and Schizophrenia. J Neurosci, 37(43), 10389-10397. doi:10.1523/JNEUROSCI.1040-
17.2017
Huang, C. L., Hsiao, S., Hwu, H. G., & Howng, S. L. (2013). Are there differential deficits in facial
emotion recognition between paranoid and non-paranoid schizophrenia? A signal detection
analysis. Psychiatry Res, 209(3), 424-430. doi:10.1016/j.psychres.2013.03.026
Hulvershorn, L. A., Karne, H., Gunn, A. D., Hartwick, S. L., Wang, Y., Hummer, T. A., & Anand, A.
(2012). Neural activation during facial emotion processing in unmedicated bipolar depression,
euthymia, and mania. Biol Psychiatry, 71(7), 603-610. doi:10.1016/j.biopsych.2011.10.038
Humphreys, G. W., Donnelly, N., & Riddoch, M. J. (1993). Expression is computed separately from
facial identity, and it is computed separately for moving and static faces: neuropsychological
evidence. Neuropsychologia, 31(2), 173-181.
Humphreys, K., Avidan, G., & Behrmann, M. (2007). A detailed investigation of facial expression
processing in congenital prosopagnosia as compared to acquired prosopagnosia. Exp Brain
Res, 176(2), 356-373. doi:10.1007/s00221-006-0621-5
Ibanez, A., Riveros, R., Hurtado, E., Gleichgerrcht, E., Urquina, H., Herrera, E., . . . Manes, F. (2012).
The face and its emotion: right N170 deficits in structural processing and early emotional
discrimination in schizophrenic patients and relatives. Psychiatry Res, 195(1-2), 18-26.
doi:10.1016/j.psychres.2011.07.027
Irani, F., Platek, S. M., Panyavin, I. S., Calkins, M. E., Kohler, C., Siegel, S. J., . . . Gur, R. C. (2006).
Self-face recognition and theory of mind in patients with schizophrenia and first-degree
relatives. Schizophr Res, 88(1-3), 151-160. doi:10.1016/j.schres.2006.07.016
Irani, F., Seligman, S., Kamath, V., Kohler, C., & Gur, R. C. (2012). A meta-analysis of emotion
perception and functional outcomes in schizophrenia. Schizophr Res, 137(1-3), 203-211.
doi:10.1016/j.schres.2012.01.023
Izard, C. E. (1971). The face of emotion: Appleton-Century-Crofts.
Jahshan, C. S., & Sergi, M. J. (2007). Theory of mind, neurocognition, and functional status in
schizotypy. Schizophr Res, 89(1-3), 278-286. doi:10.1016/j.schres.2006.09.004
Jaworska, N., Yang, X. R., Knott, V., & MacQueen, G. (2015). A review of fMRI studies during
visual emotive processing in major depressive disorder. World J Biol Psychiatry, 16(7), 448-
471. doi:10.3109/15622975.2014.885659
Jia, H., Yang, J., Zhu, H., Liu, J., & Barnaby, N. (2015). Self-face recognition in the ultra-high risk
for psychosis population. Early Interv Psychiatry, 9(2), 126-132. doi:10.1111/eip.12097
Johansson, G. (1973). Visual perception of biological motion and a model for its analysis. Perception
& Psychophysics, 14(2), 201-211. doi:10.3758/bf03212378
Johnson, S. C., Lowery, N., Kohler, C., & Turetsky, B. I. (2005). Global-local visual processing in
schizophrenia: evidence for an early visual processing deficit. Biol Psychiatry, 58(12), 937-
946. doi:10.1016/j.biopsych.2005.04.053
Johnston, P. J., Devir, H., & Karayanidis, F. (2006). Facial emotion processing in schizophrenia: no
evidence for a deficit specific to negative emotions in a differential deficit design. Psychiatry
Res, 143(1), 51-61. doi:10.1016/j.psychres.2005.08.006
Johnston, P. J., Enticott, P. G., Mayes, A. K., Hoy, K. E., Herring, S. E., & Fitzgerald, P. B. (2010).
Symptom correlates of static and dynamic facial affect processing in schizophrenia: evidence
of a double dissociation? Schizophr Bull, 36(4), 680-687. doi:10.1093/schbul/sbn136
190
Jokisch, D., Daum, I., Suchan, B., & Troje, N. F. (2005). Structural encoding and recognition of
biological motion: evidence from event-related potentials and source analysis. Behav Brain
Res, 157(2), 195-204. doi:10.1016/j.bbr.2004.06.025
Joormann, J., & Gotlib, I. H. (2006). Is this happiness I see? Biases in the identification of emotional
facial expressions in depression and social phobia. J Abnorm Psychol, 115(4), 705-714.
doi:10.1037/0021-843X.115.4.705
Joshua, N., & Rossell, S. (2009). Configural face processing in schizophrenia. Schizophr Res, 112(1-
3), 99-103. doi:10.1016/j.schres.2009.03.033
Jung, H. T., Kim, D. W., Kim, S., Im, C. H., & Lee, S. H. (2012). Reduced source activity of event-
related potentials for affective facial pictures in schizophrenia patients. Schizophr Res, 136(1-
3), 150-159. doi:10.1016/j.schres.2011.10.023
Kamachi, M., Bruce, V., Mukaida, S., Gyoba, J., Yoshikawa, S., & Akamatsu, S. (2001). Dynamic
properties influence the perception of facial expressions. Perception, 30(7), 875-887.
doi:10.1068/p3131
Kan, Y., Mimura, M., Kamijima, K., & Kawamura, M. (2004). Recognition of emotion from moving
facial and prosodic stimuli in depressed patients. J Neurol Neurosurg Psychiatry, 75(12),
1667-1671. doi:10.1136/jnnp.2004.036079
Kanade, T., Cohn, J. F., & Tian, Y. (2000). Comprehensive database for facial expression analysis.
Paper presented at the Automatic Face and Gesture Recognition, 2000. Proceedings. Fourth
IEEE International Conference on.
Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform face area: a module in human
extrastriate cortex specialized for face perception. J Neurosci, 17(11), 4302-4311.
Kanwisher, N., & Yovel, G. (2006). The fusiform face area: a cortical region specialized for the
perception of faces. Philos Trans R Soc Lond B Biol Sci, 361(1476), 2109-2128.
doi:10.1098/rstb.2006.1934
Kaufman, J., & Johnston, P. J. (2014). Facial motion engages predictive visual mechanisms. PLoS
One, 9(3), e91038. doi:10.1371/journal.pone.0091038
Kay, S. R., Fiszbein, A., & Opler, L. A. (1987). The positive and negative syndrome scale (PANSS)
for schizophrenia. Schizophr Bull, 13(2), 261-276.
Kee, K. S., Green, M. F., Mintz, J., & Brekke, J. S. (2003). Is emotion processing a predictor of
functional outcome in schizophrenia? Schizophr Bull, 29(3), 487-497.
Keefe, R. S., Eesley, C. E., & Poe, M. P. (2005). Defining a cognitive function decrement in
schizophrenia. Biol Psychiatry, 57(6), 688-691. doi:10.1016/j.biopsych.2005.01.003
Keefe, R. S., Fox, K. H., Harvey, P. D., Cucchiaro, J., Siu, C., & Loebel, A. (2011). Characteristics of
the MATRICS Consensus Cognitive Battery in a 29-site antipsychotic schizophrenia clinical
trial. Schizophr Res, 125(2-3), 161-168. doi:10.1016/j.schres.2010.09.015
Keefe, R. S., & Harvey, P. D. (2012). Cognitive impairment in schizophrenia. Handb Exp
Pharmacol(213), 11-37. doi:10.1007/978-3-642-25758-2_2
Keener, M. T., Fournier, J. C., Mullin, B. C., Kronhaus, D., Perlman, S. B., LaBarbara, E., . . .
Phillips, M. L. (2012). Dissociable patterns of medial prefrontal and amygdala activity to face
identity versus emotion in bipolar disorder. Psychol Med, 42(9), 1913-1924.
doi:10.1017/S0033291711002935
Kelleher, I., & Cannon, M. (2011). Psychotic-like experiences in the general population:
characterizing a high-risk group for psychosis. Psychol Med, 41(1), 1-6.
doi:10.1017/S0033291710001005
Kelleher, I., & Cannon, M. (2016). Putting Psychosis in Its Place. Am J Psychiatry, 173(10), 951-952.
doi:10.1176/appi.ajp.2016.16070810
Kelleher, I., Corcoran, P., Keeley, H., Wigman, J. T., Devlin, N., Ramsay, H., . . . Cannon, M. (2013).
Psychotic symptoms and population risk for suicide attempt: a prospective cohort study.
JAMA Psychiatry, 70(9), 940-948. doi:10.1001/jamapsychiatry.2013.140
Kelleher, I., Devlin, N., Wigman, J. T., Kehoe, A., Murtagh, A., Fitzpatrick, C., & Cannon, M.
(2014). Psychotic experiences in a mental health clinic sample: implications for suicidality,
multimorbidity and functioning. Psychol Med, 44(8), 1615-1624.
doi:10.1017/S0033291713002122
191
Kennerley, S. W., Walton, M. E., Behrens, T. E., Buckley, M. J., & Rushworth, M. F. (2006). Optimal
decision making and the anterior cingulate cortex. Nat Neurosci, 9(7), 940-947.
doi:10.1038/nn1724
Kerns, J. G. (2005). Positive schizotypy and emotion processing. J Abnorm Psychol, 114(3), 392-401.
doi:10.1037/0021-843X.114.3.392
Kerr, S. L., & Neale, J. M. (1993). Emotion perception in schizophrenia: specific deficit or further
evidence of generalized poor performance? J Abnorm Psychol, 102(2), 312-318.
Keshavan, M. S., Nasrallah, H. A., & Tandon, R. (2011). Schizophrenia, "Just the Facts" 6. Moving
ahead with the schizophrenia concept: from the elephant to the mouse. Schizophr Res, 127(1-
3), 3-13. doi:10.1016/j.schres.2011.01.011
Kessler, H., Doyen-Waldecker, C., Hofer, C., Hoffmann, H., Traue, H. C., & Abler, B. (2011). Neural
correlates of the perception of dynamic versus static facial expressions of emotion. Psychosoc
Med, 8, Doc03. doi:10.3205/psm000072
Kessler, R. C., Chiu, W. T., Demler, O., Merikangas, K. R., & Walters, E. E. (2005). Prevalence,
severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity
Survey Replication. Arch Gen Psychiatry, 62(6), 617-627. doi:10.1001/archpsyc.62.6.617
Kilts, C. D., Egan, G., Gideon, D. A., Ely, T. D., & Hoffman, J. M. (2003). Dissociable neural
pathways are involved in the recognition of emotion in static and dynamic facial expressions.
Neuroimage, 18(1), 156-168.
Kim, H. S., Shin, N. Y., Choi, J. S., Jung, M. H., Jang, J. H., Kang, D. H., & Kwon, J. S. (2010).
Processing of facial configuration in individuals at ultra-high risk for schizophrenia.
Schizophr Res, 118(1-3), 81-87. doi:10.1016/j.schres.2010.01.003
Kim, J., Doop, M. L., Blake, R., & Park, S. (2005). Impaired visual recognition of biological motion
in schizophrenia. Schizophr Res, 77(2-3), 299-307. doi:10.1016/j.schres.2005.04.006
Kircher, T. T., Seiferth, N. Y., Plewnia, C., Baar, S., & Schwabe, R. (2007). Self-face recognition in
schizophrenia. Schizophr Res, 94(1-3), 264-272. doi:10.1016/j.schres.2007.04.029
Klimidis, S., Stuart, G. W., Minas, I. H., Copolov, D. L., & Singh, B. S. (1993). Positive and negative
symptoms in the psychoses. Re-analysis of published SAPS and SANS global ratings.
Schizophr Res, 9(1), 11-18.
Kline, J. S., Smith, J. E., & Ellis, H. C. (1992). Paranoid and nonparanoid schizophrenic processing of
facially displayed affect. J Psychiatr Res, 26(3), 169-182.
Knyazev, G. G., Slobodskoj-Plusnin, J. Y., & Bocharov, A. V. (2009). Event-related delta and theta
synchronization during explicit and implicit emotion processing. Neuroscience, 164(4), 1588-
1600. doi:10.1016/j.neuroscience.2009.09.057
Kober, H., Barrett, L. F., Joseph, J., Bliss-Moreau, E., Lindquist, K., & Wager, T. D. (2008).
Functional grouping and cortical-subcortical interactions in emotion: a meta-analysis of
neuroimaging studies. Neuroimage, 42(2), 998-1031. doi:10.1016/j.neuroimage.2008.03.059
Kohler, C. G., Bilker, W., Hagendoorn, M., Gur, R. E., & Gur, R. C. (2000). Emotion recognition
deficit in schizophrenia: association with symptomatology and cognition. Biol Psychiatry,
48(2), 127-136.
Kohler, C. G., Hoffman, L. J., Eastman, L. B., Healey, K., & Moberg, P. J. (2011). Facial emotion
perception in depression and bipolar disorder: a quantitative review. Psychiatry Res, 188(3),
303-309. doi:10.1016/j.psychres.2011.04.019
Kohler, C. G., & Martin, E. A. (2006). Emotional processing in schizophrenia. Cogn Neuropsychiatry,
11(3), 250-271. doi:10.1080/13546800500188575
Kohler, C. G., Richard, J. A., Brensinger, C. M., Borgmann-Winter, K. E., Conroy, C. G., Moberg, P.
J., . . . Calkins, M. E. (2014). Facial emotion perception differs in young persons at genetic
and clinical high-risk for psychosis. Psychiatry Res, 216(2), 206-212.
doi:10.1016/j.psychres.2014.01.023
Kohler, C. G., Turner, T. H., Bilker, W. B., Brensinger, C. M., Siegel, S. J., Kanes, S. J., . . . Gur, R.
C. (2003). Facial emotion recognition in schizophrenia: intensity effects and error pattern. Am
J Psychiatry, 160(10), 1768-1774. doi:10.1176/appi.ajp.160.10.1768
Kohler, C. G., Walker, J. B., Martin, E. A., Healey, K. M., & Moberg, P. J. (2010). Facial emotion
perception in schizophrenia: a meta-analytic review. Schizophr Bull, 36(5), 1009-1019.
doi:10.1093/schbul/sbn192
192
Korponay, C., Nitzburg, G. C., Malhotra, A. K., & DeRosse, P. (2014). Positive and Negative
Subclinical Symptoms and Mccb Performance in Non-Psychiatric Controls. Schizophr Res
Cogn, 1(4), 175-179. doi:10.1016/j.scog.2014.09.002
Kosaka, H., Omori, M., Murata, T., Iidaka, T., Yamada, H., Okada, T., . . . Wada, Y. (2002).
Differential amygdala response during facial recognition in patients with schizophrenia: an
fMRI study. Schizophr Res, 57(1), 87-95.
Kosmidis, M. H., Bozikas, V. P., Giannakou, M., Anezoulaki, D., Fantie, B. D., & Karavatos, A.
(2007). Impaired emotion perception in schizophrenia: a differential deficit. Psychiatry Res,
149(1-3), 279-284. doi:10.1016/j.psychres.2004.09.011
Krawiecka, M., Goldberg, D., & Vaughan, M. (1977). A standardized psychiatric assessment scale for
rating chronic psychotic patients. Acta Psychiatrica Scandinavica, 55(4), 299-308.
doi:10.1111/j.1600-0447.1977.tb00174.x
Kress, T., & Daum, I. (2003). Event-related potentials reflect impaired face recognition in patients
with congenital prosopagnosia. Neurosci Lett, 352(2), 133-136.
Kret, M. E., Denollet, J., Grezes, J., & de Gelder, B. (2011). The role of negative affectivity and social
inhibition in perceiving social threat: an fMRI study. Neuropsychologia, 49(5), 1187-1193.
doi:10.1016/j.neuropsychologia.2011.02.007
Kret, M. E., & Ploeger, A. (2015). Emotion processing deficits: a liability spectrum providing insight
into comorbidity of mental disorders. Neurosci Biobehav Rev, 52, 153-171.
doi:10.1016/j.neubiorev.2015.02.011
Krombholz, A., Schaefer, F., & Boucsein, W. (2007). Modification of N170 by different emotional
expression of schematic faces. Biol Psychol, 76(3), 156-162.
doi:10.1016/j.biopsycho.2007.07.004
Kucharska-Pietura, K., David, A. S., Masiak, M., & Phillips, M. L. (2005). Perception of facial and
vocal affect by people with schizophrenia in early and late stages of illness. Br J Psychiatry,
187, 523-528. doi:10.1192/bjp.187.6.523
Kuepper, R., Ceccarini, J., Lataster, J., van Os, J., van Kroonenburgh, M., van Gerven, J. M., . . .
Henquet, C. (2013). Delta-9-tetrahydrocannabinol-induced dopamine release as a function of
psychosis risk: 18F-fallypride positron emission tomography study. PLoS One, 8(7), e70378.
doi:10.1371/journal.pone.0070378
Kumari, S., Malik, M., Florival, C., Manalai, P., & Sonje, S. (2017). An Assessment of Five (PANSS,
SAPS, SANS, NSA-16, CGI-SCH) commonly used Symptoms Rating Scales in
Schizophrenia and Comparison to Newer Scales (CAINS, BNSS). J Addict Res Ther, 8(324),
2.
Kurylo, D. D., Pasternak, R., Silipo, G., Javitt, D. C., & Butler, P. D. (2007). Perceptual organization
by proximity and similarity in schizophrenia. Schizophr Res, 95(1-3), 205-214.
doi:10.1016/j.schres.2007.07.001
LaBar, K. S., Crupain, M. J., Voyvodic, J. T., & McCarthy, G. (2003). Dynamic perception of facial
affect and identity in the human brain. Cereb Cortex, 13(10), 1023-1033.
Lahera, G., Herrera, S., Fernandez, C., Bardon, M., de los Angeles, V., & Fernandez-Liria, A. (2014).
Familiarity and face emotion recognition in patients with schizophrenia. Compr Psychiatry,
55(1), 199-205. doi:10.1016/j.comppsych.2013.06.006
Lambert, M., Naber, D., Schacht, A., Wagner, T., Hundemer, H. P., Karow, A., . . . Schimmelmann,
B. G. (2008). Rates and predictors of remission and recovery during 3 years in 392 never-
treated patients with schizophrenia. Acta Psychiatr Scand, 118(3), 220-229.
doi:10.1111/j.1600-0447.2008.01213.x
Landgraf, S., Amado, I., Purkhart, R., Ries, J., Olie, J. P., & van der Meer, E. (2011). Visuo-spatial
cognition in schizophrenia: confirmation of a preference for local information processing.
Schizophr Res, 127(1-3), 163-170. doi:10.1016/j.schres.2010.11.027
Langenecker, S. A., Bieliauskas, L. A., Rapport, L. J., Zubieta, J. K., Wilde, E. A., & Berent, S.
(2005). Face emotion perception and executive functioning deficits in depression. J Clin Exp
Neuropsychol, 27(3), 320-333. doi:10.1080/13803390490490515720
Langenecker, S. A., Saunders, E. F., Kade, A. M., Ransom, M. T., & McInnis, M. G. (2010).
Intermediate: cognitive phenotypes in bipolar disorder. J Affect Disord, 122(3), 285-293.
doi:10.1016/j.jad.2009.08.018
193
Laprevote, V., Oliva, A., Delerue, C., Thomas, P., & Boucart, M. (2010). Patients with schizophrenia
are biased toward low spatial frequency to decode facial expression at a glance.
Neuropsychologia, 48(14), 4164-4168. doi:10.1016/j.neuropsychologia.2010.10.017
Laprevote, V., Oliva, A., Ternois, A. S., Schwan, R., Thomas, P., & Boucart, M. (2013). Low Spatial
Frequency Bias in Schizophrenia is Not Face Specific: When the Integration of Coarse and
Fine Information Fails. Front Psychol, 4, 248. doi:10.3389/fpsyg.2013.00248
Lawrence, A. D., Calder, A. J., McGowan, S. W., & Grasby, P. M. (2002). Selective disruption of the
recognition of facial expressions of anger. Neuroreport, 13(6), 881-884.
Lawrie, S. M., Hall, J., McIntosh, A. M., Owens, D. G., & Johnstone, E. C. (2010). The 'continuum of
psychosis': scientifically unproven and clinically impractical. Br J Psychiatry, 197(6), 423-
425. doi:10.1192/bjp.bp.109.072827
Le Masurier, M., Cowen, P. J., & Harmer, C. J. (2007). Emotional bias and waking salivary cortisol in
relatives of patients with major depression. Psychol Med, 37(3), 403-410.
doi:10.1017/S0033291706009184
Lee, A. (2012). VirtualDub version 1.10.4. [Computer software] retrieved from
http://www.virtualdub.org.
Lee, H., Simpson, G. V., Logothetis, N. K., & Rainer, G. (2005). Phase locking of single neuron
activity to theta oscillations during working memory in monkey extrastriate visual cortex.
Neuron, 45(1), 147-156. doi:10.1016/j.neuron.2004.12.025
Lee, J., Kern, R. S., Harvey, P. O., Horan, W. P., Kee, K. S., Ochsner, K., . . . Green, M. F. (2013).
An intact social cognitive process in schizophrenia: situational context effects on perception
of facial affect. Schizophr Bull, 39(3), 640-647. doi:10.1093/schbul/sbs063
Lee, J., Kwon, J. S., Shin, Y. W., Lee, K. J., & Park, S. (2007). Visual self-recognition in patients
with schizophrenia. Schizophr Res, 94(1-3), 215-220. doi:10.1016/j.schres.2007.03.032
Lee, S. H., Kim, E. Y., Kim, S., & Bae, S. M. (2010). Event-related potential patterns and gender
effects underlying facial affect processing in schizophrenia patients. Neurosci Res, 67(2),
172-180. doi:10.1016/j.neures.2010.03.001
Leitman, D. I., Foxe, J. J., Butler, P. D., Saperstein, A., Revheim, N., & Javitt, D. C. (2005). Sensory
contributions to impaired prosodic processing in schizophrenia. Biol Psychiatry, 58(1), 56-61.
doi:10.1016/j.biopsych.2005.02.034
Leitman, D. I., Wolf, D. H., Laukka, P., Ragland, J. D., Valdez, J. N., Turetsky, B. I., . . . Gur, R. C.
(2011). Not pitch perfect: sensory contributions to affective communication impairment in
schizophrenia. Biol Psychiatry, 70(7), 611-618. doi:10.1016/j.biopsych.2011.05.032
Lembke, A., & Ketter, T. A. (2002). Impaired recognition of facial emotion in mania. Am J
Psychiatry, 159(2), 302-304. doi:10.1176/appi.ajp.159.2.302
Leonards, U., & Mohr, C. (2009). Schizotypal personality traits influence idiosyncratic initiation of
saccadic face exploration. Vision Res, 49(19), 2404-2413. doi:10.1016/j.visres.2009.07.013
Leppanen, J. M. (2006). Emotional information processing in mood disorders: a review of behavioral
and neuroimaging findings. Curr Opin Psychiatry, 19(1), 34-39.
doi:10.1097/01.yco.0000191500.46411.00
Leppanen, J. M., Milders, M., Bell, J. S., Terriere, E., & Hietanen, J. K. (2004). Depression biases the
recognition of emotionally neutral faces. Psychiatry Res, 128(2), 123-133.
doi:10.1016/j.psychres.2004.05.020
Leppanen, J. M., Niehaus, D. J., Koen, L., Du Toit, E., Schoeman, R., & Emsley, R. (2006).
Emotional face processing deficit in schizophrenia: A replication study in a South African
Xhosa population. Schizophr Res, 84(2-3), 323-330. doi:10.1016/j.schres.2006.02.007
Levkovitz, Y., Lamy, D., Ternochiano, P., Treves, I., & Fennig, S. (2003). Perception of dyadic
relationship and emotional states in patients with affective disorder. J Affect Disord, 75(1),
19-28.
Lewis, S. F., & Garver, D. L. (1995). Treatment and diagnostic subtype in facial affect recognition in
schizophrenia. J Psychiatr Res, 29(1), 5-11.
Li, H., Chan, R. C., McAlonan, G. M., & Gong, Q. Y. (2010). Facial emotion processing in
schizophrenia: a meta-analysis of functional neuroimaging data. Schizophr Bull, 36(5), 1029-
1039. doi:10.1093/schbul/sbn190
194
Lichtenstein, P., Yip, B. H., Bjork, C., Pawitan, Y., Cannon, T. D., Sullivan, P. F., & Hultman, C. M.
(2009). Common genetic determinants of schizophrenia and bipolar disorder in Swedish
families: a population-based study. Lancet, 373(9659), 234-239. doi:10.1016/S0140-
6736(09)60072-6
Lincoln, T. M., & Keller, E. (2008). Delusions and hallucinations in students compared to the general
population. Psychol Psychother, 81(Pt 3), 231-235. doi:10.1348/147608308X297096
Liu, J., Blond, B. N., van Dyck, L. I., Spencer, L., Wang, F., & Blumberg, H. P. (2012). Trait and
state corticostriatal dysfunction in bipolar disorder during emotional face processing. Bipolar
Disord, 14(4), 432-441. doi:10.1111/j.1399-5618.2012.01018.x
Loughland, C. M., Williams, L. M., & Gordon, E. (2002a). Schizophrenia and affective disorder show
different visual scanning behavior for faces: a trait versus state-based distinction? Biol
Psychiatry, 52(4), 338-348.
Loughland, C. M., Williams, L. M., & Gordon, E. (2002b). Visual scanpaths to positive and negative
facial emotions in an outpatient schizophrenia sample. Schizophr Res, 55(1-2), 159-170.
Loula, F., Prasad, S., Harber, K., & Shiffrar, M. (2005). Recognizing people from their movement. J
Exp Psychol Hum Percept Perform, 31(1), 210-220. doi:10.1037/0096-1523.31.1.210
Macmillan, N. A., & Creelman, C. D. (1991). Detection Theory: A User's Guide: Cambridge
University Press.
Macrae, C. N., & Lewis, H. L. (2002). Do I know you? Processing orientation and face recognition.
Psychol Sci, 13(2), 194-196. doi:10.1111/1467-9280.00436
Macrae, C. N., & Martin, D. (2007). A boy primed Sue: feature‐based processing and person
construal. European Journal of Social Psychology, 37(5), 793-805.
Mah, L., & Pollock, B. G. (2010). Emotional processing deficits in late-life depression. Am J Geriatr
Psychiatry, 18(7), 652-656.
Maher, S., Ekstrom, T., Holt, D., Ongur, D., & Chen, Y. (2016). The Core Brain Region for Face
Processing in Schizophrenia Lacks Face Selectivity. Schizophr Bull, 42(3), 666-674.
doi:10.1093/schbul/sbv140
Maher, S., Mashhoon, Y., Ekstrom, T., Lukas, S., & Chen, Y. (2016). Deficient cortical face-sensitive
N170 responses and basic visual processing in schizophrenia. Schizophr Res, 170(1), 87-94.
doi:10.1016/j.schres.2015.12.005
Mak, A. K., Hu, Z. G., Zhang, J. X., Xiao, Z. W., & Lee, T. M. (2009). Neural correlates of regulation
of positive and negative emotions: an fmri study. Neurosci Lett, 457(2), 101-106.
doi:10.1016/j.neulet.2009.03.094
Malaspina, D., Owen, M. J., Heckers, S., Tandon, R., Bustillo, J., Schultz, S., . . . Carpenter, W.
(2013). Schizoaffective Disorder in the DSM-5. Schizophr Res, 150(1), 21-25.
doi:10.1016/j.schres.2013.04.026
Malla, A., Norman, R., Bechard-Evans, L., Schmitz, N., Manchanda, R., & Cassidy, C. (2008).
Factors influencing relapse during a 2-year follow-up of first-episode psychosis in a
specialized early intervention service. Psychol Med, 38(11), 1585-1593.
doi:10.1017/S0033291707002656
Mandal, M. K., Jain, A., Haque-Nizamie, S., Weiss, U., & Schneider, F. (1999). Generality and
specificity of emotion-recognition deficit in schizophrenic patients with positive and negative
symptoms. Psychiatry Res, 87(1), 39-46.
Mandal, M. K., Pandey, R., & Prasad, A. B. (1998). Facial expressions of emotions and
schizophrenia: a review. Schizophr Bull, 24(3), 399-412.
Manor, B. R., Gordon, E., Williams, L. M., Rennie, C. J., Bahramali, H., Latimer, C. R., . . . Meares,
R. A. (1999). Eye movements reflect impaired face processing in patients with schizophrenia.
Biol Psychiatry, 46(7), 963-969.
Martin, D., & Macrae, C. N. (2007). A face with a cue: Exploring the inevitability of person
categorization. European Journal of Social Psychology, 37(5), 806-816.
Martin, D., Slessor, G., Allen, R., Phillips, L. H., & Darling, S. (2012). Processing orientation and
emotion recognition. Emotion, 12(1), 39-43. doi:10.1037/a0024775
Martin, F., Baudouin, J. Y., Tiberghien, G., & Franck, N. (2005). Processing emotional expression
and facial identity in schizophrenia. Psychiatry Res, 134(1), 43-53.
doi:10.1016/j.psychres.2003.12.031
195
Marwick, K., & Hall, J. (2008). Social cognition in schizophrenia: a review of face processing. Br
Med Bull, 88(1), 43-58. doi:10.1093/bmb/ldn035
Mason, O., & Claridge, G. (1999). Individual differences in schizotypy and reduced asymmetry using
the chimeric faces task. Cogn Neuropsychiatry, 4(4), 289-301.
doi:10.1080/135468099395846
Mason, O., & Claridge, G. (2006). The Oxford-Liverpool Inventory of Feelings and Experiences (O-
LIFE): further description and extended norms. Schizophr Res, 82(2-3), 203-211.
doi:10.1016/j.schres.2005.12.845
Mason, O., Claridge, G., & Jackson, M. (1995). New scales for the assessment of schizotypy.
Personality and Individual Differences, 18(1), 7-13. doi:http://dx.doi.org/10.1016/0191-
8869(94)00132-C
Matheson, S. L., Shepherd, A. M., Pinchbeck, R. M., Laurens, K. R., & Carr, V. J. (2013). Childhood
adversity in schizophrenia: a systematic meta-analysis. Psychol Med, 43(2), 225-238.
doi:10.1017/S0033291712000785
Matsumoto, D. (2001). Culture and Emotion. In D. Matsumoto (Ed.), The handbook of culture and
psychology (pp. 171-194). New York: Oxford University Press.
Matsumoto, D., Keltner, D., Shiota, M. N., Frank, M. G., & O'Sullivan, M. (2008). What's in a face?
Facial expressions as signals of discrete emotions. In M. Lewis, J. M. Haviland, & L.
Feldman Barrett (Eds.), Handbook of emotions (pp. 211-234). New York: Guildford Press.
Matsumoto, D., & Willingham, B. (2009). Spontaneous facial expressions of emotion of congenitally
and noncongenitally blind individuals. J Pers Soc Psychol, 96(1), 1-10.
doi:10.1037/a0014037
Matsumoto, K., Samson, G. T., O'Daly, O. D., Tracy, D. K., Patel, A. D., & Shergill, S. S. (2006).
Prosodic discrimination in patients with schizophrenia. Br J Psychiatry, 189, 180-181.
doi:10.1192/bjp.bp.105.009332
Maurer, D., Grand, R. L., & Mondloch, C. J. (2002). The many faces of configural processing. Trends
Cogn Sci, 6(6), 255-260.
McBain, R., Norton, D., & Chen, Y. (2010). A female advantage in basic face recognition is absent in
schizophrenia. Psychiatry Res, 177(1-2), 12-17. doi:10.1016/j.psychres.2009.02.005
McCleery, A., Lee, J., Joshi, A., Wynn, J. K., Hellemann, G. S., & Green, M. F. (2015). Meta-
analysis of face processing event-related potentials in schizophrenia. Biol Psychiatry, 77(2),
116-126. doi:10.1016/j.biopsych.2014.04.015
McGorry, P. D. (1989). The Royal Park Multidiagnostic Instrument for Psychosis (RPMIP) : a
comprehensive assessment procedure for the acute psychotic episode : glossary and
guidelines for administration: National Health and Medical Research Council, Schizophrenia
Research Unit.
McGorry, P. D., Bell, R. C., Dudgeon, P. L., & Jackson, H. J. (1998). The dimensional structure of
first episode psychosis: an exploratory factor analysis. Psychol Med, 28(4), 935-947.
McGrath, J., Saha, S., Chant, D., & Welham, J. (2008). Schizophrenia: a concise overview of
incidence, prevalence, and mortality. Epidemiol Rev, 30, 67-76. doi:10.1093/epirev/mxn001
McGrath, J. J., Saha, S., Al-Hamzawi, A., Andrade, L., Benjet, C., Bromet, E. J., . . . Kessler, R. C.
(2016). The Bidirectional Associations Between Psychotic Experiences and DSM-IV Mental
Disorders. Am J Psychiatry, 173(10), 997-1006. doi:10.1176/appi.ajp.2016.15101293
McKone, E. (2009). Holistic processing for faces operates over a wide range of sizes but is strongest
at identification rather than conversational distances. Vision Res, 49(2), 268-283.
doi:10.1016/j.visres.2008.10.020
McKone, E., Aimola Davies, A., Fernando, D., Aalders, R., Leung, H., Wickramariyaratne, T., &
Platow, M. J. (2010). Asia has the global advantage: Race and visual attention. Vision Res,
50(16), 1540-1549. doi:10.1016/j.visres.2010.05.010
McKone, E., & Robbins, R. (2011). Are faces special? In A. J. Calder, G. Rhodes, M. H. Johnson, &
J. V. Haxby (Eds.), The Oxford Handbook of Face Perception (pp. 149-176). New York:
Oxford University Press.
McKone, E., & Yovel, G. (2009). Why does picture-plane inversion sometimes dissociate perception
of features and spacing in faces, and sometimes not? Toward a new theory of holistic
processing. Psychon Bull Rev, 16(5), 778-797. doi:10.3758/PBR.16.5.778
196
Meehl, P. E. (1970). Nuisance variables and the ex post facto design. In M. Radner & S. Winocur
(Eds.), Minnesota studies in the philosophy of science (pp. 373-402). Minneapolis: University
of Minnesota Press.
Meehl, P. E. (1990). Toward an integrated theory of schizotaxia, schizotypy, and schizophrenia.
Journal of Personality Disorders, 4(1), 1-99. doi:10.1521/pedi.1990.4.1.1
Megreya, A. M. (2016). Face perception in schizophrenia: a specific deficit. Cogn Neuropsychiatry,
21(1), 60-72. doi:10.1080/13546805.2015.1133407
Mendoza, R., Cabral-Calderin, Y., Dominguez, M., Garcia, A., Borrego, M., Caballero, A., . . . Reyes,
M. M. (2011). Impairment of emotional expression recognition in schizophrenia: a Cuban
familial association study. Psychiatry Res, 185(1-2), 44-48.
doi:10.1016/j.psychres.2009.10.006
Mesholam-Gately, R. I., Giuliano, A. J., Goff, K. P., Faraone, S. V., & Seidman, L. J. (2009).
Neurocognition in first-episode schizophrenia: a meta-analytic review. Neuropsychology,
23(3), 315-336. doi:10.1037/a0014708
Meyer, M. B., & Kurtz, M. M. (2009). Elementary neurocognitive function, facial affect recognition
and social-skills in schizophrenia. Schizophr Res, 110(1-3), 173-179.
doi:10.1016/j.schres.2009.03.015
Michalopoulou, P. G., Surguladze, S., Morley, L. A., Giampietro, V. P., Murray, R. M., & Shergill, S.
S. (2008). Facial fear processing and psychotic symptoms in schizophrenia: functional
magnetic resonance imaging study. Br J Psychiatry, 192(3), 191-196.
doi:10.1192/bjp.bp.106.032649
Mikhailova, E. S., Vladimirova, T. V., Iznak, A. F., Tsusulkovskaya, E. J., & Sushko, N. V. (1996).
Abnormal recognition of facial expression of emotions in depressed patients with major
depression disorder and schizotypal personality disorder. Biol Psychiatry, 40(8), 697-705.
doi:10.1016/0006-3223(96)00032-7
Miller, B. J. (2008). A review of second-generation antipsychotic discontinuation in first-episode
psychosis. J Psychiatr Pract, 14(5), 289-300. doi:10.1097/01.pra.0000336756.65308.83
Missonnier, P., Herrmann, F. R., Zanello, A., Badan Ba, M., Curtis, L., Canovas, D., . . . Merlo, M. C.
(2012). Event-related potentials and changes of brain rhythm oscillations during working
memory activation in patients with first-episode psychosis. J Psychiatry Neurosci, 37(2), 95-
105. doi:10.1503/jpn.110033
Mitchell, A. E., Dickens, G. L., & Picchioni, M. M. (2014). Facial emotion processing in borderline
personality disorder: a systematic review and meta-analysis. Neuropsychol Rev, 24(2), 166-
184. doi:10.1007/s11065-014-9254-9
Mitchell, P. B., Frankland, A., Hadzi-Pavlovic, D., Roberts, G., Corry, J., Wright, A., . . . Breakspear,
M. (2011). Comparison of depressive episodes in bipolar disorder and in major depressive
disorder within bipolar disorder pedigrees. Br J Psychiatry, 199(4), 303-309.
doi:10.1192/bjp.bp.110.088823
Mitchell, R. L., Elliott, R., Barry, M., Cruttenden, A., & Woodruff, P. W. (2004). Neural response to
emotional prosody in schizophrenia and in bipolar affective disorder. Br J Psychiatry, 184,
223-230.
Monk, C. S., Telzer, E. H., Mogg, K., Bradley, B. P., Mai, X., Louro, H. M., . . . Pine, D. S. (2008).
Amygdala and ventrolateral prefrontal cortex activation to masked angry faces in children and
adolescents with generalized anxiety disorder. Arch Gen Psychiatry, 65(5), 568-576.
doi:10.1001/archpsyc.65.5.568
Montagne, B., Kessels, R. P., De Haan, E. H., & Perrett, D. I. (2007). The Emotion Recognition Task:
a paradigm to measure the perception of facial emotional expressions at different intensities.
Percept Mot Skills, 104(2), 589-598. doi:10.2466/pms.104.2.589-598
Morgan, V. A., Waterreus, A., Jablensky, A., Mackinnon, A., McGrath, J. J., Carr, V., . . . Saw, S.
(2012). People living with psychotic illness in 2010: the second Australian national survey of
psychosis. Aust N Z J Psychiatry, 46(8), 735-752. doi:10.1177/0004867412449877
Morris, R. W., Weickert, C. S., & Loughland, C. M. (2009). Emotional face processing in
schizophrenia. Curr Opin Psychiatry, 22(2), 140-146. doi:10.1097/YCO.0b013e328324f895
Mortimer, A. M. (2007). Symptom rating scales and outcome in schizophrenia. Br J Psychiatry Suppl,
50, s7-14.
197
Mothersill, O., Morris, D. W., Kelly, S., Rose, E. J., Bokde, A., Reilly, R., . . . Donohoe, G. (2014).
Altered medial prefrontal activity during dynamic face processing in schizophrenia spectrum
patients. Schizophr Res, 157(1-3), 225-230. doi:10.1016/j.schres.2014.05.023
Mueser, K. T., Doonan, R., Penn, D. L., Blanchard, J. J., Bellack, A. S., Nishith, P., & DeLeon, J.
(1996). Emotion recognition and social competence in chronic schizophrenia. J Abnorm
Psychol, 105(2), 271-275.
Mukherjee, P., Whalley, H. C., McKirdy, J. W., McIntosh, A. M., Johnstone, E. C., Lawrie, S. M., &
Hall, J. (2012). Lower effective connectivity between amygdala and parietal regions in
response to fearful faces in schizophrenia. Schizophr Res, 134(2-3), 118-124.
doi:10.1016/j.schres.2011.09.033
Murphy, J., Ipser, A., Gaigg, S. B., & Cook, R. (2015). Exemplar variance supports robust learning of
facial identity. J Exp Psychol Hum Percept Perform, 41(3), 577-581.
doi:10.1037/xhp0000049
Murray, R. M., Paparelli, A., Morrison, P. D., Marconi, A., & Di Forti, M. (2013). What can we learn
about schizophrenia from studying the human model, drug-induced psychosis? Am J Med
Genet B Neuropsychiatr Genet, 162B(7), 661-670. doi:10.1002/ajmg.b.32177
Najt, P., Bayer, U., & Hausmann, M. (2012). Atypical lateralisation in emotional prosody in men with
schizotypy. Laterality: Asymmetries of Body, Brain and Cognition, 17(5), 533-548.
doi:10.1080/1357650X.2011.586702
Navon, D. (1977). Forest before trees: The precedence of global features in visual perception.
Cognitive psychology, 9(3), 353-383.
Nguyen, V. T., & Cunnington, R. (2014). The superior temporal sulcus and the N170 during face
processing: single trial analysis of concurrent EEG-fMRI. Neuroimage, 86, 492-502.
doi:10.1016/j.neuroimage.2013.10.047
Niemi-Pynttari, J. A., Sund, R., Putkonen, H., Vorma, H., Wahlbeck, K., & Pirkola, S. P. (2013).
Substance-induced psychoses converting into schizophrenia: a register-based study of 18,478
Finnish inpatient cases. J Clin Psychiatry, 74(1), e94-99. doi:10.4088/JCP.12m07822
Norton, D., McBain, R., Holt, D. J., Ongur, D., & Chen, Y. (2009). Association of impaired facial
affect recognition with basic facial and visual processing deficits in schizophrenia. Biol
Psychiatry, 65(12), 1094-1098. doi:10.1016/j.biopsych.2009.01.026
Noton, D., & Stark, L. (1971). Eye movements and visual perception. Sci Am, 224(6), 35-43.
O'Toole, A. J., Deffenbacher, K. A., Valentin, D., McKee, K., Huff, D., & Abdi, H. (1998). The
perception of face gender: the role of stimulus structure in recognition and classification.
Mem Cognit, 26(1), 146-160.
O'Toole, A. J., Phillips, P. J., Weimer, S., Roark, D. A., Ayyad, J., Barwick, R., & Dunlop, J. (2011).
Recognizing people from dynamic and static faces and bodies: dissecting identity with a
fusion approach. Vision Res, 51(1), 74-83. doi:10.1016/j.visres.2010.09.035
Obayashi, C., Nakashima, T., Onitsuka, T., Maekawa, T., Hirano, Y., Hirano, S., . . . Tobimatsu, S.
(2009). Decreased spatial frequency sensitivities for processing faces in male patients with
chronic schizophrenia. Clin Neurophysiol, 120(8), 1525-1533.
doi:10.1016/j.clinph.2009.06.016
Obayashi, S., Matsushima, E., Okubo, Y., Ohkura, T., Kojima, T., & Kakuma, T. (2001). Relationship
between exploratory eye movements and clinical course in schizophrenic patients. Eur Arch
Psychiatry Clin Neurosci, 251(5), 211-216.
Oldfield, R. C. (1971). The assessment and analysis of handedness: the Edinburgh inventory.
Neuropsychologia, 9(1), 97-113.
Onitsuka, T., Niznikiewicz, M. A., Spencer, K. M., Frumin, M., Kuroki, N., Lucia, L. C., . . .
McCarley, R. W. (2006). Functional and structural deficits in brain regions subserving face
perception in schizophrenia. Am J Psychiatry, 163(3), 455-462.
doi:10.1176/appi.ajp.163.3.455
Overall, J. E., & Gorham, D. R. (1962). The Brief Psychiatric Rating Scale. Psychological Reports,
10(3), 799-812. doi:10.2466/pr0.1962.10.3.799
Palermo, R., & Rhodes, G. (2007). Are you always on my mind? A review of how face perception
and attention interact. Neuropsychologia, 45(1), 75-92.
doi:10.1016/j.neuropsychologia.2006.04.025
198
Palermo, R., Willis, M. L., Rivolta, D., McKone, E., Wilson, C. E., & Calder, A. J. (2011). Impaired
holistic coding of facial expression and facial identity in congenital prosopagnosia.
Neuropsychologia, 49(5), 1226-1235. doi:10.1016/j.neuropsychologia.2011.02.021
Pankow, A., Friedel, E., Sterzer, P., Seiferth, N., Walter, H., Heinz, A., & Schlagenhauf, F. (2013).
Altered amygdala activation in schizophrenia patients during emotion processing. Schizophr
Res, 150(1), 101-106. doi:10.1016/j.schres.2013.07.015
Pantic, M., Valstar, M., Rademaker, R., & Maat, L. (2005, 6-8 July 2005). Web-based database for
facial expression analysis. Paper presented at the 2005 IEEE International Conference on
Multimedia and Expo.
Paparelli, A., Di Forti, M., Morrison, P. D., & Murray, R. M. (2011). Drug-induced psychosis: how to
avoid star gazing in schizophrenia research by looking at more obvious sources of light. Front
Behav Neurosci, 5, 1. doi:10.3389/fnbeh.2011.00001
Patterson, J. C., 2nd, Ungerleider, L. G., & Bandettini, P. A. (2002). Task-independent functional
brain activity correlation with skin conductance changes: an fMRI study. Neuroimage, 17(4),
1797-1806.
Peluso, M. A., Glahn, D. C., Matsuo, K., Monkul, E. S., Najt, P., Zamarripa, F., . . . Soares, J. C.
(2009). Amygdala hyperactivation in untreated depressed individuals. Psychiatry Res, 173(2),
158-161. doi:10.1016/j.pscychresns.2009.03.006
Penn, D. L., Combs, D. R., Ritchie, M., Francis, J., Cassisi, J., Morris, S., & Townsend, M. (2000).
Emotion recognition in schizophrenia: further investigation of generalized versus specific
deficit models. J Abnorm Psychol, 109(3), 512-516.
Perfect, T. J. (2003). Local processing bias impairs lineup performance. Psychol Rep, 93(2), 393-394.
doi:10.2466/pr0.2003.93.2.393
Perugi, G., Fornaro, M., & Akiskal, H. S. (2011). Are atypical depression, borderline personality
disorder and bipolar II disorder overlapping manifestations of a common cyclothymic
diathesis? World Psychiatry, 10(1), 45-51.
Peterman, J., Darke, H., Carter, O., Sundram, S., & Park, S. (2017). M124. Perception of Socio-
Emotional Information in Dynamic Gait in Inpatients With Schizophrenia Spectrum
Disorders (Vol. 43).
Phan, K. L., Wager, T., Taylor, S. F., & Liberzon, I. (2002). Functional neuroanatomy of emotion: a
meta-analysis of emotion activation studies in PET and fMRI. Neuroimage, 16(2), 331-348.
doi:10.1006/nimg.2002.1087
Phillips, L. K., & Seidman, L. J. (2008). Emotion processing in persons at risk for schizophrenia.
Schizophr Bull, 34(5), 888-903. doi:10.1093/schbul/sbn085
Phillips, M. L. (2011). Facial expression perception in schizophrenia and mood disorders. In G.
Rhodes, A. Calder, M. Johnson, & J. V. Haxby (Eds.), Oxford handbook of face perception
(pp. 793-812). New York: Oxford University Press.
Phillips, M. L., & David, A. S. (1998). Abnormal visual scan paths: a psychophysiological marker of
delusions in schizophrenia. Schizophr Res, 29(3), 235-245.
Phillips, M. L., Williams, L., Senior, C., Bullmore, E. T., Brammer, M. J., Andrew, C., . . . David, A.
S. (1999). A differential neural response to threatening and non-threatening negative facial
expressions in paranoid and non-paranoid schizophrenics. Psychiatry Res, 92(1), 11-31.
Pierce, L. J., Scott, L. S., Boddington, S., Droucker, D., Curran, T., & Tanaka, J. W. (2011). The n250
brain potential to personally familiar and newly learned faces and objects. Front Hum
Neurosci, 5, 111. doi:10.3389/fnhum.2011.00111
Pinheiro, A. P., Del Re, E., Mezin, J., Nestor, P. G., Rauber, A., McCarley, R. W., . . . Niznikiewicz,
M. A. (2013). Sensory-based and higher-order operations contribute to abnormal emotional
prosody processing in schizophrenia: an electrophysiological investigation. Psychol Med,
43(3), 603-618. doi:10.1017/S003329171200133X
Pinheiro, A. P., Rezaii, N., Rauber, A., Liu, T., Nestor, P. G., McCarley, R. W., . . . Niznikiewicz, M.
A. (2014). Abnormalities in the processing of emotional prosody from single words in
schizophrenia. Schizophr Res, 152(1), 235-241. doi:10.1016/j.schres.2013.10.042
Pinkham, A. E., Brensinger, C., Kohler, C., Gur, R. E., & Gur, R. C. (2011). Actively paranoid
patients with schizophrenia over attribute anger to neutral faces. Schizophr Res, 125(2-3),
174-178. doi:10.1016/j.schres.2010.11.006
199
Pinkham, A. E., Hopfinger, J. B., Pelphrey, K. A., Piven, J., & Penn, D. L. (2008). Neural bases for
impaired social cognition in schizophrenia and autism spectrum disorders. Schizophr Res,
99(1-3), 164-175. doi:10.1016/j.schres.2007.10.024
Pinkham, A. E., Penn, D. L., Perkins, D. O., & Lieberman, J. (2003). Implications for the neural basis
of social cognition for the study of schizophrenia. Am J Psychiatry, 160(5), 815-824.
doi:10.1176/appi.ajp.160.5.815
Pinsk, M. A., DeSimone, K., Moore, T., Gross, C. G., & Kastner, S. (2005). Representations of faces
and body parts in macaque temporal cortex: a functional MRI study. Proc Natl Acad Sci U S
A, 102(19), 6996-7001. doi:10.1073/pnas.0502605102
Piskulic, D., Olver, J. S., Norman, T. R., & Maruff, P. (2007). Behavioural studies of spatial working
memory dysfunction in schizophrenia: a quantitative literature review. Psychiatry Res,
150(2), 111-121. doi:10.1016/j.psychres.2006.03.018
Place, E. J., & Gilmore, G. C. (1980). Perceptual organization in schizophrenia. J Abnorm Psychol,
89(3), 409-418.
Plana, I., Lavoie, M. A., Battaglia, M., & Achim, A. M. (2014). A meta-analysis and scoping review
of social cognition performance in social phobia, posttraumatic stress disorder and other
anxiety disorders. J Anxiety Disord, 28(2), 169-177. doi:10.1016/j.janxdis.2013.09.005
Poirel, N., Brazo, P., Turbelin, M. R., Lecardeur, L., Simon, G., Houde, O., . . . Dollfus, S. (2010).
Meaningfulness and global-local processing in schizophrenia. Neuropsychologia, 48(10),
3062-3068. doi:10.1016/j.neuropsychologia.2010.06.017
Pomarol-Clotet, E., Hynes, F., Ashwin, C., Bullmore, E. T., McKenna, P. J., & Laws, K. R. (2010).
Facial emotion processing in schizophrenia: a non-specific neuropsychological deficit?
Psychol Med, 40(6), 911-919. doi:10.1017/S0033291709991309
Poole, J. H., Tobias, F. C., & Vinogradov, S. (2000). The functional relevance of affect recognition
errors in schizophrenia. J Int Neuropsychol Soc, 6(6), 649-658.
Poreh, A. M., Whitman, R. D., Weber, M., & Ross, T. (1994). Facial recognition in hypothetically
schizotypic college students. The role of generalized poor performance. J Nerv Ment Dis,
182(9), 503-507.
Poulton, R., Caspi, A., Moffitt, T. E., Cannon, M., Murray, R., & Harrington, H. (2000). Children's
self-reported psychotic symptoms and adult schizophreniform disorder: a 15-year longitudinal
study. Arch Gen Psychiatry, 57(11), 1053-1058.
Puce, A., Allison, T., Asgari, M., Gore, J. C., & McCarthy, G. (1996). Differential sensitivity of
human visual cortex to faces, letterstrings, and textures: a functional magnetic resonance
imaging study. J Neurosci, 16(16), 5205-5215.
Purcell, S. M., Wray, N. R., Stone, J. L., Visscher, P. M., O'Donovan, M. C., Sullivan, P. F., & Sklar,
P. (2009). Common polygenic variation contributes to risk of schizophrenia and bipolar
disorder. Nature, 460(7256), 748-752. doi:10.1038/nature08185
Quintana, J., Lee, J., Marcus, M., Kee, K., Wong, T., & Yerevanian, A. (2011). Brain dysfunctions
during facial discrimination in schizophrenia: selective association to affect decoding.
Psychiatry Res, 191(1), 44-50. doi:10.1016/j.pscychresns.2010.09.005
Quintana, J., Wong, T., Ortiz-Portillo, E., Marder, S. R., & Mazziotta, J. C. (2003). Right lateral
fusiform gyrus dysfunction during facial information processing in schizophrenia. Biol
Psychiatry, 53(12), 1099-1112.
Rabinowicz, E. F., Opler, L. A., Owen, D. R., & Knight, R. A. (1996). Dot Enumeration Perceptual
Organization Task (DEPOT): evidence for a short-term visual memory deficit in
schizophrenia. J Abnorm Psychol, 105(3), 336-348.
Radua, J., Phillips, M. L., Russell, T., Lawrence, N., Marshall, N., Kalidindi, S., . . . Surguladze, S. A.
(2010). Neural response to specific components of fearful faces in healthy and schizophrenic
adults. Neuroimage, 49(1), 939-946. doi:10.1016/j.neuroimage.2009.08.030
Raine, A. (1991). The SPQ: a scale for the assessment of schizotypal personality based on DSM-III-R
criteria. Schizophr Bull, 17(4), 555-564.
Rajimehr, R., Young, J. C., & Tootell, R. B. (2009). An anterior temporal face patch in human cortex,
predicted by macaque maps. Proc Natl Acad Sci U S A, 106(6), 1995-2000.
doi:10.1073/pnas.0807304106
200
Ramos-Loyo, J., Gonzalez-Garrido, A. A., Sanchez-Loyo, L. M., Medina, V., & Basar-Eroglu, C.
(2009). Event-related potentials and event-related oscillations during identity and facial
emotional processing in schizophrenia. Int J Psychophysiol, 71(1), 84-90.
doi:10.1016/j.ijpsycho.2008.07.008
Ramsay, I. S., & MacDonald, A. W., 3rd. (2015). Brain Correlates of Cognitive Remediation in
Schizophrenia: Activation Likelihood Analysis Shows Preliminary Evidence of Neural Target
Engagement. Schizophr Bull, 41(6), 1276-1284. doi:10.1093/schbul/sbv025
Rauch, A. V., Reker, M., Ohrmann, P., Pedersen, A., Bauer, J., Dannlowski, U., . . . Suslow, T.
(2010). Increased amygdala activation during automatic processing of facial emotion in
schizophrenia. Psychiatry Res, 182(3), 200-206. doi:10.1016/j.pscychresns.2010.03.005
Rawlings, D., & Goldberg, M. (2001). Correlating a measure of sustained attention with a multi-
dimensional measure of schizotypal traits. Personality and Individual Differences, 31(3), 421-
431. doi:http://dx.doi.org/10.1016/S0191-8869(00)00147-1
Recio, G., Sommer, W., & Schacht, A. (2011). Electrophysiological correlates of perceiving and
evaluating static and dynamic facial emotional expressions. Brain Res, 1376, 66-75.
doi:10.1016/j.brainres.2010.12.041
Redfern, A. S., & Benton, C. P. (2017a). Expression Dependence in the Perception of Facial Identity.
Iperception, 8(3), 2041669517710663. doi:10.1177/2041669517710663
Redfern, A. S., & Benton, C. P. (2017b). Expressive Faces Confuse Identity. Iperception, 8(5),
2041669517731115. doi:10.1177/2041669517731115
Redfern, A. S., & Benton, C. P. (2018). Representation of facial identity includes expression
variability. Vision Res. doi:10.1016/j.visres.2018.05.004
Regenbogen, C., Kellermann, T., Seubert, J., Schneider, D. A., Gur, R. E., Derntl, B., . . . Habel, U.
(2015). Neural responses to dynamic multimodal stimuli and pathology-specific impairments
of social cognition in schizophrenia and depression. British Journal of Psychiatry, 206(3),
198-205. doi:10.1192/bjp.bp.113.143040
Reinhart, R. M., Heitz, R. P., Purcell, B. A., Weigand, P. K., Schall, J. D., & Woodman, G. F. (2012).
Homologous mechanisms of visuospatial working memory maintenance in macaque and
human: properties and sources. J Neurosci, 32(22), 7711-7722.
doi:10.1523/JNEUROSCI.0215-12.2012
Reininghaus, U., Bohnke, J. R., Hosang, G., Farmer, A., Burns, T., McGuffin, P., & Bentall, R. P.
(2016). Evaluation of the validity and utility of a transdiagnostic psychosis dimension
encompassing schizophrenia and bipolar disorder. Br J Psychiatry, 209(2), 107-113.
doi:10.1192/bjp.bp.115.167882
Rice, A., Phillips, P. J., Natu, V., An, X., & O'Toole, A. J. (2013). Unaware person recognition from
the body when face identification fails. Psychol Sci, 24(11), 2235-2243.
doi:10.1177/0956797613492986
Richler, J. J., Cheung, O. S., & Gauthier, I. (2011). Holistic processing predicts face recognition.
Psychol Sci, 22(4), 464-471. doi:10.1177/0956797611401753
Ridderinkhof, K. R., Ullsperger, M., Crone, E. A., & Nieuwenhuis, S. (2004). The role of the medial
frontal cortex in cognitive control. Science, 306(5695), 443-447.
doi:10.1126/science.1100301
Riddoch, M. J., Johnston, R. A., Bracewell, R. M., Boutsen, L., & Humphreys, G. W. (2008). Are
faces special? A case of pure prosopagnosia. Cogn Neuropsychol, 25(1), 3-26.
doi:10.1080/02643290801920113
Rief, W. (1991). Visual perceptual organization in schizophrenic patients. Br J Clin Psychol, 30 ( Pt
4), 359-366.
Ritchie, K. L., & Burton, A. M. (2017). Learning faces from variability. Q J Exp Psychol (Hove),
70(5), 897-905. doi:10.1080/17470218.2015.1136656
Rock, P. L., Goodwin, G. M., & Harmer, C. J. (2010). The common adolescent bipolar phenotype
shows positive biases in emotional processing. Bipolar Disord, 12(6), 606-615.
doi:10.1111/j.1399-5618.2010.00859.x
Roddy, S., Tiedt, L., Kelleher, I., Clarke, M. C., Murphy, J., Rawdon, C., . . . Cannon, M. (2012).
Facial emotion recognition in adolescents with psychotic-like experiences: a school-based
201
sample from the general population. Psychol Med, 42(10), 2157-2166.
doi:10.1017/S0033291712000311
Rosse, R. B., Schwartz, B. L., Johri, S., & Deutsch, S. I. (1998). Visual scanning of faces correlates
with schizophrenia symptomatology. Prog Neuropsychopharmacol Biol Psychiatry, 22(6),
971-979.
Rossell, S. L., Van Rheenen, T. E., Groot, C., Gogos, A., O'Regan, A., & Joshua, N. R. (2013).
Investigating affective prosody in psychosis: a study using the Comprehensive Affective
Testing System. Psychiatry Res, 210(3), 896-900. doi:10.1016/j.psychres.2013.07.037
Rossion, B. (2008). Picture-plane inversion leads to qualitative changes of face perception. Acta
Psychol (Amst), 128(2), 274-289. doi:10.1016/j.actpsy.2008.02.003
Rossion, B., & Caharel, S. (2011). ERP evidence for the speed of face categorization in the human
brain: Disentangling the contribution of low-level visual cues from face perception. Vision
Res, 51(12), 1297-1311. doi:10.1016/j.visres.2011.04.003
Rossion, B., Delvenne, J. F., Debatisse, D., Goffaux, V., Bruyer, R., Crommelinck, M., & Guerit, J.
M. (1999). Spatio-temporal localization of the face inversion effect: an event-related
potentials study. Biol Psychol, 50(3), 173-189.
Rossler, W., Riecher-Rossler, A., Angst, J., Murray, R., Gamma, A., Eich, D., . . . Gross, V. A.
(2007). Psychotic experiences in the general population: a twenty-year prospective
community study. Schizophr Res, 92(1-3), 1-14. doi:10.1016/j.schres.2007.01.002
Rotshtein, P., Malach, R., Hadar, U., Graif, M., & Hendler, T. (2001). Feeling or features: different
sensitivity to emotion in high-order visual cortex and amygdala. Neuron, 32(4), 747-757.
Rubinow, D. R., & Post, R. M. (1992). Impaired recognition of affect in facial expression in
depressed patients. Biol Psychiatry, 31(9), 947-953.
Ruocco, A. C., Amirthavasagam, S., Choi-Kain, L. W., & McMain, S. F. (2013). Neural correlates of
negative emotionality in borderline personality disorder: an activation-likelihood-estimation
meta-analysis. Biol Psychiatry, 73(2), 153-160. doi:10.1016/j.biopsych.2012.07.014
Ruocco, A. C., Reilly, J. L., Rubin, L. H., Daros, A. R., Gershon, E. S., Tamminga, C. A., . . .
Sweeney, J. A. (2014). Emotion recognition deficits in schizophrenia-spectrum disorders and
psychotic bipolar disorder: Findings from the Bipolar-Schizophrenia Network on Intermediate
Phenotypes (B-SNIP) study. Schizophr Res, 158(1-3), 105-112.
doi:10.1016/j.schres.2014.07.001
Russell, T. A., Reynaud, E., Kucharska-Pietura, K., Ecker, C., Benson, P. J., Zelaya, F., . . . Phillips,
M. L. (2007). Neural responses to dynamic expressions of fear in schizophrenia.
Neuropsychologia, 45(1), 107-123. doi:10.1016/j.neuropsychologia.2006.04.026
Rymarczyk, K., Biele, C., Grabowska, A., & Majczynski, H. (2011). EMG activity in response to
static and dynamic facial expressions. Int J Psychophysiol, 79(2), 330-333.
doi:10.1016/j.ijpsycho.2010.11.001
Sachs, G., Moser, E., Katschnig, H., Gur, R. C., Gur, R. E., & Erfurth, A. (2012). No amygdala
attenuation in schizophrenic patients treated with atypical antipsychotics. Psychiatry Res,
202(2), 168-171. doi:10.1016/j.pscychresns.2012.02.009
Sachs, G., Steger-Wuchse, D., Kryspin-Exner, I., Gur, R. C., & Katschnig, H. (2004). Facial
recognition deficits and cognition in schizophrenia. Schizophr Res, 68(1), 27-35.
doi:10.1016/S0920-9964(03)00131-2
Sachse, M., Schlitt, S., Hainz, D., Ciaramidaro, A., Walter, H., Poustka, F., . . . Freitag, C. M. (2014).
Facial emotion recognition in paranoid schizophrenia and autism spectrum disorder.
Schizophr Res, 159(2-3), 509-514. doi:10.1016/j.schres.2014.08.030
Salem, J. E., Kring, A. M., & Kerr, S. L. (1996). More evidence for generalized poor performance in
facial emotion perception in schizophrenia. J Abnorm Psychol, 105(3), 480-483.
Salthouse, T. A. (2000). Aging and measures of processing speed. Biol Psychol, 54(1-3), 35-54.
Sato, W., Fujimura, T., & Suzuki, N. (2008). Enhanced facial EMG activity in response to dynamic
facial expressions. Int J Psychophysiol, 70(1), 70-74. doi:10.1016/j.ijpsycho.2008.06.001
Sato, W., Kochiyama, T., Yoshikawa, S., Naito, E., & Matsumura, M. (2004). Enhanced neural
activity in response to dynamic facial expressions of emotion: an fMRI study. Brain Res Cogn
Brain Res, 20(1), 81-91. doi:10.1016/j.cogbrainres.2004.01.008
202
Sato, W., & Yoshikawa, S. (2007). Spontaneous facial mimicry in response to dynamic facial
expressions. Cognition, 104(1), 1-18. doi:10.1016/j.cognition.2006.05.001
Savla, G. N., Vella, L., Armstrong, C. C., Penn, D. L., & Twamley, E. W. (2013). Deficits in domains
of social cognition in schizophrenia: a meta-analysis of the empirical evidence. Schizophr
Bull, 39(5), 979-992. doi:10.1093/schbul/sbs080
Schaefer, K. L., Baumann, J., Rich, B. A., Luckenbaugh, D. A., & Zarate, C. A., Jr. (2010).
Perception of facial emotion in adults with bipolar or unipolar depression and controls. J
Psychiatr Res, 44(16), 1229-1235. doi:10.1016/j.jpsychires.2010.04.024
Schmalzl, L., Palermo, R., & Coltheart, M. (2008). Cognitive heterogeneity in genetically based
prosopagnosia: a family study. J Neuropsychol, 2(Pt 1), 99-117.
Schmidt-Hansen, M., Killcross, A. S., & Honey, R. C. (2009). Latent inhibition, learned irrelevance,
and schizotypy: Assessing their relationship. Cognitive Neuropsychiatry, 14(1), 11-29.
doi:10.1080/13546800802664539
Schmiedt, C., Brand, A., Hildebrandt, H., & Basar-Eroglu, C. (2005). Event-related theta oscillations
during working memory tasks in patients with schizophrenia and healthy controls. Brain Res
Cogn Brain Res, 25(3), 936-947. doi:10.1016/j.cogbrainres.2005.09.015
Schneider, F., Gur, R. C., Gur, R. E., & Shtasel, D. L. (1995). Emotional processing in schizophrenia:
neurobehavioral probes in relation to psychopathology. Schizophr Res, 17(1), 67-75.
Schneider, F., Gur, R. C., Koch, K., Backes, V., Amunts, K., Shah, N. J., . . . Habel, U. (2006).
Impairment in the specificity of emotion processing in schizophrenia. Am J Psychiatry,
163(3), 442-447. doi:10.1176/appi.ajp.163.3.442
Schneider, F., Weiss, U., Kessler, C., Salloum, J. B., Posse, S., Grodd, W., & Muller-Gartner, H. W.
(1998). Differential amygdala activation in schizophrenia during sadness. Schizophr Res,
34(3), 133-142.
Scholten, M. R., Aleman, A., Montagne, B., & Kahn, R. S. (2005). Schizophrenia and processing of
facial emotions: sex matters. Schizophr Res, 78(1), 61-67. doi:10.1016/j.schres.2005.06.019
Schwartz, B. L., Marvel, C. L., Drapalski, A., Rosse, R. B., & Deutsch, S. I. (2002). Configural
processing in face recognition in schizophrenia. Cogn Neuropsychiatry, 7(1), 15-39.
doi:10.1080/13546800143000113
Seiferth, N. Y., Pauly, K., Habel, U., Kellermann, T., Shah, N. J., Ruhrmann, S., . . . Kircher, T.
(2008). Increased neural response related to neutral faces in individuals at risk for psychosis.
Neuroimage, 40(1), 289-297. doi:10.1016/j.neuroimage.2007.11.020
Selemon, L. D., Rajkowska, G., & Goldman-Rakic, P. S. (1995). Abnormally high neuronal density in
the schizophrenic cortex. A morphometric analysis of prefrontal area 9 and occipital area 17.
Arch Gen Psychiatry, 52(10), 805-818; discussion 819-820.
Sergent, J., Ohta, S., Macdonald, B., & Zuck, E. (1994). Segregated processing of facial identity and
emotion in the human brain: A PET study. Visual Cognition, 1(2-3), 349-369.
Sergi, M. J., Rassovsky, Y., Widmark, C., Reist, C., Erhart, S., Braff, D. L., . . . Green, M. F. (2007).
Social cognition in schizophrenia: relationships with neurocognition and negative symptoms.
Schizophr Res, 90(1-3), 316-324. doi:10.1016/j.schres.2006.09.028
Shean, G., Bell, E., & Cameron, C. D. (2007). Recognition of nonverbal affect and schizotypy. J
Psychol, 141(3), 281-291. doi:10.3200/JRLP.141.3.281-292
Shenton, M. E., Dickey, C. C., Frumin, M., & McCarley, R. W. (2001). A review of MRI findings in
schizophrenia. Schizophr Res, 49(1-2), 1-52.
Shimizu, T., Shimizu, A., Yamashita, K., Iwase, M., Kajimoto, O., & Kawasaki, T. (2000).
Comparison of eye-movement patterns in schizophrenic and normal adults during
examination of facial affect displays. Percept Mot Skills, 91(3 Pt 2), 1045-1056.
doi:10.2466/pms.2000.91.3f.1045
Shin, Y. W., Na, M. H., Ha, T. H., Kang, D. H., Yoo, S. Y., & Kwon, J. S. (2008). Dysfunction in
configural face processing in patients with schizophrenia. Schizophr Bull, 34(3), 538-543.
doi:10.1093/schbul/sbm118
Shrira, A., & Tsakanikos, E. (2009). Latent inhibition as a function of positive and negative
schizotypal symptoms: Evidence for a bi-directional model. Personality and Individual
Differences, 47(5), 434-438. doi:http://dx.doi.org/10.1016/j.paid.2009.04.015
203
Silver, H., Bilker, W., & Goodman, C. (2009). Impaired recognition of happy, sad and neutral
expressions in schizophrenia is emotion, but not valence, specific and context dependent.
Psychiatry Res, 169(2), 101-106. doi:10.1016/j.psychres.2008.11.017
Silver, H., & Shlomo, N. (2001). Perception of facial emotions in chronic schizophrenia does not
correlate with negative symptoms but correlates with cognitive and motor dysfunction.
Schizophr Res, 52(3), 265-273.
Silver, H., Shlomo, N., Turner, T., & Gur, R. C. (2002). Perception of happy and sad facial
expressions in chronic schizophrenia: evidence for two evaluative systems. Schizophr Res,
55(1-2), 171-177.
Silverstein, S. M., Keane, B. P., Barch, D. M., Carter, C. S., Gold, J. M., Kovacs, I., . . . Strauss, M. E.
(2012). Optimization and validation of a visual integration test for schizophrenia research.
Schizophr Bull, 38(1), 125-134. doi:10.1093/schbul/sbr141
Silverstein, S. M., Kovacs, I., Corry, R., & Valone, C. (2000). Perceptual organization, the
disorganization syndrome, and context processing in chronic schizophrenia. Schizophr Res,
43(1), 11-20.
Simpson, C., Pinkham, A. E., Kelsven, S., & Sasson, N. J. (2013). Emotion recognition abilities
across stimulus modalities in schizophrenia and the role of visual attention. Schizophr Res,
151(1-3), 102-106. doi:10.1016/j.schres.2013.09.026
Soria Bauser, D., Thoma, P., Aizenberg, V., Brune, M., Juckel, G., & Daum, I. (2012). Face and body
perception in schizophrenia: a configural processing deficit? Psychiatry Res, 195(1-2), 9-17.
doi:10.1016/j.psychres.2011.07.017
Souto, T., Baptista, A., Tavares, D., Queirós, C., & António, M. (2013). Facial emotional recognition
in schizophrenia: preliminary results of the virtual reality program for facial emotional
recognition. Archives of Clinical Psychiatry (São Paulo), 40(4), 129-134.
Spezio, M. L., Huang, P. Y., Castelli, F., & Adolphs, R. (2007). Amygdala damage impairs eye
contact during conversations with real people. J Neurosci, 27(15), 3994-3997.
doi:10.1523/JNEUROSCI.3789-06.2007
Stahl, S. M., & Buckley, P. F. (2007). Negative symptoms of schizophrenia: a problem that will not
go away. Acta Psychiatr Scand, 115(1), 4-11. doi:10.1111/j.1600-0447.2006.00947.x
Starzer, M. S. K., Nordentoft, M., & Hjorthoj, C. (2018). Rates and Predictors of Conversion to
Schizophrenia or Bipolar Disorder Following Substance-Induced Psychosis. Am J Psychiatry,
175(4), 343-350. doi:10.1176/appi.ajp.2017.17020223
Stekelenburg, J. J., & de Gelder, B. (2004). The neural correlates of perceiving human bodies: an ERP
study on the body-inversion effect. Neuroreport, 15(5), 777-780.
Strauss, G. P., Jetha, S. S., Ross, S. A., Duke, L. A., & Allen, D. N. (2010). Impaired facial affect
labeling and discrimination in patients with deficit syndrome schizophrenia. Schizophr Res,
118(1-3), 146-153. doi:10.1016/j.schres.2010.01.016
Streit, M., Wolwer, W., Brinkmeyer, J., Ihl, R., & Gaebel, W. (2000). Electrophysiological correlates
of emotional and structural face processing in humans. Neurosci Lett, 278(1-2), 13-16.
Streit, M., Wolwer, W., Brinkmeyer, J., Ihl, R., & Gaebel, W. (2001). EEG-correlates of facial affect
recognition and categorisation of blurred faces in schizophrenic patients and healthy
volunteers. Schizophr Res, 49(1-2), 145-155.
Streit, M., Wolwer, W., & Gaebel, W. (1997). Facial-affect recognition and visual scanning behaviour
in the course of schizophrenia. Schizophr Res, 24(3), 311-317.
Studio ArchiDOM. (2011). 3D free model cars. Retrieved from
http://www.archidom.net/ddd/mod_drive.htm.
Stuhrmann, A., Suslow, T., & Dannlowski, U. (2011). Facial emotion processing in major depression:
a systematic review of neuroimaging findings. Biol Mood Anxiety Disord, 1(1), 10.
doi:10.1186/2045-5380-1-10
Summers, M., Papadopoulou, K., Bruno, S., Cipolotti, L., & Ron, M. A. (2006). Bipolar I and bipolar
II disorder: cognition and emotion processing. Psychol Med, 36(12), 1799-1809.
doi:10.1017/S0033291706008804
Surguladze, S., Russell, T., Kucharska-Pietura, K., Travis, M. J., Giampietro, V., David, A. S., &
Phillips, M. L. (2006). A reversal of the normal pattern of parahippocampal response to
204
neutral and fearful faces is associated with reality distortion in schizophrenia. Biol Psychiatry,
60(5), 423-431. doi:10.1016/j.biopsych.2005.11.021
Surguladze, S. A., Marshall, N., Schulze, K., Hall, M. H., Walshe, M., Bramon, E., . . . McDonald, C.
(2010). Exaggerated neural response to emotional faces in patients with bipolar disorder and
their first-degree relatives. Neuroimage, 53(1), 58-64. doi:10.1016/j.neuroimage.2010.05.069
Surguladze, S. A., Young, A. W., Senior, C., Brebion, G., Travis, M. J., & Phillips, M. L. (2004).
Recognition accuracy and response bias to happy and sad facial expressions in patients with
major depression. Neuropsychology, 18(2), 212-218. doi:10.1037/0894-4105.18.2.212
Suslow, T., Dannlowski, U., Lalee-Mentzel, J., Donges, U. S., Arolt, V., & Kersting, A. (2004).
Spatial processing of facial emotion in patients with unipolar depression: a longitudinal study.
J Affect Disord, 83(1), 59-63. doi:10.1016/j.jad.2004.03.003
Swartz, M. S., Wagner, H. R., Swanson, J. W., Stroup, T. S., McEvoy, J. P., Canive, J. M., . . .
Lieberman, J. A. (2006). Substance use in persons with schizophrenia: baseline prevalence
and correlates from the NIMH CATIE study. J Nerv Ment Dis, 194(3), 164-172.
doi:10.1097/01.nmd.0000202575.79453.6e
Tamminga, C. A., Pearlson, G., Keshavan, M., Sweeney, J., Clementz, B., & Thaker, G. (2014).
Bipolar and schizophrenia network for intermediate phenotypes: outcomes across the
psychosis continuum. Schizophr Bull, 40 Suppl 2, S131-137. doi:10.1093/schbul/sbt179
Tan, H. K., Jones, G. V., & Watson, D. G. (2009). Encouraging the perceptual underdog: positive
affective priming of nonpreferred local-global processes. Emotion, 9(2), 238-247.
doi:10.1037/a0014713
Tanaka, J. W., Curran, T., Porterfield, A. L., & Collins, D. (2006). Activation of preexisting and
acquired face representations: the N250 event-related potential as an index of face familiarity.
J Cogn Neurosci, 18(9), 1488-1497. doi:10.1162/jocn.2006.18.9.1488
Tandon, R., Gaebel, W., Barch, D. M., Bustillo, J., Gur, R. E., Heckers, S., . . . Carpenter, W. (2013).
Definition and description of schizophrenia in the DSM-5. Schizophr Res, 150(1), 3-10.
doi:10.1016/j.schres.2013.05.028
Tandon, R., Nasrallah, H. A., & Keshavan, M. S. (2009). Schizophrenia, "just the facts" 4. Clinical
features and conceptualization. Schizophr Res, 110(1-3), 1-23.
doi:10.1016/j.schres.2009.03.005
Taylor, S. F., Kang, J., Brege, I. S., Tso, I. F., Hosanagar, A., & Johnson, T. D. (2012). Meta-analysis
of functional neuroimaging studies of emotion perception and experience in schizophrenia.
Biol Psychiatry, 71(2), 136-145. doi:10.1016/j.biopsych.2011.09.007
The International Schizophrenia Consortium. (2009). Common polygenic variation contributes to risk
of schizophrenia and bipolar disorder. Nature, 460, 748. doi:10.1038/nature08185
Thompson, A., Sullivan, S., Heron, J., Thomas, K., Zammit, S., Horwood, J., . . . Harrison, G. (2011).
Childhood facial emotion recognition and psychosis-like symptoms in a nonclinical
population at 12 years of age: results from the ALSPAC birth cohort. Cogn Neuropsychiatry,
16(2), 136-157. doi:10.1080/13546805.2010.510040
Thompson, J. L., Urban, N., Slifstein, M., Xu, X., Kegeles, L. S., Girgis, R. R., . . . Abi-Dargham, A.
(2013). Striatal dopamine release in schizophrenia comorbid with substance dependence. Mol
Psychiatry, 18(8), 909-915. doi:10.1038/mp.2012.109
Toh, W. L., Rossell, S. L., & Castle, D. J. (2011). Current visual scanpath research: a review of
investigations into the psychotic, anxiety, and mood disorders. Compr Psychiatry, 52(6), 567-
579. doi:10.1016/j.comppsych.2010.12.005
Tomkins, S. S., & McCarter, R. (1964). What and Where are the Primary Affects? Some Evidence for
a Theory. Perceptual and Motor Skills, 18(1), 119-158. doi:doi:10.2466/pms.1964.18.1.119
Toomey, R., Seidman, L. J., Lyons, M. J., Faraone, S. V., & Tsuang, M. T. (1999). Poor perception of
nonverbal social-emotional cues in relatives of schizophrenic patients. Schizophr Res, 40(2),
121-130.
Tottenham, N., Tanaka, J. W., Leon, A. C., McCarry, T., Nurse, M., Hare, T. A., . . . Nelson, C.
(2009). The NimStim set of facial expressions: judgments from untrained research
participants. Psychiatry Res, 168(3), 242-249. doi:10.1016/j.psychres.2008.05.006
Tranel, D., Damasio, A. R., & Damasio, H. (1988). Intact recognition of facial expression, gender,
and age in patients with impaired recognition of face identity. Neurology, 38(5), 690-696.
205
Trautmann, S. A., Fehr, T., & Herrmann, M. (2009). Emotions in motion: dynamic compared to static
facial expressions of disgust and happiness reveal more widespread emotion-specific
activations. Brain Res, 1284, 100-115. doi:10.1016/j.brainres.2009.05.075
Tsao, D. Y., Freiwald, W. A., Knutsen, T. A., Mandeville, J. B., & Tootell, R. B. (2003). Faces and
objects in macaque cerebral cortex. Nat Neurosci, 6(9), 989-995. doi:10.1038/nn1111
Tsao, D. Y., Moeller, S., & Freiwald, W. A. (2008). Comparing face patch systems in macaques and
humans. Proc Natl Acad Sci U S A, 105(49), 19514-19519. doi:10.1073/pnas.0809662105
Tsao, D. Y., Schweers, N., Moeller, S., & Freiwald, W. A. (2008). Patches of face-selective cortex in
the macaque frontal lobe. Nat Neurosci, 11(8), 877-879. doi:10.1038/nn.2158
Tsoneva, T., Baldo, D., Lema, V., & Garcia-Molina, G. (2011). EEG-rhythm dynamics during a 2-
back working memory task and performance. Conf Proc IEEE Eng Med Biol Soc, 2011,
3828-3831. doi:10.1109/IEMBS.2011.6090952
Tsunoda, T., Kanba, S., Ueno, T., Hirano, Y., Hirano, S., Maekawa, T., & Onitsuka, T. (2012).
Altered face inversion effect and association between face N170 reduction and social
dysfunction in patients with schizophrenia. Clin Neurophysiol, 123(9), 1762-1768.
doi:10.1016/j.clinph.2012.01.024
Turetsky, B. I., Kohler, C. G., Indersmitten, T., Bhati, M. T., Charbonnier, D., & Gur, R. C. (2007).
Facial emotion recognition in schizophrenia: when and why does it go awry? Schizophr Res,
94(1-3), 253-263. doi:10.1016/j.schres.2007.05.001
Uhlhaas, P. J., Phillips, W. A., Mitchell, G., & Silverstein, S. M. (2006). Perceptual grouping in
disorganized schizophrenia. Psychiatry Res, 145(2-3), 105-117.
doi:10.1016/j.psychres.2005.10.016
Upthegrove, R., Marwaha, S., & Birchwood, M. (2017). Depression and Schizophrenia: Cause,
Consequence, or Trans-diagnostic Issue? Schizophr Bull, 43(2), 240-244.
doi:10.1093/schbul/sbw097
Utama, N. P., Takemoto, A., Koike, Y., & Nakamura, K. (2009). Phased processing of facial emotion:
an ERP study. Neurosci Res, 64(1), 30-40. doi:10.1016/j.neures.2009.01.009
Valstar, M., & Pantic, M. (2010). Induced disgust, happiness and surprise: an addition to the mmi
facial expression database. Paper presented at the Proc. 3rd Intern. Workshop on EMOTION
(satellite of LREC): Corpora for Research on Emotion and Affect.
van 't Wout, M., Aleman, A., Kessels, R. P., Cahn, W., de Haan, E. H., & Kahn, R. S. (2007).
Exploring the nature of facial affect processing deficits in schizophrenia. Psychiatry Res,
150(3), 227-235. doi:10.1016/j.psychres.2006.03.010
van 't Wout, M., Aleman, A., Kessels, R. P., Laroi, F., & Kahn, R. S. (2004). Emotional processing in
a non-clinical psychosis-prone sample. Schizophr Res, 68(2-3), 271-281.
doi:10.1016/j.schres.2003.09.006
Van den Stock, J., Tamietto, M., Hervais-Adelman, A., Pegna, A. J., & de Gelder, B. (2015). Body
recognition in a patient with bilateral primary visual cortex lesions. Biol Psychiatry, 77(7),
e31-33. doi:10.1016/j.biopsych.2013.06.023
van der Schalk, J., Hawk, S. T., Fischer, A. H., & Doosje, B. (2011). Moving faces, looking places:
validation of the Amsterdam Dynamic Facial Expression Set (ADFES). Emotion, 11(4), 907-
920. doi:10.1037/a0023853
van Os, J., & Kapur, S. (2009). Schizophrenia. Lancet, 374(9690), 635-645. doi:10.1016/S0140-
6736(09)60995-8
van Os, J., Linscott, R. J., Myin-Germeys, I., Delespaul, P., & Krabbendam, L. (2009). A systematic
review and meta-analysis of the psychosis continuum: evidence for a psychosis proneness-
persistence-impairment model of psychotic disorder. Psychol Med, 39(2), 179-195.
doi:10.1017/S0033291708003814
van Os, J., & Reininghaus, U. (2016). Psychosis as a transdiagnostic and extended phenotype in the
general population. World Psychiatry, 15(2), 118-124. doi:10.1002/wps.20310
Van Rheenen, T. E., & Rossell, S. L. (2013a). Auditory-prosodic processing in bipolar disorder; from
sensory perception to emotion. J Affect Disord, 151(3), 1102-1107.
doi:10.1016/j.jad.2013.08.039
206
Van Rheenen, T. E., & Rossell, S. L. (2013b). Is the non-verbal behavioural emotion-processing
profile of bipolar disorder impaired? A critical review. Acta Psychiatr Scand, 128(3), 163-
178. doi:10.1111/acps.12125
Van Rheenen, T. E., & Rossell, S. L. (2014). Let's face it: facial emotion processing is impaired in
bipolar disorder. J Int Neuropsychol Soc, 20(2), 200-208. doi:10.1017/S1355617713001367
van Rijn, S., Schothorst, P., Wout, M., Sprong, M., Ziermans, T., van Engeland, H., . . . Swaab, H.
(2011). Affective dysfunctions in adolescents at risk for psychosis: emotion awareness and
social functioning. Psychiatry Res, 187(1-2), 100-105. doi:10.1016/j.psychres.2010.10.007
van Veluw, S. J., & Chance, S. A. (2014). Differentiating between self and others: an ALE meta-
analysis of fMRI studies of self-recognition and theory of mind. Brain Imaging Behav, 8(1),
24-38. doi:10.1007/s11682-013-9266-8
Vaskinn, A., Sundet, K., Friis, S., Simonsen, C., Birkenaes, A. B., Engh, J. A., . . . Andreassen, O. A.
(2007). The effect of gender on emotion perception in schizophrenia and bipolar disorder.
Acta Psychiatr Scand, 116(4), 263-270. doi:10.1111/j.1600-0447.2007.00991.x
Vederman, A. C., Weisenbach, S. L., Rapport, L. J., Leon, H. M., Haase, B. D., Franti, L. M., . . .
McInnis, M. G. (2012). Modality-specific alterations in the perception of emotional stimuli in
Bipolar Disorder compared to Healthy Controls and Major Depressive Disorder. Cortex,
48(8), 1027-1034. doi:10.1016/j.cortex.2011.03.017
Venables, P. H., & Rector, N. A. (2000). The content and structure of schizotypy: a study using
confirmatory factor analysis. Schizophr Bull, 26(3), 587-602.
Venn, H. R., Gray, J. M., Montagne, B., Murray, L. K., Michael Burt, D., Frigerio, E., . . . Young, A.
H. (2004). Perception of facial expressions of emotion in bipolar disorder. Bipolar Disord,
6(4), 286-293. doi:10.1111/j.1399-5618.2004.00121.x
Ventura, J., Thames, A. D., Wood, R. C., Guzik, L. H., & Hellemann, G. S. (2010). Disorganization
and reality distortion in schizophrenia: a meta-analysis of the relationship between positive
symptoms and neurocognitive deficits. Schizophr Res, 121(1-3), 1-14.
doi:10.1016/j.schres.2010.05.033
Ventura, J., Wood, R. C., Jimenez, A. M., & Hellemann, G. S. (2013). Neurocognition and symptoms
identify links between facial recognition and emotion processing in schizophrenia: meta-
analytic findings. Schizophr Res, 151(1-3), 78-84. doi:10.1016/j.schres.2013.10.015
Versace, A., Thompson, W. K., Zhou, D., Almeida, J. R., Hassel, S., Klein, C. R., . . . Phillips, M. L.
(2010). Abnormal left and right amygdala-orbitofrontal cortical functional connectivity to
emotional faces: state versus trait vulnerability markers of depression in bipolar disorder. Biol
Psychiatry, 67(5), 422-431. doi:10.1016/j.biopsych.2009.11.025
Victoria Institute of Strategic Economic Studies. (2016). The economic cost of serious mental illness
and comorbidities in Australia and New Zealand. Melbourne: Royal Australian and New
Zealand College of Psychiatrists ; Australian Health Policy Collaboration Retrieved from
http://vuir.vu.edu.au/31128/.
Vizueta, N., Rudie, J. D., Townsend, J. D., Torrisi, S., Moody, T. D., Bookheimer, S. Y., & Altshuler,
L. L. (2012). Regional fMRI hypoactivation and altered functional connectivity during
emotion processing in nonmedicated depressed patients with bipolar II disorder. Am J
Psychiatry, 169(8), 831-840. doi:10.1176/appi.ajp.2012.11030349
Vollema, M. G., & Hoijtink, H. (2000). The multidimensionality of self-report schizotypy in a
psychiatric population: an analysis using multidimensional Rasch models. Schizophr Bull,
26(3), 565-575.
Vuilleumier, P., & Pourtois, G. (2007). Distributed and interactive brain mechanisms during emotion
face perception: evidence from functional neuroimaging. Neuropsychologia, 45(1), 174-194.
doi:10.1016/j.neuropsychologia.2006.06.003
Waldeck, T. L., & Miller, L. S. (2000). Social skills deficits in schizotypal personality disorder.
Psychiatry Res, 93(3), 237-246.
Walker, D. M., & Standen, P. J. (2011). Level and pattern of neuropsychological functioning in early-
onset psychoses. Eur Child Adolesc Psychiatry, 20(11-12), 551-560. doi:10.1007/s00787-
011-0219-8
Wallhoff, F. (2006). Facial Expressions and Emotion Database. http://cotesys.mmk.e-technik.tu-
muenchen.de/isg/content/feed-database
207
Walther, S., Federspiel, A., Horn, H., Bianchi, P., Wiest, R., Wirth, M., . . . Muller, T. J. (2009).
Encoding deficit during face processing within the right fusiform face area in schizophrenia.
Psychiatry Res, 172(3), 184-191. doi:10.1016/j.pscychresns.2008.07.009
Ward, J. H. J. (1963). Hierarchical Grouping to Optimize an Objective Function (Vol. 58).
Wechsler, D. (1997). Wechsler Memory Scale - Third Edition (WMS-III) administration and scoring
manual. San Antonio, TX: The Psychological Corporation.
Wechsler, D. (2008). Wechsler adult intelligence scale–Fourth Edition (WAIS–IV). San Antonio, TX:
NCS Pearson, 22, 498.
Wehrle, T., Kaiser, S., Schmidt, S., & Scherer, K. R. (2000). Studying the dynamics of emotional
expression using synthesized facial muscle movements. J Pers Soc Psychol, 78(1), 105-119.
Weniger, G., Lange, C., Ruther, E., & Irle, E. (2004). Differential impairments of facial affect
recognition in schizophrenia subtypes and major depression. Psychiatry Res, 128(2), 135-146.
doi:10.1016/j.psychres.2003.12.027
Weyers, P., Muhlberger, A., Hefele, C., & Pauli, P. (2006). Electromyographic responses to static and
dynamic avatar emotional facial expressions. Psychophysiology, 43(5), 450-453.
doi:10.1111/j.1469-8986.2006.00451.x
Whittaker, J. F., Deakin, J. F., & Tomenson, B. (2001). Face processing in schizophrenia: defining the
deficit. Psychol Med, 31(3), 499-507.
Wigman, J. T., de Vos, S., Wichers, M., van Os, J., & Bartels-Velthuis, A. A. (2017). A
Transdiagnostic Network Approach to Psychosis. Schizophr Bull, 43(1), 122-132.
doi:10.1093/schbul/sbw095
Wigman, J. T., van Nierop, M., Vollebergh, W. A., Lieb, R., Beesdo-Baum, K., Wittchen, H. U., &
van Os, J. (2012). Evidence that psychotic symptoms are prevalent in disorders of anxiety and
depression, impacting on illness onset, risk, and severity--implications for diagnosis and ultra-
high risk research. Schizophr Bull, 38(2), 247-257. doi:10.1093/schbul/sbr196
Williams, B. T., Henry, J. D., & Green, M. J. (2007). Facial affect recognition and schizotypy. Early
Intervention in Psychiatry, 1(2), 177-182. doi:10.1111/j.1751-7893.2007.00023.x
Williams, L. M., Loughland, C. M., Gordon, E., & Davidson, D. (1999). Visual scanpaths in
schizophrenia: is there a deficit in face recognition? Schizophr Res, 40(3), 189-199.
Winston, J. S., Henson, R. N., Fine-Goulden, M. R., & Dolan, R. J. (2004). fMRI-adaptation reveals
dissociable neural representations of identity and expression in face perception. J
Neurophysiol, 92(3), 1830-1839. doi:10.1152/jn.00155.2004
Wolwer, W., Streit, M., Polzer, U., & Gaebel, W. (1996). Facial affect recognition in the course of
schizophrenia. Eur Arch Psychiatry Clin Neurosci, 246(3), 165-170.
Woodberry, K. A., Giuliano, A. J., & Seidman, L. J. (2008). Premorbid IQ in schizophrenia: a meta-
analytic review. Am J Psychiatry, 165(5), 579-587. doi:10.1176/appi.ajp.2008.07081242
Wright, I. C., Rabe-Hesketh, S., Woodruff, P. W., David, A. S., Murray, R. M., & Bullmore, E. T.
(2000). Meta-analysis of regional brain volumes in schizophrenia. Am J Psychiatry, 157(1),
16-25. doi:10.1176/ajp.157.1.16
Wynn, J. K., Jahshan, C., Altshuler, L. L., Glahn, D. C., & Green, M. F. (2013). Event-related
potential examination of facial affect processing in bipolar disorder and schizophrenia.
Psychol Med, 43(1), 109-117. doi:10.1017/S0033291712001006
Wynn, J. K., Lee, J., Horan, W. P., & Green, M. F. (2008). Using event related potentials to explore
stages of facial affect recognition deficits in schizophrenia. Schizophr Bull, 34(4), 679-687.
doi:10.1093/schbul/sbn047
Xu, Y., Liu, J., & Kanwisher, N. (2005). The M170 is selective for faces, not for expertise.
Neuropsychologia, 43(4), 588-597. doi:10.1016/j.neuropsychologia.2004.07.016
Yang, E., Tadin, D., Glasser, D. M., Hong, S. W., Blake, R., & Park, S. (2013). Visual context
processing in schizophrenia. Clin Psychol Sci, 1(1), 5-15. doi:10.1177/2167702612464618
Yin, R. K. (1969). Looking at upside-down faces. Journal of Experimental Psychology, 81, 141-145.
doi:10.1037/h0027474
Yoon, J. H., D'Esposito, M., & Carter, C. S. (2006). Preserved function of the fusiform face area in
schizophrenia as revealed by fMRI. Psychiatry Res, 148(2-3), 205-216.
doi:10.1016/j.pscychresns.2006.06.002
208
Yoon, J. H., Maddock, R. J., Rokem, A., Silver, M. A., Minzenberg, M. J., Ragland, J. D., & Carter,
C. S. (2010). GABA concentration is reduced in visual cortex in schizophrenia and correlates
with orientation-specific surround suppression. J Neurosci, 30(10), 3777-3781.
doi:10.1523/JNEUROSCI.6158-09.2010
Yoshikawa, S., & Sato, W. (2006). Enhanced perceptual, emotional, and motor processing in response
to dynamic facial expressions of emotion. Japanese Psychological Research, 48(3), 213-222.
Young, A. W., Newcombe, F., de Haan, E. H., Small, M., & Hay, D. C. (1993). Face perception after
brain injury. Selective impairments affecting identity and expression. Brain, 116 ( Pt 4), 941-
959.
Yovel, G., Levy, J., & Yovel, I. (2001). Hemispheric asymmetries for global and local visual
perception: effects of stimulus and task factors. J Exp Psychol Hum Percept Perform, 27(6),
1369-1385.
Yun, J. Y., Hur, J. W., Jung, W. H., Jang, J. H., Youn, T., Kang, D. H., . . . Kwon, J. S. (2014).
Dysfunctional role of parietal lobe during self-face recognition in schizophrenia. Schizophr
Res, 152(1), 81-88. doi:10.1016/j.schres.2013.07.010
Zambelli, A. E. (2016). A data-driven approach to estimating the number of clusters in hierarchical
clustering. F1000Research, 5, ISCB Comm J-2809. doi:10.12688/f1000research.10103.1
Zanarini, M. C., Barison, L. K., Frankenburg, F. R., Reich, D. B., & Hudson, J. I. (2009). Family
history study of the familial coaggregation of borderline personality disorder with axis I and
nonborderline dramatic cluster axis II disorders. J Pers Disord, 23(4), 357-369.
doi:10.1521/pedi.2009.23.4.357
Zvyagintsev, M., Parisi, C., Chechko, N., Nikolaev, A. R., & Mathiak, K. (2013). Attention and
multisensory integration of emotions in schizophrenia. Front Hum Neurosci, 7, 674.
doi:10.3389/fnhum.2013.00674
Appendices
209
Appendix A: Participant information and consent form (Healthy control)
210
211
212
213
Appendix B: Questionnaire for healthy controls
214
Appendix C: NART word card and scoring form
215
216
Appendix D: The Oxford-Liverpool Inventory of Feelings and Experiences
217
218
219
220
221
222
Appendix E: Participant information and consent form (Inpatient)
223
224
225
226
227
228
Appendix F: Patient demographic questionnaire
229
230
231
232
233
Appendix G: Edinburgh handedness inventory
234
Appendix H: The SCI-PANSS interview and rating form
235
236
237
238
239
240
241
242
243
244
245
246
247
Minerva Access is the Institutional Repository of The University of Melbourne
Author/s:
Darke, Hayley
Title:
Face processing impairments in schizophrenia and other psychiatric disorders
Date:
2018
Persistent Link:
http://hdl.handle.net/11343/217206
File Description:
Face Processing Impairments in Schizophrenia and Other Psychiatric Disorders - Amended
thesis
Terms and Conditions:
Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the
copyright owner. The work may not be altered without permission from the copyright owner.
Readers may only download, print and save electronic copies of whole works for their own
personal non-commercial use. Any use that exceeds these limits requires permission from
the copyright owner. Attribution is essential when quoting or paraphrasing from these works.