Impulsivity, aggression and brain structure in high and low lethalitysuicide attempters with borderline personality disorder

Paul Soloff a,n, Richard White b, Vaibhav A. Diwadkar b

a Department of Psychiatry, Western Psychiatric Institute and Clinic, University of Pittsburgh, 3811 O'Hara Street, Pittsburgh, PA 15213, USAb Department of Psychiatry and Behavioral Neurosciences, Wayne State University School of Medicine, Detroit, MI, USA

a r t i c l e i n f o

Article history:Received 10 April 2013Received in revised form5 November 2013Accepted 7 February 2014

Keywords:Magnetic resonance imagingImpulsive–aggressionSuicidal behaviorBorderline personality disorder

a b s t r a c t

Impulsivity and aggressiveness are trait dispositions associated with the vulnerability to suicidalbehavior across diagnoses. They are associated with structural and functional abnormalities in brainnetworks involved in regulation of mood, impulse and behavior. They are also core characteristics ofborderline personality disorder (BPD), a disorder defined, in part, by recurrent suicidal behavior. Weassessed the relationships between personality traits, brain structure and lethality of suicide attempts in51 BPD attempters using multiple regression analyses on structural MRI data. BPD was diagnosed by theDiagnostic Interview for Borderline Patients-revised, impulsivity by the Barratt Impulsiveness Scale (BIS),aggression by the Brown–Goodwin Lifetime History of Aggression (LHA), and high lethality by a score of4 or more on the Lethality Rating Scale (LRS). Sixteen High Lethality attempters were compared to 35Low Lethality attempters, with no significant differences noted in gender, co-morbidity, childhood abuse,BIS or LHA scores. Degree of medical lethality (LRS) was negatively related to gray matter volumes acrossmultiple fronto-temporal–limbic regions. Effects of impulsivity and aggression on gray matter volumesdiscriminated High from Low Lethality attempters and differed markedly within lethality groups.Lethality of suicide attempts in BPD may be related to the mediation of these personality traits byspecific neural networks.

& 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Personality traits such as impulsivity and aggressiveness areassociated with suicidal behavior across diagnoses. In a stress–diathesis model of suicide, they represent vulnerable tempera-ments, and predispositions to impulsive and aggressive behaviorin response to specific trigger events (for review, see Mannet al.,1999; Mann, 2003). Trait dispositions such as impulsivityand aggressiveness may be heritable (e.g., as endophenotypes), oracquired in the course of development (e.g., through childhoodabuse). In neuroimaging studies, they have been associated withvariations in the structure and function of brain networks thatregulate mood, impulse and behavior. At times of emotional stress,dysfunction in these neural networks may result in interferencewith executive cognitive functions, such as response inhibition,conflict resolution, and recall of episodic memory (for review, seeFertuck et al., 2006). As a result, problem solving and adaptivecoping are impaired, increasing the likelihood of impulsive oraggressive behavior. We study the relationship between personality

characteristics, brain function and suicidal behavior in the contextof borderline personality disorder (BPD), a personality disorderdefined, in part, by recurrent suicidal behavior, impulsivity andaggression. With a suicide rate of 3–10% and a community pre-valence estimated at 1% of the population, BPD is a clinicallyrelevant model for the study of suicide (Swartz et al., 1990).

There is a paucity of neuroimaging studies in BPD subjectsascertained specifically for suicidal behavior. In a voxel-basedmorphometry study (VBM) comparing BPD suicide attempterswith BPD non-attempters, we recently reported specific structuraldifferences in BPD subjects associated with suicidal behavior, anddifferences between High Lethality and Low Lethality suicideattempters (Soloff et al., 2012). BPD attempters had diminishedgray matter concentrations in left insular cortex compared withBPD non-attempters. High Lethality attempters had diminishedgray matter compared with Low Lethality attempters in anextensive fronto-limbic network including the following regions:right middle-inferior orbital frontal cortex, right middle-superiortemporal cortex, right insular cortex, left fusiform gyrus, leftlingual gyrus, and right parahippocampal gyrus. These areas arebroadly involved in emotion regulation, behavioral control, andadaptive responding to social situations. Suicide researchers havelong maintained that suicide attempters and completers represent

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/psychresns

Psychiatry Research: Neuroimaging

http://dx.doi.org/10.1016/j.pscychresns.2014.02.0060925-4927 & 2014 Elsevier Ireland Ltd. All rights reserved.

n Corresponding author. Tel.: þ1 412 687 2646; fax þ1 412 621 2308.E-mail address: soloffph@upmc.edu (P. Soloff).

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

separate but overlapping populations, with differing clinical char-acteristics (Maris et al., 2000). High Lethality attempters sharemany clinical characteristics with patients who complete suicideand may share neurobiological vulnerabilities related to high-riskpersonality traits such as impulsivity and aggressiveness.

To assess the relationships between personality traits, brainstructure, and suicidal behavior, we used a multiple regressionanalysis of VBM data in High and Low Lethality BPD attempters tomap the relationships between impulsivity, aggression and graymatter in key brain structures.

2. Methods

2.1. Subjects

Subjects for this study were recruited by advertisement from the outpatientprograms of the Western Psychiatric Institute and Clinic and surrounding commu-nity to participate in a longitudinal study of suicidal behavior in BPD. The study wasapproved by the Institutional Review Board of the University of Pittsburgh, andfunded by the National Institute of Mental Health. All subjects gave writteninformed consent for participation.

Diagnoses were determined by Master's prepared research raters usingstructured interviews. Axis I disorders were diagnosed using the Structured ClinicalInterview for DSM III-R or DSM IV (SCID) (Spitzer et al., 1988; First et al., 2005).(Because this is a longitudinal study, DSM-IV was added when it first becameavailable.) Axis II diagnoses were established using the International PersonalityDisorders Examination (IPDE), which has a lifetime framework (Loranger et al.,1997). The Diagnostic Interview for Borderlines (DIB) (Gunderson et al., 1981) wasadministered as an independent measure of diagnosis and recent symptomseverity, with a timeframe of 3 months to 2 years for individual subscales. (TheDIB was used to preserve continuity with the longitudinal study; however, theDiagnostic Interview for Borderlines-Revised (DIB-R) was added and scoredconcurrently when it became available (Zanarini et al., 1989).) For inclusion,participants had to meet diagnostic criteria for BPD on the IPDE (probable ordefinite), have a score of 7 or more (definite) on the DIB, and 8 or more (definite) onthe DIB-R. Exclusion criteria included any past or current Axis I diagnosis ofschizophrenia, delusional (paranoid) disorder, schizoaffective disorder, bipolardisorder, or psychotic depression. Subjects were also excluded for physicaldisorders of known psychiatric consequence (e.g., hypothyroidism, seizure disorder,or brain injury) and borderline mental retardation. Medical records were reviewedwhere available to confirm inclusion and exclusion criteria. Final diagnoses weredetermined by consensus of raters using all available data. Control subjects werefree of all Axes I and II disorders. Attempter status and medical lethality of attemptswere obtained by interview using the Columbia Suicide History Form and LethalityRating Scale (Oquendo et al., 2003). Scans were obtained from newly recruitedsubjects and from subjects already enrolled in the longitudinal study at the time oftheir annual follow-up assessment. As a result, all subjects had updated SCIDinterviews for current diagnoses. Aggression was assessed by interview on theBrown–Goodwin Lifetime History of Aggression (LHA) (Brown et al., 1979) and traitimpulsivity by self-report questionnaire using the Barratt Impulsiveness Scale (BIS)(Barratt, 1965). The 24 item Hamilton Rating Scale for Depression (HamD) wasobtained before the scan as a measure of depressed mood (Guy, 1976). HighLethality status among attempters was defined as a lifetime maximum LethalityRating Scale score (LRS) of 4 or more. (For example, for a suicide attempt byoverdose with sedative drugs, an LRS score of 4 is defined as “comatose; injurysufficient for hospitalization.”)

All subjects were physically healthy, free of drugs of abuse and alcohol for atleast 1 week before the scan. Female subjects were required to have a negativescreen for pregnancy. All subjects had a negative urine toxicology screen for drugsof abuse immediately before the scan. Some BPD subjects were taking psychoactivemedication.

2.2. Imaging method

Magnetic resonance imaging (MRI) was performed with a 1.5T GE SignaImaging System running version Signa 5.4.3 software (General Electric MedicalSystems, Milwaukee, WI, USA). A T1-weighted sagittal scout image was obtained forgraphic prescription of the coronal and axial images. Three-dimensional gradientecho imaging (Spoiled Gradient Recalled Acquisition, SPGR) was performed in thecoronal plane (repetition time¼25 ms, echo time¼5 ms, nutation angle¼401, fieldof view¼24 cm, slice thickness¼1.5 mm, number of excitations¼1, and matrixsize¼256�192) to obtain 124 images covering the entire brain. Additionally, adouble echo–spin echo sequence was used to obtain T2 and proton density imagesin the axial plane to screen for neuroradiological abnormalities.

Structural MR images were preprocessed using the Statistical ParametricMapping (SPM) diffeomorphic image registration algorithm (DARTEL) in SPM8(Friston et al., 1995; Ashburner, 2007; Diwadkar et al., 2011). DARTEL optimizes thefidelity of shape-based deformations applied to fit native images in stereotacticspace, and performs favorably relative to other non-linear deformation algorithms(Klein et al., 2009). It is therefore optimized for assessing structural changes withina stereotactic framework, and well suited for VBM analyses. Following resampling(2 mm3) and segmentation of T1-weighted images, a rigid gray matter templatewas created representing the average shape and size of the brains of all the subjectsincluded in the study. Subjects' gray matter maps were then warped to the co-ordinate system of the template, with the Jacobian modulation used to scale nativegray matter volume from native to Montreal Neurological Institute (MNI) space(Good et al., 2001). This procedure has been extensively used in voxel-basedanalyses of gray matter images within the framework of random field methods.Structural data for 44 BPD attempters were previously included in a larger studycomparing BPD attempters, non-attempters and healthy control subjects (Soloffet al., 2012). Our findings from that structural analysis, defining deficits associatedwith lethality, were used to define regions of interest in this analysis. First, wedefined a regional mask corresponding to clusters of significance (Po0.05, clusterlevel) identifying reductions in gray matter volume in High (relative to Low)Lethality attempters (Ward, 2000). This regional mask spanned nine regions ofinterest including the following: the middle-inferior orbital frontal cortex, anteriorcingulate cortex, middle-superior temporal cortex, insula, hippocampus, parahip-pocampus, fusiform gyrus, lingual gyrus and amygdala (Soloff et al., 2008;Leichsenring et al., 2011).

Personality trait variables (LHA and BIS) were entered individually in multipleregression analyses as co-variates of interest to investigate the positive andnegative effects of these clinical measures on brain structures in High and LowLethality suicidal subjects (Friston et al., 1995). These methods follow previousinvestigations of effects of symptom or personality dimensions on regional brainstructure (Banissy et al., 2012). Cluster level correction (Pco0.05) was used tooptimize sensitivity to detect clusters with minimal extent (cluster formingthreshold, Pthro0.05) (Ward, 2000).

3. Results

3.1. Sample characteristics

There were 51 BPD attempters subdivided as follows: 16 HighLethality (5 male, 11 female) and 35 Low Lethality attempters (5male, 30 female), with no significant group differences by gender,race or socioeconomic status (Table 1). The mean (S.D.) age of thesample was 30.1 (8.1) years with a range of 18–47 years. HighLethality attempters were significantly older (36.1 (9.2) years) thanLow Lethality attempters (27.4 (5.9) years, t¼3.47, d.f.¼20.95,

Table 1Characteristics of the samplea.

HighLethality

Low Lethality Statistics (t, d.f., Por χ2, d.f., P)nn

N (M/F) 16 (5 M/11 F) 35 (5 M/30 F) χ2¼2.00, d.f.¼1, P¼n.s.Age (years, S.D.) 36.1 (9.2) 27.4 (5.9) t¼3.47, d.f.¼20.95,

P¼0.002Race (%Cau.) 75 74.3 χ2¼0.003, d.f.¼1, P¼n.s.Educ.(%412 years)

56.3 62.9 χ2¼0.20, d.f.¼1, P¼n.s.

SESLow¼2þ3 10 23High¼4þ5 6 12 χ2¼0.05, d.f.¼1, P¼n.s

Abused (%yes) 43.8 37.1 χ2¼0.20 d.f.¼1, P¼n.s.MDD (%yes) 75 60 χ2¼1.08, d.f.¼1, P¼n.s.Alcohol (%yes) 6.3 17.1 χ2¼1.10, d.f.¼1, P¼n.sOther drugs (%yes) 31.3 14.3 χ2¼2.01, d.f.¼1, P¼n.s.Anxiety dx. (%yes) 50 65.7 χ2¼1.14, d.f.¼1, P¼n.s.PTSD (%yes) 18.8 8.6 χ2¼1.10, d.f.¼1, P¼n.s.Psych. meds (%yes) 50.0 37.1 χ2¼0.75, d.f.¼1, P¼n.s.HamD 18.1 (10.0) 16.3 (7.9) t¼0.68, d.f.¼49, P¼n.sBIS 73.5 (5.2) 74.7 (4.8) t¼0.82, d.f.¼49, P¼n.s.LHA 15.6 (8.4) 16.5 (6.3) t¼0.39, d.f.¼49, P¼n.s

nn t, d.f., P: Student's t-test, 2 tailed. χ2, d.f., P: Chi Square test, 2 tailed.a SCID Axis I co-morbidity at time of scan. Alcohol and Other drugs include

abuse and/or dependence, anxiety dx includes any anxiety disorder except PTSD.

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎2

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

P¼0.002). The two groups did not differ significantly in proportionof subjects with current or lifetime co-morbid Axis I disorders,including the following: major depressive disorder (MDD), alcoholabuse or dependence, other drug abuse or dependence, anyanxiety disorder, or post-traumatic stress disorder. The mostfrequent current co-morbid diagnosis, MDD, was found in 12 HighLethality (75%) and 21 Low Lethality attempters (60%), with nosignificant difference between groups in severity of depressedmood (HamD) at the time of the scan (Table 1). Similarly, therewere no significant differences between groups for BIS and LHAscores, and no significant correlation between these scores acrossall attempters. A history of childhood abuse was found in 7 HighLethality (43.8%) and 13 Low Lethality attempters (37.1%), with nosignificant difference between groups. Of the 51 subjects, 21(41.5%) were taking psychoactive medication at the time of thescan, with no proportional difference between High and LowLethality subjects.

Among all attempters, Lethality Rating Scale (LRS) scoresranged from 0 to 7 with a mean (S.D.)¼2.80 (1.9) and a medianof 3.0. There were no significant correlations between LRS scores,BIS scores and LHA scores across all attempters. The number oflifetime suicide attempts differed greatly between High and LowLethality attempters. High Lethality subjects had a mean (S.D.) of7.1 (7.0) lifetime attempts compared with 2.3 (1.4) lifetimeattempts for Low Lethality subjects (t¼2.76, d.f.¼15.5, P¼0.014).

However, the groups did not differ significantly in violence of thesuicide method. Overdose, a non-violent method, was the solemeans used by 41 attempters (80.4%), while 10 subjects usedviolent methods on at least one occasion (e.g. cutting (5), hanging(3), jumping (1), and drowning (1)). The mean (S.D.) time from thelast attempt to the scan did not significantly differ between groups(High Lethality: 50.3 (56.2) months; Low Lethality: 72.6 (78.9)months, t¼�1.01, d.f.¼48, P¼n.s.).

3.2. Lethality Rating Scale scores and gray matter volumes

The relationship between LRS scores and gray matter volumeswas assessed by regression analysis in all regions of interest (ROIs).LRS scores were negatively related to gray matter volumes in8 ROIs, with some variations in laterality. Higher degrees oflethality were significantly associated with diminished gray mattervolumes across multiple fronto-temporal–limbic regions, whichincluded (in order of cluster size) the following: bilateral middle-superior temporal cortex, left lingual gyrus, bilateral middle-inferior orbitofrontal cortes, right insula, bilateral fusiform gyrus,right parahippocampal gyrus, left anterior cingulate, and lefthippocampus (Table 2, Fig. 1). There were no significant positivecorrelations between LRS scores and gray matter volumes.

Fig. 1. In ALL BPD attempters: significant clusters (Po0.05, cluster level) showing a negative association between Lethality Rating Scale scores and gray matter arerendered on (a) bilateral lateral cortical surface views, (b) ventral view of the cortical surface, and (c) successive axial slices (z¼�36 to 20). Significant decreases wereobserved in Mid-Inf OFC (1), fusiform (1), insula (2), hippocampus (3), parahippocampus (4), Mid-Inf OFC (5), ACC (6), lingual (7), and mid-sup temp (8).

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

3.3. Personality interactions in High Lethality attempters(Tables 3a and 3b)

Among High Lethality attempters, aggression (LHA) was posi-tively associated with gray matter volumes in large bilateral areasof the middle-inferior orbital frontal cortex (BA 11) and theanterior cingulate cortex. Significant, though much smaller, posi-tive effects were also noted in the right middle-superior temporalcortex (BA 22), right insula, right lingual gyrus, bilateral fusiformgyrus, and right parahippocampus (Table 3a, Fig. 2). There were nosignificant negative effects of aggression on gray matter volumesamong High Lethality attempters.

Impulsivity (BIS) had a positive effect on gray matter volumesin the right middle-superior temporal cortex, with lesser effects onthe left fusiform gyrus and bilateral parahippocampus. A smallnegative effect of impulsivity on gray matter was noted in the rightinsula.

3.4. Personality interactions in Low Lethality attempters

Among Low Lethality attempters, aggression was also positivelyassociated with gray matter volumes, although differing greatlyfrom High Lethality attempters in anatomical locations and clustersizes (i.e., size of correlated area). The most robust findings were inthe right insula, and bilateral fusiform gyrus. Smaller areas ofpositive correlation with aggression were also noted in the righthippocampus, left middle-superior temporal lobe (BA 21), rightparahippocampus, bilateral middle-inferior orbital frontal cortex(BA 11), left lingual gyrus, and right amygdala (Tables 3a and 3b).As with the High Lethality attempters, there were no significant

negative correlations between aggression and gray mattervolumes among Low Lethality attempters.

Impulsivity (BIS) was negatively associated with gray mattervolumes in nine ROIs among Low Lethality attempters, mostwidely in the right middle-superior temporal cortex, bilateralinsula, and bilateral lingual gyrus. Smaller negative effects ofimpulsivity on gray matter were also noted in the left anteriorcingulate, left fusiform gyrus, bilateral parahippocampus, leftmiddle-inferior orbitofrontal cortex, right hippocampus, and rightamygdala (Tables 3a and 3b). There were no positive effects ofimpulsivity on gray matter volumes among Low Lethality attemp-ters. Thus, High and Low Lethality attempters differed in theeffects of impulsivity on gray matter in direction of association(e.g. negative vs. positive correlations), anatomical locations, andcluster sizes (e.g., insula) (Fig. 3).

4. Discussion

4.1. Overview

Neuroimaging studies, using MRI, positron emission tomogra-phy (PET), and functional MRI (fMRI) techniques, have demon-strated abnormalities in structure, metabolism and function insubjects with BPD compared with healthy controls (for review, seeSchmahl and Bremner, 2006). Structural (MRI) and metabolic (PET)abnormalities have been described in impulsive subjects with BPDin areas of the prefrontal cortex (PFC), especially orbitofrontal andventromedial PFC, anterior cingulate, and other fronto-limbicstructures involved in regulation of mood, impulse and behavior(Goyer et al., 1994; De La Fuenta et al., 1997; Juengling et al., 2003;

Fig. 2. Significant clusters (Po0.05, cluster level) in High Lethality attempters (red) and Low Lethality attempters (blue) showing a positive association between graymatter and aggression (LHA) are rendered on (a) bilateral lateral cortical surface views, (b) ventral view of the cortical surface and (c) successive axial slices (z¼�36 to 20).Significant associations shown in fusiform (1), ACC (2), Mid-Inf OFC (3), and insula (4). (For interpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎4

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

Schmahl et al., 2003; Tebartz van Elst et al., 2003; Soloff et al.,2003, 2008; Hazlett et al., 2005). Subjects with BPD (and otherimpulsive PDs), have diminished metabolic responses to seroto-nergic pharmacologic activation by fenfluramine (FEN) or meta-chloropiperazine (mCPP) in related areas, suggesting a neurobio-logical basis for impulsivity and aggression in these patients(Siever et al., 1999; Soloff et al., 2000, 2005; New et al., 2002).Among PD subjects ascertained specifically for impulsive–aggres-sion (e.g., subjects with intermittent explosive disorder), fMRIstudies using angry faces demonstrate increased activation of theamygdala and diminished activation in orbitofrontal cortex (OFC).Among aggressive subjects, there is a loss of the normal amyg-dala–OFC coupling that facilitates cognitive control over affectivearousal (Coccaro et al., 2007). Functional MRI studies in BPDsubjects also demonstrate hyper-arousal of amygdala in responseto negatively valenced faces or pictures (Herpertz et al., 2001;Donegan et al., 2003), and loss of connectivity between theamygdala and the OFC (New et al., 2007). Affective interferencewith executive cognitive functions in BPD is associated with afailure of “top down” cortical inhibition in the presence of “bottomup” limbic hyper-arousal resulting in the characteristic emotionaldysregulation of the BPD patient (Silberzweig et al., 2007). Emo-tional dysregulation increases the vulnerability to impulsive,aggressive and suicidal behavior in BPD.

Our main aim in this study was to map the relationshipbetween impulsivity, aggression and suicidal behavior in key brainstructures among suicidal sub-groups distinguished by the leth-ality of their attempts. We found that higher degrees of medicallethality were related to diminished gray matter volumes acrossmultiple fronto-temporal–limbic regions. The relationships

between impulsivity, aggression and brain structure differedmarkedly between High and Low Lethality attempters. Similarly,the effects of impulsivity and aggression on brain structurediffered markedly within lethality groups.

4.2. Aggression

The effects of aggression on brain structure differed by lethalitystatus in anatomical location and cluster size (a measure of thearea of significant correlation). Since there were no significantdifferences between groups in baseline aggression scores, lethalityof suicidal behavior in BPD may be related to the mediation ofaggression by specific neural networks. Among High Lethalityattempters, increased aggression is most widely associated withincreased gray matter volumes in the middle-inferior OFC and theanterior cingulate gyrus. The OFC is associated with executivecognitive functions such as response inhibition, selective attention,conflict resolution, and monitoring and regulating the limbicsystem (Bonelli and Cummings, 2007). It is part of a complexneural network (along with the fusiform gyrus and amygdala) thatassesses emotion in facial expression and is activated by angryfaces (Blair et al., 1999; Blair, 2004). Injury to the OFC results indisinhibition and in impulsive and aggressive behavior. (Thefamous 1848 case of Phineas Gage illustrates this point [Damasioet al., 1994].) Structural MRI studies in BPD subjects have reporteddiminished volumes in the OFC in adults and adolescents com-pared with healthy controls, and an association between increasedimpulsivity (BIS), diminished gray matter (BA 10), and increasedwhite matter (BA 47) in areas of the OFC (Tebartz van Elst et al.,2003; Hazlett et al., 2005; Chanen et al., 2008). PET studies in BPD

Fig. 3. Low Lethality subjects: significant clusters (Po0.05, cluster level) showing a negative association between impulsivity (BIS) and gray matter are rendered on(a) bilateral lateral cortical surface views, (b) ventral view of the cortical surface and (c) successive axial slices (z¼�36 to 20). Significant decreases are shown in insula (1),fusiform (2), Mid-Inf OFC (3), parahippocampus (4), hippocampus (5), lingual (6), amygdala (7), mid-sup temp (8), and ACC (9).

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

subjects report diminished fluorodeoxyglucose uptake in medialOFC (BA 9, 10, 11) compared with healthy controls. Co-varying forimpulsivity (BIS) rendered these differences non-significant (Soloffet al., 2003).

BPD subjects also have diminished alpha [C-11] mTrptophantrapping (mTrp) (a measure of serotonin synthesis) in the OFCcompared with healthy controls (Leyton et al., 2001) (in thesestudies, impulsivity was inversely related to alpha[C-11] mTrptrapping in the medial frontal gyrus, anterior cingulate, temporalcortex and striatum). Structural and functional imaging studiessuggest a neural mediation for impulsive–aggressive behavior inBPD, related, in part, to diminished serotonergic function inthe OFC.

The OFC has extensive connectivity with the amygdala and, inconcert with other prefrontal structures (such as the DLPFC),moderates amygdala hyper-arousal (Silberzweig et al., 2007). Attimes of negative emotional stress, prefrontal cortical inhibitionfrom the OFC is decreased in subjects with BPD compared withcontrol subjects (Silberzweig et al., 2007; New et al., 2009; Krauset al., 2010). OFC activation decreases in BPD subjects whileimagining self-injurious acts (Kraus et al., 2010). The resultingaffective dysregulation of cognitive control increases the like-lihood of impulsive–aggressive behavior in response to triggerevents.

Among High Lethality attempters, increased aggression wasalso related to increased gray matter volumes in the anteriorcingulate cortex. The anterior cingulate is activated by tasksrequiring choices between competing stimuli (e.g., conflict resolu-tion in the X version of the continuous performance task) (Carteret al., 2000). It is also involved in recognition of affective states,execution of affect-related operations, and self-regulation of emo-tion (Beauregard et al., 2001; Devinsky et al., 1995; Hazlett et al.,2005). In concert with the amygdala, the anterior cingulate isactivated by negative emotions in healthy subjects (e.g., especiallyin dorsal, middle cingulate, perigenual and subgenual cingulate)where it acts to modulate emotion. Among subjects with BPD,negative emotion activates the posterior cingulate gyrus, withdecreased activation of dorsal and right subgenual divisions (forreview, see Ruocco et al., 2012). When presented with fearful facesin an fMRI task, BPD subjects show greater deactivation (comparedwith controls) in the subgenual anterior cingulate cortex bilater-ally, while demonstrating increased activation in the right amyg-dala. This result supports the hypothesis of fronto-limbic

dysfunction in BPD with exaggerated amygdala response anddiminished emotion regulation by the anterior cingulate(Minzenberg et al., 2007). Unfortunately, this study defined onlytwo ROIs: amygdala and anterior cingulate. Other fMRI studieswith BPD subjects have also noted diminished activation of theOFC with negative affective stimuli (Silberzweig et al., 2007; Newet al., 2009; Kraus et al., 2010).

Table 2Relationship between Lethality Rating Scale scores and gray matter concentrationsin BPD attempters.

Anat. ROI Cluster ext. Ind. cluster ext. P value Voxel peak

(TAL) Po0.05 (x, y, z)

ROIs with significant negative correlationsa

Mid-Sup Tp 419 1139 o0.002 (45, �10, �16)Mid-Sup Tp(L) 419 920 o0.002 (�63, �35, 23)Lingual g (L) 238 801 o0.003 (�14, �85, 0)Mid-Inf OFC(L) 274 686 o0.004 (�21, 49, �9)Mid-Inf OFC 274 321 o0.001 (38, 44, �3)Insula 123 439 o0.002 (45, �3, 2)Fusiform g 183 421 o0.009 (30, �57, �5)Fusiform g (L) 183 409 o0.006 (�24, �31, �21)Parahipp. g 99 330 o0.004 (25, �24, �22)ACC(L) 160 281 o0.007 (�9, 26, �3)Hipp.(L) 43 64 o0.013 (�19, �28, �6)

a No significant positive correlations were found between LRS scores and graymatter concentrations. Mid-Sup Tp¼middle superior temporal cortex, mid-infOFC¼middle inferior orbital frontal cortex, Parahipp. g¼parahippocampal gyrus,ACC¼anterior cingulate cortex, and Hipp.¼hippocampus

Table 3aRelationships between aggression (LHA) and gray matter concentrations in High,and Low Lethality attemptersa.

Anat. ROI Cluster ext. Ind. cluster ext. P value Voxel peak(TAL) Po0.05 (x, y, z)

A. High Lethality LHA-positive correlationsMid-Inf OFC 352 4019 o0.003 (24, 37, �11)Mid-Inf OFC(L) 352 1011 o0.003 (�27, 29, �19)ACC 176 3572 o0.002 (9, 37, �4)ACC(L) 176 3572 o0.002 (�11, 42, 0)Mid Sup Tp 398 771 o0.001 (58, �46, 5)Insula 173 568 o0.007 (45, 11, 2)Lingual 165 277 o0.006 (13, �79, �2)Fusiform(L) 136 230 o0.001 (�36, �76, �10)[Fusiform] 136 189 o0.009 (32, �7, �30)Parahipp. 76 201 o0.009 (23, �46, �2)

B. High Lethality LHA-negative correlations: none

C. Low Lethality LHA-positive correlationsInsula 166 868 o0.006 (35, 19, 9)Fusiform(L) 156 643 o0.001 (�35, �24, �19)[Fusiform] 156 601 o0.005 (37, �24, �19)Hippocampus 127 479 o0.014 (32, �23, �10)Mid Sup Tp(L) 390 438 o0.001 (�45, �18, �17)Parahipp 101 347 o0.008 (34, �23, �19)Mid-Inf OFC(L) 213 327 o0.016 (�27, 38, �16)[Mid-Inf OFC] 213 255 o0.002 (27, 49, �15)Lingual(L) 173 206 o0.007 (�4, �57, 5)Amygdala 59 64 o0.019 (23, �8, �6)

D. Low Lethality LHA-negative correlations: none

a ROIs are listed in order of descending ind. cluster ext., except for bilateralROIs, which remain paired.

Table 3bRelationships between impulsivity (BIS) and grey matter concentrations in Highand Low Lethality attempters.

Anat. ROI Cluster ext. Ind. cluster ext. P value Voxel peak(TAL) Po0.05 (x, y, z)

A. High Lethality BIS-positive correlationsMid Sup Tp 390 771 o0.001 (58, �46, 5)Parahipp.(L) 65 232 o0.004 (�16, �28, �8)[Parahipp.] 65 140 o0.001 (14, �28, �6)Fusiform(L) 140 208 o0.005 (�20, �4, �35)

B. High Lethality BIS-negative correlationsInsula 162 162 o0.008 (42, 10, 1)

C. Low Lethality BIS-Positive: none

D. Low Lethality BIS-negative correlationsMid Sup Tp 466 1709 o0.001 (52, �48, �2)Insula 177 1158 o0.002 (40, �3, 1)[Insula(L)] 177 775 o0.007 (�35, 2, �1)Lingual 237 1020 o0.002 (22, �88, �7)[Lingual(L)] 237 327 o0.012 (�17, �63, 0)ACC(L) 151 534 o0.008 (�2, 48, 0)Fusiform(L) 149 503 o0.008 (�27, �46, �2)Parahipp 105 451 o0.007 (31, �28, �7)[Parahipp(L)] 105 159 o0.008 (�27, �45, �2)Mid-Inf OFC(L) 148 336 o0.001 (�28, 31, �3)Hippocampus 96 271 o0.006 (26, �31, 2)Amygdala 60 176 o0.019 (32, �4, �19)

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎6

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

Some, but not all, structural MRI studies in BPD subjects havereported diminished volumes in anterior and ventral cingulatecompared to healthy control subjects (De La Fuenta et al., 1997;Tebartz van Elst et al., 2003; Hazlett et al., 2005; Soloff et al., 2008).Hazlett et al. (2005) found that diminished gray matter volume inthe left anterior cingulate (BA 25), and diminished white mattervolume in the right posterior cingulate (BA 23), correlated withincreased impulsivity (BIS) in subjects with BPD. Although theseresults appear contrary to our own, our VBM analyses did not defineseparate ROIs within the cingulate gyrus.

In marked contrast, aggression among Low Lethality attempterswas most widely associated with gray matter volumes in the rightinsula and bilateral fusiform gyrus. The insula is a limbic “integra-tion area” that is involved in recognition of one's own internalemotional state and perceived emotion in others (i.e. “empathy”)(New et al., 2008). It is also involved in representing negativeemotional states, such as disgust (Augustine, 1996). In structuralstudies, BPD subjects have diminished gray matter in insularcortex compared with healthy controls (Soloff et al., 2008). Highlethality BPD attempters have diminished gray matter in the rightinsula compared with Low Lethality attempters (Soloff et al.,2012). In fMRI studies of BPD subjects, the insula is activated bynegative emotional stimuli (for review, see Ruocco et al., 2012), byparadigms that mimic social exclusion (Eisenberger et al., 2003),by aversive memories (Schnell et al., 2007), and by recall ofunresolved autobiographical memories (Beblo et al., 2006). Socialtasks requiring co-operation activate the anterior insula, which isresponsive to violations of social norms (King-Casas et al., 2008).BPD subjects show diminished activation in the anterior insulacompared with control subjects during an economic exchangegame requiring social co-operation (King-Casas et al., 2008).Impairment in the function of the insula could lead to misjudgingothers' intentions and disinhibited responses to perceived rejec-tion, a core characteristic of BPD and common precipitant forimpulsive suicide attempts.

Aggression among Low Lethality attempters also had a sig-nificant effect on the fusiform gyrus, bilaterally. The fusiform gyrusis primarily associated with facial recognition, and is a componentpart of a complex face-processing system that includes the OFCand the superior temporal cortex, lingual and parahippocampalgyrii, and the amygdala (Radua et al., 2010). This networkprocesses facial recognition and emotion, and it analyzes bodilymovement to assess the intentions of others. (The parahippocam-pus is involved in memory encoding and retrieval of informationconcerning familiarity of social settings, complementing the func-tion of the fusiform face area by adding social and emotionalcontext to a scene [Radua et al., 2010].) Deficits in these functionswould impair social functioning.

Abnormalities in the structure and function of the fusiformgyrus have previously been reported in BPD compared withcontrol subjects. In previous VBM studies, fusiform and parahip-pocampal gyrii were both diminished in volume in High comparedwith Low Lethality BPD attempters (Soloff et al., 2012). In somePET studies (though not all), BPD subjects have demonstrateddiminished metabolism or cerebral blood flow in the fusiformgyrus compared with healthy controls (Leyton et al., 2001;Schmahl et al., 2003; Lange et al., 2005). The fusiform gyrus isactivated bilaterally (with the amygdala) in BPD subjects com-pared with healthy controls in response to aversive stimuli in fMRIprotocols (Herpertz et al., 2001).

Although there is considerable overlap in areas positivelyassociated with aggression in High and Low Lethality attempters,regions with the largest cluster sizes differ markedly betweengroups. Prefrontal regions primarily associated with regulation ofaffect and impulsive–aggressive behavior are correlated withaggression in High Lethality attempters, while aggression in Low

Lethality attempters is associated with limbic areas involved inempathy, social acceptance and co-operation. High Lethalitysuicide attempts are characterized by a subjective intent to die,while Low lethality attempts are generally communicative ges-tures intended to convey emotional distress and coerce a caringresponse.

4.3. Impulsivity

Impulsivity in both High and Low Lethality attempters wascorrelated with gray matter volumes in the middle-superiortemporal cortex, but in opposite directions and markedly differingcluster sizes. The superior temporal cortex (with the insula and thelingual and fusiform gyrii) is involved in facial processing, analysisof intentions of others through bodily movement (Frith and Frith,1999; Allison and McCarthy, 2000), and hyper-vigilance in attach-ment relationships (Buchheim et al., 2008). The superior temporalcortex is part of a complex circuit that mediates reflexiveresponses to visual social inputs, especially negative visual stimuli,such as angry, disgruntled faces (Koenigsberg et al., 2009; Iidakaet al., 2001).

The effects of impulsivity and aggression on brain structurediscriminated High from Low Lethality attempters, despite theabsence of significant differences on psychological measures ofimpulsivity and aggression (BIS and LHA). Differences betweenHigh and Low Lethality attempters are found in the neuralmediation of these personality traits, supporting the view thatthey represent two separate but overlapping groups. High Leth-ality attempters are a commonly used model for completedsuicides and may share with them specific neurobiological vulner-abilities associated with personality risk factors such as impulsiv-ity and aggression. We also demonstrated differences in the effectsof impulsivity and aggression on brain structure within eachlethality group. These differences in affected regions and directionof association may suggest separate neural mediation for thesetraits.

Impulsivity and aggressiveness are related trait dispositions,but not identical constructs. Distinctions between these traits havebeen obscured in the literature through the use of terms such asimpulsive–aggression or aggressive impulsivity, implying a unitarybehavioral dimension for non-premeditated aggression. Coccaro(1992, 1998) has advanced the view that impulsive–aggression is asingle dimensional trait, a dysregulation of impulse control relatedto diminished serotonergic function in the brain. In this view,impulsive–aggression is an endophenotype that increases thelikelihood of aggressive behavior given environmental triggers(McCloskey et al., 2009). In studies of BPD patients, and in large-sample surveys of non-clinical subjects, impulsivity and aggres-siveness have been shown to be separable traits. Impulsivity is ahigher order trait predisposing to non-premeditated aggression,which is a lower order observable behavior (Garcia-Forero et al.,2009; Critchfield et al., 2004). Impulsivity and aggression are riskfactors for suicidal behavior in BPD, with differing effects onspecific neural networks. Our results provide evidence of theneural substrates of impulsivity and aggression in the context ofsuicidality. Specific neural substrates may functionally mediatethese personality traits, a hypothesis that can be more stronglytested using fMRI studies. Lethality of suicidal behavior in BPDmay be determined, in part, by the mediation of these personalitytraits by specific neural networks involved in the regulation ofmood, impulse and behavior.

4.4. Limitations

This study focused on effects of impulsivity and aggression ongray matter volumes in suicidal subjects with BPD and may not be

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

generalizable to suicidal behavior in other disorders. Similarly,personality traits associated with suicidal behavior in otherdisorders (e.g., hopelessness/pessimism in MDD) may have theirown unique relationships to neural structures and lethality, notfound in BPD.

Could differences between High and Low Lethality attemptersbe related to brain injury incurred in high lethality attempts? Thisis a valid concern in all imaging studies of high lethality suicidalsubjects. We are unable to resolve this issue. Although we did notconduct formal neuropsychological testing, all of our subjects wereable to give informed consent, demonstrated normal cognitivefunctioning during diagnostic interviews, self-reports and scan-ning procedures, and had no apparent cognitive impairment intheir everyday lives.

This was a correlational study. Correlation does not provecausation. Functional MRI studies are needed to test effects ofimpulsivity and aggression on brain activation in High and LowLethality suicide attempters with BPD.

Acknowledgments

The research reported was supported by the NIMH grant MH48463 (PHS), the Brain Behavior Research Foundation, the Chil-dren's Hospital Foundation, and the Prechter Pediatric BipolarProgram World Heritage Foundation (VAD).

References

Allison, T., McCarthy, G., 2000. Social perception from visual cues: role of the STSregion. Trends in Cognitive Science 4 (7), 267–277.

Ashburner, J., 2007. A fast diffeomorphic image registration algorithm. NeuroImage38 (1), 95–113.

Augustine, J.R., 1996. Circuitry and functional aspects of the insular lobe in primatesincluding humans. Brain Research Reviews 22, 229–244.

Banissy, M.J., Kanai, R., Walsh, V., Rees, G., 2012. Inter-individual differences inempathy are reflected in human brain structure. NeuroImage 62 (3),2034–2039.

Barratt, E.S., 1965. Factor analysis of some psychometric measures of impulsivenessand anxiety. Psychological Reports 16, 547–554.

Beauregard, M., Levesque, J., Bourgouin, P., 2001. Neural correlates of conscious self-regulation of emotion. Journal of Neuroscience 21, RC165.

Beblo, T., Driessen, M., Mertens, M., Wingenfeld, K., Piefke, M., Rullkoetter, N., Silva-Saavedra, A., Mensebach, C., Reddemann, L., Rau, H., Markowitsch, H.J., Wulff,H., Lange, W., Berea, C., Ollech, I., Woermann, F.G., 2006. Functional MRIcorrelates of the recall of unresolved life events in borderline personalitydisorder. Psychological Medicine 36, 845–856.

Blair, R.J., 2004. The roles of orbital frontal cortex in the modulation of antisocialbehavior. Brain and Cognition 55, 198–208.

Blair, R.J., Morris, J.S., Frith, C.D., Perrett, D.J., Dolan, R.J., 1999. Dissociable neuralresponses to facial expressions of sadness and anger. Brain 122, 883–893.

Bonelli, R.M., Cummings, J.L., 2007. Frontal subcortical circuitry and behavior.Dialogues in Clinical Neuroscience 9 (2), 141–151.

Brown, G.L., Goodwin, F.K., Ballenger, J.C., Goyer, P.F., Major, L.K., 1979. Aggression inhumans correlates with cerebrospinal fluid amine metabolites. PsychiatryResearch 1, 131–139.

Buchheim, A., Erk, S., George, C., Kachele, H., Kircher, T., Martius, P., Pokorny, D.,Ruchsow, M., Spitzer, M., Walter, H., 2008. Neural correlates of attachmenttrauma in borderline personality disorder: a functional magnetic resonanceimaging study. Psychiatry Research: Neuroimaging 163, 223–235.

Carter, C.S., Macdonald, A.M., Botvinik, M., Ross, L.L., Stenger, V.A., Noll, D., Cohen, J.D., 2000. Parsing executive processes: strategic versus evaluative functions ofthe anterior cingulate cortex. Proceedings of the National Academy of Sciencesof the United States of America 97, 1944–1948.

Chanen, A.M., Velakoulis, D., Carison, K., Gaunson, K., Wood, S.J., Yuen, H.P., Yücel,M., Jackson, H.J., McGorry, P.D., Pantelis, C., 2008. Orbitofrontal, amygdala andhippocampal volumes in teenagers with first presentation borderline person-ality disorder. Psychiatry Research: Neuroimaging 163, 116–125.

Coccaro, E., 1992. Impulsive aggression and central serotonergic system function inhumans: an example of a dimensional brain–behavior relationship. Interna-tional Clinical Psychopharmacology 7, 3–12.

Coccaro, E.F., 1998. Impulsive aggression: a behavior in search of a clinicaldefinition. Harvard Review of Psychiatry 5, 1–4.

Coccaro, E.F., McCloskey, M.S., Fitzgerald, D.A., Phan, K.L., 2007. Amygdala andorbitofrontal reactivity to social threat in individuals with impulsive aggres-sion. Biological Psychiatry 62, 168–178.

Critchfield, K., Levy, K., Clarkin, J., 2004. The relationship between impulsivity,aggression, and impulsive-aggression in borderline personality disorder: anempirical analysis of self-report measures. Journal of Personality Disorders 18,555–570.

Damasio, H., Grabowski, T., Frank, R., Galaburda, A., Damasio, A.R., 1994. The returnof Phineas Gage: clues about the brain from the skull of a famous patient.Science 264, 1102–1105.

De La Fuenta, J.M., Goldman, S., Stanus, E., Vizuete, C., Morlan, I., Bobes, J.,Mendlewicz, J., 1997. Brain glucose metabolism in borderline personalitydisorder. Journal of Psychiatric Research 31, 531–541.

Devinsky, O., Morrell, M., Vogt, B., 1995. Contributions of anterior cingulate cortexto behavior. Brain 118, 279–306.

Donegan, N.H., Sanislow, C.A., Blumberg, H.P., Fulbright, R.K., Lacadie, C., Skudlarski,P., Gore, J.C., Olson, I.R., McGlashan, T.H., Wexler, B.E., 2003. Amygdalahyperreactivity in borderline personality disorder: implications for emotionaldysregulation. Biological Psychiatry 54 (11), 1284–1293.

Diwadkar, V.A., Pruitt, P., Goradia, D., Murphy, E., Bakshi, N., Keshavan, M.S., Rajan,U., Reid, A., Zajac-Benitez, C., 2011. Fronto-parietal hypo-activation duringworking memory independent of structural abnormalities: conjoint fMRI andsMRI analyses in adolescent offspring of schizophrenia patients. NeuroImage 58(1), 234–241.

Eisenberger, N.I., Lieberman, M.D., Williams, K.D., 2003. Does rejection hurt? AnfMRI study of social exclusion. Science 302 (5643), 290–292.

Fertuck, E.A., Lenzenweger, M.F., Clarkin, J.F., Hoermann, S., Stanley, B., 2006.Executive neurocognition, memory systems and borderline personality dis-order. Clinical Psychology Review 26, 346–375.

First, M.B., Spitzer, R.L., Gibbon, M., & Williams, J.B.W., 2005. Structured ClinicalInterview for DSM-IV-TR Axis I Disorders-Patient Edition (SCID-I/P, 4/2005revision). Biometrics Research Department, New York State Psychiatric Insti-tute, New York, NY.

Frith, C.D., Frith, U., 1999. Interacting minds – a biological basis. Science 286,1692–1695.

Friston, K.J., Holmes, A.P., Worsely, K.J., Poline, J.B., Frith, C.D., Frackowiak, R.S.J.,1995. Statistical parametric maps in functional imaging: a general approach.Human Brain Mapping 2, 189–210.

Garcia-Forero, C., Gallardo-Pujol, D., Maydeu-Olivares, A., Andres-Pueyo, A., 2009.Disentangling impulsiveness, aggressiveness and impulsive aggression: anempirical approach using self-report measures. Psychiatry Research 168,40–49.

Good, C.D., Johnsrude, I.S., Ashburner, J., Henson, R.N., Friston, K.J., Frackowiak, R.S.,2001. A voxel-based morphometric study of ageing in 465 normal adult humanbrains. NeuroImage 14, 21–36.

Goyer, P.F., Andreason, P.J., Semple, W.E., Clayton, A.H., King, A.C., Compton-Toth, B.A., Schultz, S.C., Cohen, R.M., 1994. Positron-emission tomography and person-ality disorders. Neuropsychopharmacology 10, 21–28.

Gunderson, J.G., Kolb, J.E., Austin, V., 1981. The Diagnostic Interview for Borderlines.American Journal of Psychiatry 138, 896–903.

Guy, W., 1976. ECDEU Assessment Manual of Psychopharmacology-Revised. DHEWPublication No. ADM. 76-338. NIMH, Rockville, MD.

Hazlett, E.A., New, A.S., Newmark, R., Haznedar, M.M., Lo, J.N., Speiser, L.J., Chen, A.D., Mitropoulou, V., Minzenberg, M., Siever, L.J., Buchsbaum, M.S., 2005.Reduced anterior and posterior cingulate gray matter in borderline personalitydisorder. Biological Psychiatry 58, 614–623.

Herpertz, S.C., Dietrich, T.M., Wenning, B., Krings, T., Erberich, S.G., Willmes, K.,Thron, A., Sass, H., 2001. Evidence of abnormal amygdala functioning inborderline personality disorder: a functional MRI study. Biological Psychiatry50, 292–298.

Iidaka, T., Omori, M., Murata, T., Kosaka, H., Yonekura, Y., Okada, T., Sadato, N., 2001.Neural interaction of the amygdala with the prefrontal and temporal cortices inthe processing of facial expressions as revealed by fMRI. Journal of CognitiveNeuroscience 13 (8), 1035–1047.

Juengling, F.D., Schmahl, C., Hesslinger, B., Ebert, D., Bremner, J.D., Gostomzyk, J.,Bohus, M., Lieb, K., 2003. Positron emission tomography in female patients withborderline personality disorder. Journal of Psychiatric Research 37, 109–115.

King-Casas, B., Sharp, C., Loman-Breqam, L., Lohrenz, T., Fonagy, P., Montague, P.R.,2008. The rupture and repair of co-operation in borderline personality disorder.Science 321, 806–810.

Klein, A., Andersson, J., Ardekani, B.A., Ashburner, J., Avants, B., Chiang, M.C.,Christensen, G.E., Collins, D.L., Gee, J., Hellier, P., Song, J.H., Jenkinson, M.,Lepage, C., Rueckert, D., Thompson, P., Vercauteren, T., Woods, R.P., Mann, J.J.,Parsey, R.V., 2009. Evaluation of 14 nonlinear deformation algorithms appliedto human brain MRI registration. NeuroImage 46 (3), 786–802.

Koenigsberg, H.W., Siever, L.J., Lee, H., Pizzarello, S., New, A.S., Goodman, M., Cheng,H., Flory, J., Prohovnik, I., 2009. Neural correlates of emotion processing inborderline personality disorder. Psychiatry Research: Neuroimaging 172,192–199.

Kraus, A., Valerius, G., Seifritz, E., Ruf, M., Bremner, J.D., Bohus, M., Schmahl, C.,2010. Script-driven imagery of self-injurious behavior in patients with border-line personality disorder: a pilot fMRI study. Acta Psychiatrica Scandinavica121, 41–51.

Lange, C., Kracht, L., Herholz, K., Sachsse, U., Irle, E., 2005. Reduced glucosemetabolism in temporo-parietal cortices of women with borderline personalitydisorder. Psychiatry Research: Neuroimaging 139, 115–126.

Leichsenring, F., Leibing, E., Kruse, J., New, A.S., Leweke, F., 2011. Borderlinepersonality disorder. Lancet 377 (9759), 74–84.

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎8

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i

Leyton, M., Okazawa, H., Diksic, M., Paris, J., Rosa, P., Mzengeza, S., Young, S.N., Blier,B., Benkelfat, C., 2001. Brain regional alpha-[11C] methyl-L-tryptophan trappingin impulsive subjects with borderline personality disorder. American Journal ofPsychiatry 158, 775–782.

Loranger, A.W., Janca, A., Sartorius, N., 1997. Assessment and Diagnosis ofPersonality Disorder: The ICD-10 International Personality Disorder Examina-tion (IPDE). Cambridge University Press, Cambridge

Mann, J.J., 2003. Neurobiology of suicidal behavior. Nature Reviews 4, 819–828.Mann, J.J., Waternaux, C., Haas, G.L., Malone, K.M., 1999. Toward a clinical model of

suicidal behavior in psychiatric patients. American Journal of Psychiatry 156,181–189.

Maris, R.W., Berman, A.L., Silverman, M.M., 2000. Comprehensive Textbook ofSuicidology. Guildford Press, New York p. 19

McCloskey, M.S., New, A.S., Siever, L.J., Goodman, M., Koenigsberg, H.W., Flory, J.D.,Coccaro, E.F., 2009. Evaluation of behavioral impulsivity and aggression tasks asendophenotypes for borderline personality disorder. Journal of PsychiatricResearch 43, 1036–1048.

Minzenberg, M.J., Fan, J., New, A.S., Tang, C.Y., Siever, L.J., 2007. Fronto-limbicdysfunction in response to facial emotion in borderline personality disorder: anevent-related fMRI study. Psychiatry Research: Neuroimaging 155, 231–243.

New, A.S., Goodman, M., Triebwasser, J., Siever, L.J., 2008. Recent advances in thebiological study of personality disorders. Psychiatric Clinics of North America31, 441–461.

New, A.S., Hazlett, E.A., Buchsbaum, M.S., Goodman, M., Mitelman, S.A., Newmark,R., Trisdorfer, R., Haznedar, M.M., Koenigsberg, H.W., Flory, J., Siever, L.J., 2007.Amygdala-prefrontal disconnecrtion in borderline personality disorder. Neu-ropsychopharmacology 32, 1629–1640.

New, A.S., Hazlett, E.A., Buchsbaum, M.S., Goodman, M., Reynolds, D., Mitropoulou,V., Sprung, L., Shaw, R.B., Koeningsberg, H., Platholi, J., Silverman, J., Siever, L.J.,2002. Blunted prefrontal cortical 18-flurodeoxyglucose positron emissiontomography response to meta-chlorophenylpiperazine in impulsive aggression.Archives of General Psychiatry 59, 621–629.

New, A.S., Hazlett, E.A., Newmark, R.E., Zhang, J., Triebwasser, J., Meyerson, D.,Lazarus, S., Trisdorfrer, R., Goldstein, K.E., Goodman, M., Koenigsberg, H.W.,Flory, J.D., Siever, L.J., Buchsbaum, M.S., 2009. Laboratory induced aggression: apositron emission tomography study of aggressive individuals with borderlinepersonality disorder. Biological Psychiatry 66, 1107–1114.

Oquendo, M.A., Halberstam, B., Mann, J.J., 2003. Risk factors for suicidal behavior.In: First, M.B. (Ed.), Standardized Evaluation in Clinical Practice. Review ofPsychiatry, vol. 22. American Psychiatric Publishing, Inc., Washington, DC,pp. 103–129.

Radua, J., Phillips, M.L., Russell, T., Lawrence, N.S., Marshall, N., Kalidindi, S., El-Hage,W., McDonald, C., Giampietro, V., Brammer, M.J., David, A.S., Surguladze, S.A.,2010. Neural response to specific components of fearful faces in healthy andschizophrenic adults. NeuroImage 49, 939–946.

Ruocco, A.C., Amirthavasagam, S., Choi-Kain, L.W., McMain, S.F., 2012. Neuralcorrelates of negative emotionality in borderline personality disorder: anactivation-likelihood-estimation meta-analysis. Biological Psychiatry 73,153–160.

Schmahl, C.G., Elzinga, B.M., Vermetten, E., Sanislow, C., McGlashen, T.H., Bremner,J.D., 2003. Neural correlates of memories of abandonment in women with andwithout borderline personality disorder. Biological Psychiatry 54, 142–151.

Schmahl, C., Bremner, J.D., 2006. Neuroimaging in borderline personality disorder.Journal of Psychiatric Research 40, 419–427.

Schnell, K., Dietrich, T., Schnitker, R., Daumann, J., Herpertz, S.C., 2007. Processing ofautobiographical memory retrieval cues in borderline personality disorder.Journal of Affective Disorders 97, 253–259.

Siever, L.J., Buchsbaum, M., New, A.S., Spiegel-Cohen, J., Wei, T., Hazlett, E., Sevin, E.,Nunn, M., Mitropoulou, M., 1999. d,L-Fenfluramine response in impulsivepersonality disorder assessed with [18-F] fluorodeoxyglucose positron emissiontomography. Neuropsychopharmacology 20, 413–423.

Silberzweig, D., Clarkin, J.F., Goldstein, M., Kernberg, O.F., Tuescher, O., Levy, K.N.,Brendel, G., Pan, H., Beutel, M., Pavony, M.T., Epstein, J., Lenzenweger, M.F.,Thomas, K.M., Posner, M.I., Stern, E., 2007. Failure of fronto-limbic inhibitoryfunction in the context of negative emotion in borderline personality disorder.American Journal of Psychiatry 164, 1832–1841.

Soloff, P.H., Nutche, J., Goradia, D., Diwadkar, V.A., 2008. Structural brain abnorm-alities in borderline personality disorder: a voxel-based morphometry study.Psychiatry Research: Neuroimaging 164, 223–236 (PMCID: NIHMS83690).

Soloff, P.H., Meltzer, C.C., Becker, C., Greer, P., Constantine, D., 2005. Genderdifferences in a fenfluramine-activated FDG-PET study of borderline personalitydisorder. Psychiatry Research: Neuroimaging 138, 183–195.

Soloff, P.H., Meltzer, C.C., Becker, C., Greer, P.J., Kelly, T.M., Constantine, D., 2003.Impulsivity and prefrontal hypometabolism in borderline personality disorder.Psychiatry Research: Neuroimaging 123, 153–163.

Soloff, P.H., Meltzer, C.C., Greer, P.J., Constantine, D., Kelly, T.M., 2000. Fenfluramine-activated FDG study of borderline personality disorder. Biological Psychiatry 47,540–547.

Soloff, P.H., Pruitt, P., Sharma, M., Radwan, J., White, R., Diwadkar, V.A., 2012.Structural brain abnormalities and suicidal behavior in borderline personalitydisorder. Journal of Psychiatric Research 46 (4), 516–525.

Spitzer, R.L., Williams, J.B.W., Gibbon, M., First, M.B., 1988. Instruction Manual forthe Structured Clinical Interview for DSM III-R (SCID). New York StatePsychiatric Institute, Biometrics Research, New York, NY.

Swartz, M., Blazer, D., George, L., Winfield, I., 1990. Estimating the prevalence ofborderline personality disorder in the community. Journal of PersonalityDisorders 4, 257–272.

Tebartz van Elst, L., Hesslinger, B., Thiel, T., Geiger, E., Haegele, K., Lemieux, L., Lieb,K., Bohus, M., Hennig, J., Ebert, D., 2003. Frontolimbic brain abnormalities inpatients with borderline personality disorder. A volumetric MRI study. Biolo-gical Psychiatry 54 (2), 163–171.

Ward, B.D., 2000. Simultaneous Inference for fMRI Data. Medical College ofWisconsin, Milwaukee, WI

Zanarini, M., Gunderson, J.G., Frankenburg, F.R., Chauncey, D.L., 1989. The reviseddiagnostic interview for borderlines: discriminating BPD from other Axis IIdisorders. Journal of Personality Disorders 3 (1), 10–18.

P. Soloff et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 9

Please cite this article as: Soloff, P., et al., Impulsivity, aggression and brain structure in high and low lethality suicide attempters withborderline personality disorder. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.02.006i