Reduced occipital GABA in Parkinson diseasewith visual hallucinationsMichael J Firbank PhD Jehill Parikh PhD Nicholas Murphy PhD Alison Killen MSc Charlotte L Allan MD
Daniel Collerton MSc Andrew M Blamire PhDDagger and John-Paul Taylor PhDDagger
AbstractObjectiveTo investigate the relationship between visual hallucinations in Parkinson disease (PD) andlevels of γ-aminobutyric acid (GABA) in the primary visual cortex
MethodsWe utilized magnetic resonance spectroscopy to investigate occipital GABA levels in 36 par-ticipants with PD 19 with and 17 without complex visual hallucinations together with 20healthy controls without hallucinations In addition we acquired T1-weighted MRI whole-brain fMRI during a visual task and diffusion tensor imaging
ResultsWe found lower GABA+creatine in PD with visual hallucinations (0091 plusmn 0010) vs thosewithout (0101 plusmn 0010) and controls (0099 plusmn 0010) (F249 = 45 p = 0016) Reduced graymatter in the hallucinations group was also observed in the anterior temporal lobe Althoughthere were widespread reductions in white matter integrity in the visual hallucinations groupthis was no longer significant after controlling for cognitive function
ConclusionsThe data suggest that reduced levels of GABA are associated with visual hallucinations in PDand implicate changes to the ventral visual stream in the genesis of visual hallucinationsModulation of visual cortical excitability through for example pharmacologic interventionmay be a promising treatment avenue to explore
RELATED ARTICLE
EditorialGABA and hallucinationsin Parkinson disease Whois that sitting on my chair
Page 293
These authors contributed equally to this work as first authors
DaggerThese authors contributed equally to this work as senior authors
From the Institute of Neuroscience (MJF AK CLA DC J-PT) Newcastle University Campus for Ageing and Vitality Newcastle upon Tyne and Newcastle Magnetic ResonanceCentre (JP AMB) Institute of Cellular Medicine Newcastle University Newcastle upon Tyne UK and Baylor College of Medicine (NM) Houston TX
Go to NeurologyorgN for full disclosures Funding information and disclosures deemed relevant by the authors if any are provided at the end of the article
The Article Processing Charge was funded by the Newcastle NIHR Biomedical Research Centre
This is an open access article distributed under the terms of the Creative Commons Attribution License 40 (CC BY) which permits unrestricted use distribution and reproduction in anymedium provided the original work is properly cited
Copyright copy 2018 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology e675
Visual hallucinations are common in Parkinson disease(PD)1 particularly as the disease advances and range fromrelatively simple flashes of light or color to more complexhallucinations that are typically well-formed images2 Anumber of models34 have been proposed to explain the originof complex recurrent visual hallucinations Unifying elementsacross the models include breakdown in communication be-tween cortical regions involved in visual processing and al-teration in the weighting of internal vs external input
A neuropathologic study of the visual system in dementia withLewy bodies (DLB) found reduced GABAergic activity in theprimary visual cortex5 This may be an adaption to poor visualinput or disrupted connectivity with other visual areas withreduced GABAergic inhibition maintaining the ability torecognize objects at a cost of seeing things that are not there
The aims of this study were therefore to investigate γ-ami-nobutyric acid (GABA)+ levels using magnetic resonancespectroscopy (MRS) in the occipital lobe of patients with PDwith and without complex visual hallucinations along withsimilarly aged healthy participants We also used structuraldiffusion and fMRI with a flashing checkerboard paradigm tocomprehensively investigate brain changes in people withhallucinations
We hypothesized that (1) GABA+ would be reduced in par-ticipants with visual hallucinations and correlated with hal-lucination symptom severity (2) GABA+ levels wouldinversely correlate with fMRI activation and (3) in the visualsystem there would be brain atrophy and disruption of whitematter fibers in those with visual hallucinations and thesechanges would correlate with GABA+
MethodsParticipants
Inclusion and exclusion criteriaWe prospectively recruited 45 participants between 2014 and2017 with PD aged 60 years and older with a Mini-MentalState Examination (MMSE) score gt12 from a population oflocal community-dwelling participants who had been referredto local neurology and old age psychiatry services Twenty-one healthy controls were identified from spouses and friends
of participants included in this and previous studies Diagnosisof PDwas made according to the UKBrain Bank criteria6 withany cognitive impairment either diagnosed as mild cognitiveimpairment according to the Movement Disorder Societylevel 1 criteria7 or dementia according to the diagnostic cri-teria for PD dementia8 Clinical diagnoses were confirmed byan independent and experienced clinician
Control participants in the study showed no evidence of de-mentia (from their history and CAMCOG [CambridgeCognition Examination] score gt80) For all participants ex-clusion criteria included contraindications for MRI history ofalcohol or substance misuse moderate to severe visual im-pairment significant non-PDndashrelated psychiatric or neuro-logic history moderate to severe cerebral small vessel diseaseimaging evidence of focal brain lesions or the presence ofother unstable or severe medical illness The sample size waschosen to give 85 power to detect a 125 difference inGABA concentration9
Clinical assessmentGlobal cognitive function was assessed using the CAMCOGand MMSE The presence and severity of any extrapyramidalsigns were graded using the motor component of the UnifiedParkinsonrsquos Disease Rating Scale (UPDRS-III)
All participants had their near visual acuity measured with theSnellen chart and Landolt broken rings (test distance 40 cm)after correction of any refractive errors Participants wereexcluded if they had significant visual impairment that couldnot be corrected We used the best performance across alltests for each participant as a measure of acuity in groupcomparisons Computerized tests of visuoperceptual functionincluded angle and motion discrimination tasks which haveestablished metrics in Lewy body disease and have beenreported in a number of our reports1011
As in previous work12 for assessment of visual hallucinationsthe hallucinations subscale of the Neuropsychiatric Inventory(NPIhall)13 was used with specific reference to visual halluci-nations occurring in the previous month Subsequently we usedthe derived NPIhall score (frequency times severity of hallucina-tions) in analyses For reliability purposes patients and carerswere independently asked about the occurrence of visual hal-lucinations in the month before MRI using screening questions
GlossaryBOLD = blood oxygen levelndashdependent CAMCOG = Cambridge Cognition Examination Cr = creatine DLB = dementiawith Lewy bodies DTI = diffusion tensor imaging FA = fractional anisotropy FWE = family-wise error GABA =γ-aminobutyric acid 5-HT = 5-hydroxytryptamine MAYO = Mayo Fluctuations Composite Score MD = mean diffusivityMMSE = Mini-Mental State Examination MRS = magnetic resonance spectroscopy NAA = N-acetylaspartate NPI =Neuropsychiatric Inventory NPIhall = Neuropsychiatric Inventory hallucinations subscale PD = Parkinson disease PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinatorROI = region of interestSPM = Statistical Parametric Mapping TBSS = tract-based spatial statistics TE = echo time TR = repetition time UPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VBM = voxel-based morphometry
e676 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
originating from theNorth-East Visual Hallucinations InterviewIII14 any discrepancies in the accounts of hallucinations be-tween patient and carerfamily member were discussed withboth parties and the assessor with reformulation of NPIhall testscores (with primacy given to the opinion of the caregiverwhere the patient seemed to lack insight)
We also used the noise pareidolia test15 as this has been shownto correlate with tendency to hallucinate in Lewy body de-mentia For this test participants were shown a series of 40black and white images These all contain cloud-like noiseformations and in 8 images there is a face inserted some-where (the location and face differ for each case) After 3example images participants were shown the 40 images one ata time and asked whether they could see any faces If so theywere asked to indicate the location of the face(s) We scoredthe test according to the number of illusory faces seen(ie faces indicated where there were not faces inserted in theimage) Cognitive fluctuations were quantified with the MayoFluctuations Composite Score (MAYO)16 and the ClinicianAssessment of Fluctuation scale17
Participants were classed as active visual hallucinators (PD-VH) if they had complex visual hallucinations in the monthpreceding their interview otherwise they were classed asnonhallucinators (controls and PD-nonVH) Since our fo-cus was on complex visual hallucinations those with minorhallucinations (eg passage or feeling of presence) but nocomplex hallucinations in the last month were included inthe PD-nonVH group We made this distinction sincepassage and feeling of presence probably have a differentetiologic basis to complex hallucinations even though mi-nor visual hallucinations typically precede complex visualhallucinations1819
Standard protocol approvals registrationsand patient consentsThe study was approved by the local ethics committee andwritten consent was obtained from all participants (or nom-inated independent mental capacity advocate where partici-pant lacked capacity)
MRI acquisitionParticipants were scanned on a 3T whole-body MR scanner(Achieva scanner Philips Medical Systems Best the Neth-erlands) with body coil transmission and an 8-channel headcoil receiver
Magnetic resonance spectroscopyWe used the MEGA-PRESS technique20 as previouslyreported21 with a sinc gaussian editing pulse applied alter-nately at 19 ppm (EDIT-ON) and 75 ppm (EDIT-OFF)Subtraction of the EDIT-OFF from EDIT-ON spectra allowsthe 3-ppm GABA+ signal to be separated from the overlyingcreatine peak MEGA-PRESS spectra were acquired froma voxel sized 45 times 32 times 20 mm centered on the midline of theoccipital lobe (data available from Newcastle University
e-prints [figure 1] eprintnclacuk247552) Sequenceparameters were as follows repetition time (TR) = 2000milliseconds (ms) echo time (TE) = 68 ms 320 averagesacquisition bandwidth = 1000 Hz VAPOR (variable powerradiofrequency pulses with optimized relaxation delays) watersuppression22 Macromolecular suppression editing23 was notperformed and thus our results are of GABA+ (ie GABAplus macromolecules) The magnetic resonance GABA signalis thought to reflect concentrations of metabolic GABA andlevels of ambient extracellular GABA that contribute to tonicGABAergic activity24
Structural and functional magnetic resonanceWe acquired images including a whole-brain structural3-dimensional MPRAGE (magnetization-prepared rapid-acquisition gradient echo) scan with sagittal acquisitionslice thickness 10 mm in-plane resolution 10 times 10 mm TR= 83 ms TE = 46 ms flip angle = 8deg and SENSE factor = 2fMRI data were collected with a gradient-echo echo planarimaging sequence (TR = 192 seconds TE = 40 ms field ofview 192 times 192 mm2 64 times 64 matrix size flip angle 90deg 27slices slice thickness 3mm slice gap 1mm)with 100 volumes(192 seconds) as participants looked at the checkerboardstimulus
Diffusion tensor imaging (DTI) acquisitions utilized a 2-di-mensional spin-echo echo planar imaging diffusion-weightedsequence with 59 slices TR = 6100ms TE = 70ms flip angle= 90deg field of view = 270 times 270mm pixel size = 21 times 21 mmslice thickness = 21 mm Diffusion weighting was applied in64 uniformly distributed directions (diffusion b = 1000smiddotmmminus2) and there were 6 acquisitions with no diffusionweighting (b = 0 smiddotmmminus2) We also collected an identicalimage with b = 0 smiddotmmminus2 but with the phase encoding di-rection reversed for distortion correction purposes
fMRI stimulus presentationfMRI was performed with the same checkerboard paradigm aswe have previously utilized25 fMRI-compatible goggles withlenses that ranged from minus40 to 40 diopters (05 increment)were used to correct any refractive errors that partic-ipants had
The stimulus presentation was controlled by the psycho-physics toolbox26 (psychtoolboxorg) extension for MAT-LAB (MathWorks Natick MA) A block design was usedwith a full field circular checkerboard stimulus consisting offive 192-second blocks of a black-and-white checkerboard(inverting at 75 Hz) alternating with five 192-second base-line blocks of a gray screen Participants were asked to focuson a central cross-hair
Magnetic resonance analysis
Spectroscopy data analysisGABA+ quantification was performed using the Gannettoolbox for MATLAB27 and consisted of the following steps(1) alignment of each pair (EDIT-ON and EDIT-OFF) of
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e677
spectra28 (2) subtraction of aligned spectra to produceGABA+ spectra followed by averaging across acquisitions(3) fitting a gaussian to the 3-ppm GABA+ peak to quantifyGABA+ based on the area under the curve For a typicaledited spectrum and Gaussian fit see data available fromNewcastle University e-prints (figure 1) eprintnclacuk247552
Choline creatine and NAA (N-acetylaspartate) amplitudeswere quantified from nonedited spectra only using theAMARES (Advanced Method for Accurate Robust and Ef-ficient Spectral fitting of MRS data with use of prior knowl-edge) algorithm from jMRUI (java-based magnetic resonanceuser interface)29 GABA+ and NAA were expressed as ratiosand normalized to creatine MRS fit quality was assessed by anexperienced physicist as described previously21
fMRI analysisImaging data were processed with Statistical ParametricMapping (SPM)12 (filionuclacukspm) similar to ourprevious work25 For each participant the T1 anatomicalimage was segmented and spatially normalized in SPM usingthe default parameters The fMRI data for each stimuluscondition were slice timing corrected motion corrected byaligning all functional images together and then coregisteredwith the T1 anatomical image The spatial normalizationparameters from the T1 segmentation were used to transformthe fMRI data to standard space with a voxel size of 3 times 3 times3 mm The normalized images were then smoothed with a 6 times6 times 6 mm full width at half maximum gaussian kernel A high-pass filter of 128 seconds was used and serial correlationswere removed with SPMrsquos AR(1) model
The general linear model in SPM was used to conducta whole-brain analysis of the fMRI data We created a designmatrix by convolving the time course of the checkerboardblock with the canonical hemodynamic response function andits first derivative The 6 parameters from the motion cor-rection were included in the design matrix as covariates of nointerest Individual participant and second-level (random-effects) group analyses were conducted Contrast images weregenerated from β estimates for the comparison of checker-board vs baseline Results are shown with a voxelwisethreshold of p lt 0001 (uncorrected) followed by clusterwisethreshold of p lt 005 family-wise error (FWE)-corrected formultiple comparisons
We also used a region of interest (ROI) analysis focusing onthe visual areas Five ROIs in MNI (Montreal NeurologicalInstitute) space were defined averaging across left and righthemispheres V1 V2 V3 V4 and V5 were taken from theSPM Anatomy toolbox (fz-juelichdeinminm-1DEFor-schung_docsSPMAnatomyToolboxSPMAnatomyTool-box_nodehtml) We also defined 3 ROIs from the overallactivation across all participantsmdashall voxels with activation inthe occipital lobe all voxels with associated deactivation (bothwith voxelwise threshold at p lt 005 FWE-corrected) and
a bilateral LGN (lateral geniculate nucleus) region (voxelwisep lt 0001 uncorrected)
Structural MRI analysis
Gray matterFor analysis of gray matter atrophy the T1-weighted struc-tural images were segmented with the SPM12 segment tooland then processed using the DARTEL (Diffeomorphic An-atomical Registration Through Exponentiated Lie Algebra)Toolbox to create a group-specific template to which theindividual images were spatially normalized Images weremodulated to preserve the total tissue amount during nor-malization and smoothed with an 8-mm gaussian filter Weused the SPM Anatomy toolbox30 to identify location ofsignificant clusters For each participant we also extracted thefraction of gray matter white matter and CSF within theindividual spectroscopy voxel location
Diffusion white matterDTI data were processed using FSL (fslfmriboxacukfslfslwiki) using the topup program to correct susceptibility-induced distortions using the 2 b = 0 smiddotmmminus2 images withopposite phase encoding The eddy package was then used tocorrect images for eddy current distortion movement andmotion-induced signal dropout Fractional anisotropy (FA)andmean diffusivity (MD) were then calculated with the dtifitsoftware and the TBSS (tract-based spatial statistics) pack-age31 used to align the FA images all together create a whitematter skeleton of major tracts and extract FA andMD valuesfor each participant on the white matter skeleton The imageswere visually inspected at each stage
Statistical analysisROI data and clinical variables were analyzed with the Sta-tistical Package for Social Sciences (SPSS version 19 IBMCorp Armonk NY) Independent t tests or analysis of vari-ance was used to compare groups for continuous variablesSpearman rank correlation coefficient was used to comparecontinuous variables
For the fMRI voxelwise data a 3-group analysis of variancewas performed using SPM to determine overall activationpatterns and investigate group differences Voxel-based mor-phometry (VBM) was done using SPM on the smoothedmodulated gray matter images using a 3-group analysis ofcovariance with age and intracranial volume (sum of CSFgray matter and white matter) as covariates and also with theaddition of CAMCOG as a measure of cognitive function Weinvestigated with SPM the relationship of GABA+creatine(Cr) with voxelwise gray matter volume with covariates ofage intracranial volume and group (PD-nonVH PD-VH andcontrols)
For the diffusion analysis voxelwise differences inMD and FAbetween groups (with age as a covariate) were estimated usingthe FSL randomise package We also used this to look atvoxelwise correlations with occipital GABA+Cr controlling
e678 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
for age and group (PD-nonVH PD-VH and control) Thiswas repeated with the addition of CAMCOG as a measure ofcognitive function
Data availabilityAnonymized data on which this article is based will be sharedon request with any appropriately qualified investigator
ResultsMRI scans were obtained on 20 of the controls and 36 par-ticipants with PD of whom 15 had mild cognitive impairmentand 21 had PD dementia Table 1 shows the demographics forthese participants There were no significant differences in ageor sex between hallucination groups and there were no dif-ferences in duration of PD or levodopa dose between the PD-VH and PD-nonVH groups However the PD-VH group hadworse motor function according to the UPDRS-III score (p lt0001) worse cognition on the CAMCOG scale (p lt 0001)
and were more likely to be taking cholinesterase inhibitors (p= 0025) As expected the PD-VH group had significantlyhigher hallucination scores and were more likely to havemisperceptions on the pareidolia test compared to the PD-nonVH group Data available from Newcastle Universitye-prints (table 1 eprintnclacuk247552) compare thepatients with PD with vs without dementia There were nosignificant differences in age sex years of education durationof PD or levodopa dose The participants with dementia hada higher UPDRS-III score poorer vision and a greater ten-dency to hallucinate as indicated by the NeuropsychiatricInventory (NPI) and pareidolia test
We excluded 4 participants (1 PD-VH 2 PD-nonVH 1control) from the spectroscopy analysis because they didnot meet MRS quality-assurance criteria21 There were nosignificant differences in age sex or diagnosis of dementiabetween the excluded and nonexcluded participantsTable 2 shows the ratio of GABA+ and NAA to Cr for the
Table 1 Demographics
Control (n = 20) PD-nonVH (n = 19) PD-VH (n = 17)
Age y 754 (50) 723 (51) 755 (45) F253 = 26 p = 0083
Female n () 6 (30) 2 (11) 4 (24) χ2 = 18 p = 041
Dementia n () 0 7 (37) 15 (83) FET p = 0007
Education y 135 (21) 111 (15) 116 (22) F253 = 87 p = 0001ab
Duration of PD y mdash 96 (65) 110 (74) F134 = 04 p = 054
Levodopa dose in 24 h mdash 6735 (4282) 7173 (4217) F134 = 01 p = 076
ChEI n () mdash 2 (11) 8 (47) FET p = 0025
Antipsychotics n () mdash 0 (0) 3 (18) FET p = 0095
UPDRS-III total score 22 (25) 347 (188) 559 (193) F253 = 581 p lt 0001abc
CAMCOG total score 952 (70) 845 (114) 746 (153) F253 = 147 p lt 0001abc
NPI total (A times B) hallucinations mdash 01 (02) 35 (24) F134 = 396 p lt 0001
NPI total score mdash 75 (66) 225 (178) F131 = 105 p = 0003
CAF total mdash 21 (27) 58 (38) F133 = 118 p = 0002
MAYO total mdash 14 (11) 24 (15) F133 = 60 p = 0020
Abbreviations CAF = Clinician Assessment of Fluctuation scale CAMCOG = Cambridge Cognition Examination ChEI = cholinesterase inhibitor FET = Fisherexact test MAYO = Mayo Fluctuations Composite Score MMSE = Mini-Mental State Examination NPI = Neuropsychiatric Inventory PD = Parkinson diseaseUPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VH = visual hallucinationThe pareidolia task result is the number of pareidolias seenSignificant (p lt 005) Tukey post hoc testsa Control vs PD-nonVHbControl vs PD-VHcPD-VH vs PD-nonVH
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e679
groups There was a significant group difference in theGABA+Cr ratio (figure 1) with post hoc Tukey testfinding GABA+Cr reduced in PD-VH relative to PD-nonVH The group difference remained significant afterincluding CAMCOG as a measure of cognitive ability in thelinear model (F248 = 327 p = 0047) There were no sig-nificant differences in gray or white matter proportionwithin the voxel between groups (table 2) Within the PDgroup GABA+Cr correlated with visual acuity (Spearmanρ = 04 p = 0025) MMSE (ρ = 035 p = 0047) UPDRS-III(ρ = minus0345 p = 0049) and MAYO total score (ρ = minus0627p lt 0001) There were no significant correlations (p gt 01)with disease duration CAMCOG angle or motion test orthe pareidolia test There were no significant correlationsbetween GABA+Cr and NPI hallucination score aftercontrolling for VH group
There was no significant difference in GABA+Cr levels be-tween participants with PD taking cholinesterase inhibitors vsthose not taking them (0098 SD 0012 vs 0093 SD 0008 t31= 116 p = 026) between those taking antipsychotic agents(quetiapine) vs those not (0088 SD 0007 vs 0097 SD 0011t31 = 142 p = 017) and there was no significant correlationbetween levodopa dose and GABA+Cr (ρ = 0018 p = 09)
The fMRI scans were not acquired on one PD participant whodid not tolerate the full scanning session and 4 PD participantsrsquoscans were excluded because of excessive motion leaving 17PD-VH 14 PD-nonVH and 20 controls with usable fMRI dataAll groups showed a typical activation pattern to the checker-board (data available fromNewcastle University e-prints [figure2] eprintnclacuk247552) but there were no significant dif-ferences in activation between any groups In the ROI analysis
e680 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
Visual hallucinations are common in Parkinson disease(PD)1 particularly as the disease advances and range fromrelatively simple flashes of light or color to more complexhallucinations that are typically well-formed images2 Anumber of models34 have been proposed to explain the originof complex recurrent visual hallucinations Unifying elementsacross the models include breakdown in communication be-tween cortical regions involved in visual processing and al-teration in the weighting of internal vs external input
A neuropathologic study of the visual system in dementia withLewy bodies (DLB) found reduced GABAergic activity in theprimary visual cortex5 This may be an adaption to poor visualinput or disrupted connectivity with other visual areas withreduced GABAergic inhibition maintaining the ability torecognize objects at a cost of seeing things that are not there
The aims of this study were therefore to investigate γ-ami-nobutyric acid (GABA)+ levels using magnetic resonancespectroscopy (MRS) in the occipital lobe of patients with PDwith and without complex visual hallucinations along withsimilarly aged healthy participants We also used structuraldiffusion and fMRI with a flashing checkerboard paradigm tocomprehensively investigate brain changes in people withhallucinations
We hypothesized that (1) GABA+ would be reduced in par-ticipants with visual hallucinations and correlated with hal-lucination symptom severity (2) GABA+ levels wouldinversely correlate with fMRI activation and (3) in the visualsystem there would be brain atrophy and disruption of whitematter fibers in those with visual hallucinations and thesechanges would correlate with GABA+
MethodsParticipants
Inclusion and exclusion criteriaWe prospectively recruited 45 participants between 2014 and2017 with PD aged 60 years and older with a Mini-MentalState Examination (MMSE) score gt12 from a population oflocal community-dwelling participants who had been referredto local neurology and old age psychiatry services Twenty-one healthy controls were identified from spouses and friends
of participants included in this and previous studies Diagnosisof PDwas made according to the UKBrain Bank criteria6 withany cognitive impairment either diagnosed as mild cognitiveimpairment according to the Movement Disorder Societylevel 1 criteria7 or dementia according to the diagnostic cri-teria for PD dementia8 Clinical diagnoses were confirmed byan independent and experienced clinician
Control participants in the study showed no evidence of de-mentia (from their history and CAMCOG [CambridgeCognition Examination] score gt80) For all participants ex-clusion criteria included contraindications for MRI history ofalcohol or substance misuse moderate to severe visual im-pairment significant non-PDndashrelated psychiatric or neuro-logic history moderate to severe cerebral small vessel diseaseimaging evidence of focal brain lesions or the presence ofother unstable or severe medical illness The sample size waschosen to give 85 power to detect a 125 difference inGABA concentration9
Clinical assessmentGlobal cognitive function was assessed using the CAMCOGand MMSE The presence and severity of any extrapyramidalsigns were graded using the motor component of the UnifiedParkinsonrsquos Disease Rating Scale (UPDRS-III)
All participants had their near visual acuity measured with theSnellen chart and Landolt broken rings (test distance 40 cm)after correction of any refractive errors Participants wereexcluded if they had significant visual impairment that couldnot be corrected We used the best performance across alltests for each participant as a measure of acuity in groupcomparisons Computerized tests of visuoperceptual functionincluded angle and motion discrimination tasks which haveestablished metrics in Lewy body disease and have beenreported in a number of our reports1011
As in previous work12 for assessment of visual hallucinationsthe hallucinations subscale of the Neuropsychiatric Inventory(NPIhall)13 was used with specific reference to visual halluci-nations occurring in the previous month Subsequently we usedthe derived NPIhall score (frequency times severity of hallucina-tions) in analyses For reliability purposes patients and carerswere independently asked about the occurrence of visual hal-lucinations in the month before MRI using screening questions
GlossaryBOLD = blood oxygen levelndashdependent CAMCOG = Cambridge Cognition Examination Cr = creatine DLB = dementiawith Lewy bodies DTI = diffusion tensor imaging FA = fractional anisotropy FWE = family-wise error GABA =γ-aminobutyric acid 5-HT = 5-hydroxytryptamine MAYO = Mayo Fluctuations Composite Score MD = mean diffusivityMMSE = Mini-Mental State Examination MRS = magnetic resonance spectroscopy NAA = N-acetylaspartate NPI =Neuropsychiatric Inventory NPIhall = Neuropsychiatric Inventory hallucinations subscale PD = Parkinson disease PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinatorROI = region of interestSPM = Statistical Parametric Mapping TBSS = tract-based spatial statistics TE = echo time TR = repetition time UPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VBM = voxel-based morphometry
e676 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
originating from theNorth-East Visual Hallucinations InterviewIII14 any discrepancies in the accounts of hallucinations be-tween patient and carerfamily member were discussed withboth parties and the assessor with reformulation of NPIhall testscores (with primacy given to the opinion of the caregiverwhere the patient seemed to lack insight)
We also used the noise pareidolia test15 as this has been shownto correlate with tendency to hallucinate in Lewy body de-mentia For this test participants were shown a series of 40black and white images These all contain cloud-like noiseformations and in 8 images there is a face inserted some-where (the location and face differ for each case) After 3example images participants were shown the 40 images one ata time and asked whether they could see any faces If so theywere asked to indicate the location of the face(s) We scoredthe test according to the number of illusory faces seen(ie faces indicated where there were not faces inserted in theimage) Cognitive fluctuations were quantified with the MayoFluctuations Composite Score (MAYO)16 and the ClinicianAssessment of Fluctuation scale17
Participants were classed as active visual hallucinators (PD-VH) if they had complex visual hallucinations in the monthpreceding their interview otherwise they were classed asnonhallucinators (controls and PD-nonVH) Since our fo-cus was on complex visual hallucinations those with minorhallucinations (eg passage or feeling of presence) but nocomplex hallucinations in the last month were included inthe PD-nonVH group We made this distinction sincepassage and feeling of presence probably have a differentetiologic basis to complex hallucinations even though mi-nor visual hallucinations typically precede complex visualhallucinations1819
Standard protocol approvals registrationsand patient consentsThe study was approved by the local ethics committee andwritten consent was obtained from all participants (or nom-inated independent mental capacity advocate where partici-pant lacked capacity)
MRI acquisitionParticipants were scanned on a 3T whole-body MR scanner(Achieva scanner Philips Medical Systems Best the Neth-erlands) with body coil transmission and an 8-channel headcoil receiver
Magnetic resonance spectroscopyWe used the MEGA-PRESS technique20 as previouslyreported21 with a sinc gaussian editing pulse applied alter-nately at 19 ppm (EDIT-ON) and 75 ppm (EDIT-OFF)Subtraction of the EDIT-OFF from EDIT-ON spectra allowsthe 3-ppm GABA+ signal to be separated from the overlyingcreatine peak MEGA-PRESS spectra were acquired froma voxel sized 45 times 32 times 20 mm centered on the midline of theoccipital lobe (data available from Newcastle University
e-prints [figure 1] eprintnclacuk247552) Sequenceparameters were as follows repetition time (TR) = 2000milliseconds (ms) echo time (TE) = 68 ms 320 averagesacquisition bandwidth = 1000 Hz VAPOR (variable powerradiofrequency pulses with optimized relaxation delays) watersuppression22 Macromolecular suppression editing23 was notperformed and thus our results are of GABA+ (ie GABAplus macromolecules) The magnetic resonance GABA signalis thought to reflect concentrations of metabolic GABA andlevels of ambient extracellular GABA that contribute to tonicGABAergic activity24
Structural and functional magnetic resonanceWe acquired images including a whole-brain structural3-dimensional MPRAGE (magnetization-prepared rapid-acquisition gradient echo) scan with sagittal acquisitionslice thickness 10 mm in-plane resolution 10 times 10 mm TR= 83 ms TE = 46 ms flip angle = 8deg and SENSE factor = 2fMRI data were collected with a gradient-echo echo planarimaging sequence (TR = 192 seconds TE = 40 ms field ofview 192 times 192 mm2 64 times 64 matrix size flip angle 90deg 27slices slice thickness 3mm slice gap 1mm)with 100 volumes(192 seconds) as participants looked at the checkerboardstimulus
Diffusion tensor imaging (DTI) acquisitions utilized a 2-di-mensional spin-echo echo planar imaging diffusion-weightedsequence with 59 slices TR = 6100ms TE = 70ms flip angle= 90deg field of view = 270 times 270mm pixel size = 21 times 21 mmslice thickness = 21 mm Diffusion weighting was applied in64 uniformly distributed directions (diffusion b = 1000smiddotmmminus2) and there were 6 acquisitions with no diffusionweighting (b = 0 smiddotmmminus2) We also collected an identicalimage with b = 0 smiddotmmminus2 but with the phase encoding di-rection reversed for distortion correction purposes
fMRI stimulus presentationfMRI was performed with the same checkerboard paradigm aswe have previously utilized25 fMRI-compatible goggles withlenses that ranged from minus40 to 40 diopters (05 increment)were used to correct any refractive errors that partic-ipants had
The stimulus presentation was controlled by the psycho-physics toolbox26 (psychtoolboxorg) extension for MAT-LAB (MathWorks Natick MA) A block design was usedwith a full field circular checkerboard stimulus consisting offive 192-second blocks of a black-and-white checkerboard(inverting at 75 Hz) alternating with five 192-second base-line blocks of a gray screen Participants were asked to focuson a central cross-hair
Magnetic resonance analysis
Spectroscopy data analysisGABA+ quantification was performed using the Gannettoolbox for MATLAB27 and consisted of the following steps(1) alignment of each pair (EDIT-ON and EDIT-OFF) of
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e677
spectra28 (2) subtraction of aligned spectra to produceGABA+ spectra followed by averaging across acquisitions(3) fitting a gaussian to the 3-ppm GABA+ peak to quantifyGABA+ based on the area under the curve For a typicaledited spectrum and Gaussian fit see data available fromNewcastle University e-prints (figure 1) eprintnclacuk247552
Choline creatine and NAA (N-acetylaspartate) amplitudeswere quantified from nonedited spectra only using theAMARES (Advanced Method for Accurate Robust and Ef-ficient Spectral fitting of MRS data with use of prior knowl-edge) algorithm from jMRUI (java-based magnetic resonanceuser interface)29 GABA+ and NAA were expressed as ratiosand normalized to creatine MRS fit quality was assessed by anexperienced physicist as described previously21
fMRI analysisImaging data were processed with Statistical ParametricMapping (SPM)12 (filionuclacukspm) similar to ourprevious work25 For each participant the T1 anatomicalimage was segmented and spatially normalized in SPM usingthe default parameters The fMRI data for each stimuluscondition were slice timing corrected motion corrected byaligning all functional images together and then coregisteredwith the T1 anatomical image The spatial normalizationparameters from the T1 segmentation were used to transformthe fMRI data to standard space with a voxel size of 3 times 3 times3 mm The normalized images were then smoothed with a 6 times6 times 6 mm full width at half maximum gaussian kernel A high-pass filter of 128 seconds was used and serial correlationswere removed with SPMrsquos AR(1) model
The general linear model in SPM was used to conducta whole-brain analysis of the fMRI data We created a designmatrix by convolving the time course of the checkerboardblock with the canonical hemodynamic response function andits first derivative The 6 parameters from the motion cor-rection were included in the design matrix as covariates of nointerest Individual participant and second-level (random-effects) group analyses were conducted Contrast images weregenerated from β estimates for the comparison of checker-board vs baseline Results are shown with a voxelwisethreshold of p lt 0001 (uncorrected) followed by clusterwisethreshold of p lt 005 family-wise error (FWE)-corrected formultiple comparisons
We also used a region of interest (ROI) analysis focusing onthe visual areas Five ROIs in MNI (Montreal NeurologicalInstitute) space were defined averaging across left and righthemispheres V1 V2 V3 V4 and V5 were taken from theSPM Anatomy toolbox (fz-juelichdeinminm-1DEFor-schung_docsSPMAnatomyToolboxSPMAnatomyTool-box_nodehtml) We also defined 3 ROIs from the overallactivation across all participantsmdashall voxels with activation inthe occipital lobe all voxels with associated deactivation (bothwith voxelwise threshold at p lt 005 FWE-corrected) and
a bilateral LGN (lateral geniculate nucleus) region (voxelwisep lt 0001 uncorrected)
Structural MRI analysis
Gray matterFor analysis of gray matter atrophy the T1-weighted struc-tural images were segmented with the SPM12 segment tooland then processed using the DARTEL (Diffeomorphic An-atomical Registration Through Exponentiated Lie Algebra)Toolbox to create a group-specific template to which theindividual images were spatially normalized Images weremodulated to preserve the total tissue amount during nor-malization and smoothed with an 8-mm gaussian filter Weused the SPM Anatomy toolbox30 to identify location ofsignificant clusters For each participant we also extracted thefraction of gray matter white matter and CSF within theindividual spectroscopy voxel location
Diffusion white matterDTI data were processed using FSL (fslfmriboxacukfslfslwiki) using the topup program to correct susceptibility-induced distortions using the 2 b = 0 smiddotmmminus2 images withopposite phase encoding The eddy package was then used tocorrect images for eddy current distortion movement andmotion-induced signal dropout Fractional anisotropy (FA)andmean diffusivity (MD) were then calculated with the dtifitsoftware and the TBSS (tract-based spatial statistics) pack-age31 used to align the FA images all together create a whitematter skeleton of major tracts and extract FA andMD valuesfor each participant on the white matter skeleton The imageswere visually inspected at each stage
Statistical analysisROI data and clinical variables were analyzed with the Sta-tistical Package for Social Sciences (SPSS version 19 IBMCorp Armonk NY) Independent t tests or analysis of vari-ance was used to compare groups for continuous variablesSpearman rank correlation coefficient was used to comparecontinuous variables
For the fMRI voxelwise data a 3-group analysis of variancewas performed using SPM to determine overall activationpatterns and investigate group differences Voxel-based mor-phometry (VBM) was done using SPM on the smoothedmodulated gray matter images using a 3-group analysis ofcovariance with age and intracranial volume (sum of CSFgray matter and white matter) as covariates and also with theaddition of CAMCOG as a measure of cognitive function Weinvestigated with SPM the relationship of GABA+creatine(Cr) with voxelwise gray matter volume with covariates ofage intracranial volume and group (PD-nonVH PD-VH andcontrols)
For the diffusion analysis voxelwise differences inMD and FAbetween groups (with age as a covariate) were estimated usingthe FSL randomise package We also used this to look atvoxelwise correlations with occipital GABA+Cr controlling
e678 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
for age and group (PD-nonVH PD-VH and control) Thiswas repeated with the addition of CAMCOG as a measure ofcognitive function
Data availabilityAnonymized data on which this article is based will be sharedon request with any appropriately qualified investigator
ResultsMRI scans were obtained on 20 of the controls and 36 par-ticipants with PD of whom 15 had mild cognitive impairmentand 21 had PD dementia Table 1 shows the demographics forthese participants There were no significant differences in ageor sex between hallucination groups and there were no dif-ferences in duration of PD or levodopa dose between the PD-VH and PD-nonVH groups However the PD-VH group hadworse motor function according to the UPDRS-III score (p lt0001) worse cognition on the CAMCOG scale (p lt 0001)
and were more likely to be taking cholinesterase inhibitors (p= 0025) As expected the PD-VH group had significantlyhigher hallucination scores and were more likely to havemisperceptions on the pareidolia test compared to the PD-nonVH group Data available from Newcastle Universitye-prints (table 1 eprintnclacuk247552) compare thepatients with PD with vs without dementia There were nosignificant differences in age sex years of education durationof PD or levodopa dose The participants with dementia hada higher UPDRS-III score poorer vision and a greater ten-dency to hallucinate as indicated by the NeuropsychiatricInventory (NPI) and pareidolia test
We excluded 4 participants (1 PD-VH 2 PD-nonVH 1control) from the spectroscopy analysis because they didnot meet MRS quality-assurance criteria21 There were nosignificant differences in age sex or diagnosis of dementiabetween the excluded and nonexcluded participantsTable 2 shows the ratio of GABA+ and NAA to Cr for the
Table 1 Demographics
Control (n = 20) PD-nonVH (n = 19) PD-VH (n = 17)
Age y 754 (50) 723 (51) 755 (45) F253 = 26 p = 0083
Female n () 6 (30) 2 (11) 4 (24) χ2 = 18 p = 041
Dementia n () 0 7 (37) 15 (83) FET p = 0007
Education y 135 (21) 111 (15) 116 (22) F253 = 87 p = 0001ab
Duration of PD y mdash 96 (65) 110 (74) F134 = 04 p = 054
Levodopa dose in 24 h mdash 6735 (4282) 7173 (4217) F134 = 01 p = 076
ChEI n () mdash 2 (11) 8 (47) FET p = 0025
Antipsychotics n () mdash 0 (0) 3 (18) FET p = 0095
UPDRS-III total score 22 (25) 347 (188) 559 (193) F253 = 581 p lt 0001abc
CAMCOG total score 952 (70) 845 (114) 746 (153) F253 = 147 p lt 0001abc
NPI total (A times B) hallucinations mdash 01 (02) 35 (24) F134 = 396 p lt 0001
NPI total score mdash 75 (66) 225 (178) F131 = 105 p = 0003
CAF total mdash 21 (27) 58 (38) F133 = 118 p = 0002
MAYO total mdash 14 (11) 24 (15) F133 = 60 p = 0020
Abbreviations CAF = Clinician Assessment of Fluctuation scale CAMCOG = Cambridge Cognition Examination ChEI = cholinesterase inhibitor FET = Fisherexact test MAYO = Mayo Fluctuations Composite Score MMSE = Mini-Mental State Examination NPI = Neuropsychiatric Inventory PD = Parkinson diseaseUPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VH = visual hallucinationThe pareidolia task result is the number of pareidolias seenSignificant (p lt 005) Tukey post hoc testsa Control vs PD-nonVHbControl vs PD-VHcPD-VH vs PD-nonVH
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e679
groups There was a significant group difference in theGABA+Cr ratio (figure 1) with post hoc Tukey testfinding GABA+Cr reduced in PD-VH relative to PD-nonVH The group difference remained significant afterincluding CAMCOG as a measure of cognitive ability in thelinear model (F248 = 327 p = 0047) There were no sig-nificant differences in gray or white matter proportionwithin the voxel between groups (table 2) Within the PDgroup GABA+Cr correlated with visual acuity (Spearmanρ = 04 p = 0025) MMSE (ρ = 035 p = 0047) UPDRS-III(ρ = minus0345 p = 0049) and MAYO total score (ρ = minus0627p lt 0001) There were no significant correlations (p gt 01)with disease duration CAMCOG angle or motion test orthe pareidolia test There were no significant correlationsbetween GABA+Cr and NPI hallucination score aftercontrolling for VH group
There was no significant difference in GABA+Cr levels be-tween participants with PD taking cholinesterase inhibitors vsthose not taking them (0098 SD 0012 vs 0093 SD 0008 t31= 116 p = 026) between those taking antipsychotic agents(quetiapine) vs those not (0088 SD 0007 vs 0097 SD 0011t31 = 142 p = 017) and there was no significant correlationbetween levodopa dose and GABA+Cr (ρ = 0018 p = 09)
The fMRI scans were not acquired on one PD participant whodid not tolerate the full scanning session and 4 PD participantsrsquoscans were excluded because of excessive motion leaving 17PD-VH 14 PD-nonVH and 20 controls with usable fMRI dataAll groups showed a typical activation pattern to the checker-board (data available fromNewcastle University e-prints [figure2] eprintnclacuk247552) but there were no significant dif-ferences in activation between any groups In the ROI analysis
e680 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
originating from theNorth-East Visual Hallucinations InterviewIII14 any discrepancies in the accounts of hallucinations be-tween patient and carerfamily member were discussed withboth parties and the assessor with reformulation of NPIhall testscores (with primacy given to the opinion of the caregiverwhere the patient seemed to lack insight)
We also used the noise pareidolia test15 as this has been shownto correlate with tendency to hallucinate in Lewy body de-mentia For this test participants were shown a series of 40black and white images These all contain cloud-like noiseformations and in 8 images there is a face inserted some-where (the location and face differ for each case) After 3example images participants were shown the 40 images one ata time and asked whether they could see any faces If so theywere asked to indicate the location of the face(s) We scoredthe test according to the number of illusory faces seen(ie faces indicated where there were not faces inserted in theimage) Cognitive fluctuations were quantified with the MayoFluctuations Composite Score (MAYO)16 and the ClinicianAssessment of Fluctuation scale17
Participants were classed as active visual hallucinators (PD-VH) if they had complex visual hallucinations in the monthpreceding their interview otherwise they were classed asnonhallucinators (controls and PD-nonVH) Since our fo-cus was on complex visual hallucinations those with minorhallucinations (eg passage or feeling of presence) but nocomplex hallucinations in the last month were included inthe PD-nonVH group We made this distinction sincepassage and feeling of presence probably have a differentetiologic basis to complex hallucinations even though mi-nor visual hallucinations typically precede complex visualhallucinations1819
Standard protocol approvals registrationsand patient consentsThe study was approved by the local ethics committee andwritten consent was obtained from all participants (or nom-inated independent mental capacity advocate where partici-pant lacked capacity)
MRI acquisitionParticipants were scanned on a 3T whole-body MR scanner(Achieva scanner Philips Medical Systems Best the Neth-erlands) with body coil transmission and an 8-channel headcoil receiver
Magnetic resonance spectroscopyWe used the MEGA-PRESS technique20 as previouslyreported21 with a sinc gaussian editing pulse applied alter-nately at 19 ppm (EDIT-ON) and 75 ppm (EDIT-OFF)Subtraction of the EDIT-OFF from EDIT-ON spectra allowsthe 3-ppm GABA+ signal to be separated from the overlyingcreatine peak MEGA-PRESS spectra were acquired froma voxel sized 45 times 32 times 20 mm centered on the midline of theoccipital lobe (data available from Newcastle University
e-prints [figure 1] eprintnclacuk247552) Sequenceparameters were as follows repetition time (TR) = 2000milliseconds (ms) echo time (TE) = 68 ms 320 averagesacquisition bandwidth = 1000 Hz VAPOR (variable powerradiofrequency pulses with optimized relaxation delays) watersuppression22 Macromolecular suppression editing23 was notperformed and thus our results are of GABA+ (ie GABAplus macromolecules) The magnetic resonance GABA signalis thought to reflect concentrations of metabolic GABA andlevels of ambient extracellular GABA that contribute to tonicGABAergic activity24
Structural and functional magnetic resonanceWe acquired images including a whole-brain structural3-dimensional MPRAGE (magnetization-prepared rapid-acquisition gradient echo) scan with sagittal acquisitionslice thickness 10 mm in-plane resolution 10 times 10 mm TR= 83 ms TE = 46 ms flip angle = 8deg and SENSE factor = 2fMRI data were collected with a gradient-echo echo planarimaging sequence (TR = 192 seconds TE = 40 ms field ofview 192 times 192 mm2 64 times 64 matrix size flip angle 90deg 27slices slice thickness 3mm slice gap 1mm)with 100 volumes(192 seconds) as participants looked at the checkerboardstimulus
Diffusion tensor imaging (DTI) acquisitions utilized a 2-di-mensional spin-echo echo planar imaging diffusion-weightedsequence with 59 slices TR = 6100ms TE = 70ms flip angle= 90deg field of view = 270 times 270mm pixel size = 21 times 21 mmslice thickness = 21 mm Diffusion weighting was applied in64 uniformly distributed directions (diffusion b = 1000smiddotmmminus2) and there were 6 acquisitions with no diffusionweighting (b = 0 smiddotmmminus2) We also collected an identicalimage with b = 0 smiddotmmminus2 but with the phase encoding di-rection reversed for distortion correction purposes
fMRI stimulus presentationfMRI was performed with the same checkerboard paradigm aswe have previously utilized25 fMRI-compatible goggles withlenses that ranged from minus40 to 40 diopters (05 increment)were used to correct any refractive errors that partic-ipants had
The stimulus presentation was controlled by the psycho-physics toolbox26 (psychtoolboxorg) extension for MAT-LAB (MathWorks Natick MA) A block design was usedwith a full field circular checkerboard stimulus consisting offive 192-second blocks of a black-and-white checkerboard(inverting at 75 Hz) alternating with five 192-second base-line blocks of a gray screen Participants were asked to focuson a central cross-hair
Magnetic resonance analysis
Spectroscopy data analysisGABA+ quantification was performed using the Gannettoolbox for MATLAB27 and consisted of the following steps(1) alignment of each pair (EDIT-ON and EDIT-OFF) of
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e677
spectra28 (2) subtraction of aligned spectra to produceGABA+ spectra followed by averaging across acquisitions(3) fitting a gaussian to the 3-ppm GABA+ peak to quantifyGABA+ based on the area under the curve For a typicaledited spectrum and Gaussian fit see data available fromNewcastle University e-prints (figure 1) eprintnclacuk247552
Choline creatine and NAA (N-acetylaspartate) amplitudeswere quantified from nonedited spectra only using theAMARES (Advanced Method for Accurate Robust and Ef-ficient Spectral fitting of MRS data with use of prior knowl-edge) algorithm from jMRUI (java-based magnetic resonanceuser interface)29 GABA+ and NAA were expressed as ratiosand normalized to creatine MRS fit quality was assessed by anexperienced physicist as described previously21
fMRI analysisImaging data were processed with Statistical ParametricMapping (SPM)12 (filionuclacukspm) similar to ourprevious work25 For each participant the T1 anatomicalimage was segmented and spatially normalized in SPM usingthe default parameters The fMRI data for each stimuluscondition were slice timing corrected motion corrected byaligning all functional images together and then coregisteredwith the T1 anatomical image The spatial normalizationparameters from the T1 segmentation were used to transformthe fMRI data to standard space with a voxel size of 3 times 3 times3 mm The normalized images were then smoothed with a 6 times6 times 6 mm full width at half maximum gaussian kernel A high-pass filter of 128 seconds was used and serial correlationswere removed with SPMrsquos AR(1) model
The general linear model in SPM was used to conducta whole-brain analysis of the fMRI data We created a designmatrix by convolving the time course of the checkerboardblock with the canonical hemodynamic response function andits first derivative The 6 parameters from the motion cor-rection were included in the design matrix as covariates of nointerest Individual participant and second-level (random-effects) group analyses were conducted Contrast images weregenerated from β estimates for the comparison of checker-board vs baseline Results are shown with a voxelwisethreshold of p lt 0001 (uncorrected) followed by clusterwisethreshold of p lt 005 family-wise error (FWE)-corrected formultiple comparisons
We also used a region of interest (ROI) analysis focusing onthe visual areas Five ROIs in MNI (Montreal NeurologicalInstitute) space were defined averaging across left and righthemispheres V1 V2 V3 V4 and V5 were taken from theSPM Anatomy toolbox (fz-juelichdeinminm-1DEFor-schung_docsSPMAnatomyToolboxSPMAnatomyTool-box_nodehtml) We also defined 3 ROIs from the overallactivation across all participantsmdashall voxels with activation inthe occipital lobe all voxels with associated deactivation (bothwith voxelwise threshold at p lt 005 FWE-corrected) and
a bilateral LGN (lateral geniculate nucleus) region (voxelwisep lt 0001 uncorrected)
Structural MRI analysis
Gray matterFor analysis of gray matter atrophy the T1-weighted struc-tural images were segmented with the SPM12 segment tooland then processed using the DARTEL (Diffeomorphic An-atomical Registration Through Exponentiated Lie Algebra)Toolbox to create a group-specific template to which theindividual images were spatially normalized Images weremodulated to preserve the total tissue amount during nor-malization and smoothed with an 8-mm gaussian filter Weused the SPM Anatomy toolbox30 to identify location ofsignificant clusters For each participant we also extracted thefraction of gray matter white matter and CSF within theindividual spectroscopy voxel location
Diffusion white matterDTI data were processed using FSL (fslfmriboxacukfslfslwiki) using the topup program to correct susceptibility-induced distortions using the 2 b = 0 smiddotmmminus2 images withopposite phase encoding The eddy package was then used tocorrect images for eddy current distortion movement andmotion-induced signal dropout Fractional anisotropy (FA)andmean diffusivity (MD) were then calculated with the dtifitsoftware and the TBSS (tract-based spatial statistics) pack-age31 used to align the FA images all together create a whitematter skeleton of major tracts and extract FA andMD valuesfor each participant on the white matter skeleton The imageswere visually inspected at each stage
Statistical analysisROI data and clinical variables were analyzed with the Sta-tistical Package for Social Sciences (SPSS version 19 IBMCorp Armonk NY) Independent t tests or analysis of vari-ance was used to compare groups for continuous variablesSpearman rank correlation coefficient was used to comparecontinuous variables
For the fMRI voxelwise data a 3-group analysis of variancewas performed using SPM to determine overall activationpatterns and investigate group differences Voxel-based mor-phometry (VBM) was done using SPM on the smoothedmodulated gray matter images using a 3-group analysis ofcovariance with age and intracranial volume (sum of CSFgray matter and white matter) as covariates and also with theaddition of CAMCOG as a measure of cognitive function Weinvestigated with SPM the relationship of GABA+creatine(Cr) with voxelwise gray matter volume with covariates ofage intracranial volume and group (PD-nonVH PD-VH andcontrols)
For the diffusion analysis voxelwise differences inMD and FAbetween groups (with age as a covariate) were estimated usingthe FSL randomise package We also used this to look atvoxelwise correlations with occipital GABA+Cr controlling
e678 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
for age and group (PD-nonVH PD-VH and control) Thiswas repeated with the addition of CAMCOG as a measure ofcognitive function
Data availabilityAnonymized data on which this article is based will be sharedon request with any appropriately qualified investigator
ResultsMRI scans were obtained on 20 of the controls and 36 par-ticipants with PD of whom 15 had mild cognitive impairmentand 21 had PD dementia Table 1 shows the demographics forthese participants There were no significant differences in ageor sex between hallucination groups and there were no dif-ferences in duration of PD or levodopa dose between the PD-VH and PD-nonVH groups However the PD-VH group hadworse motor function according to the UPDRS-III score (p lt0001) worse cognition on the CAMCOG scale (p lt 0001)
and were more likely to be taking cholinesterase inhibitors (p= 0025) As expected the PD-VH group had significantlyhigher hallucination scores and were more likely to havemisperceptions on the pareidolia test compared to the PD-nonVH group Data available from Newcastle Universitye-prints (table 1 eprintnclacuk247552) compare thepatients with PD with vs without dementia There were nosignificant differences in age sex years of education durationof PD or levodopa dose The participants with dementia hada higher UPDRS-III score poorer vision and a greater ten-dency to hallucinate as indicated by the NeuropsychiatricInventory (NPI) and pareidolia test
We excluded 4 participants (1 PD-VH 2 PD-nonVH 1control) from the spectroscopy analysis because they didnot meet MRS quality-assurance criteria21 There were nosignificant differences in age sex or diagnosis of dementiabetween the excluded and nonexcluded participantsTable 2 shows the ratio of GABA+ and NAA to Cr for the
Table 1 Demographics
Control (n = 20) PD-nonVH (n = 19) PD-VH (n = 17)
Age y 754 (50) 723 (51) 755 (45) F253 = 26 p = 0083
Female n () 6 (30) 2 (11) 4 (24) χ2 = 18 p = 041
Dementia n () 0 7 (37) 15 (83) FET p = 0007
Education y 135 (21) 111 (15) 116 (22) F253 = 87 p = 0001ab
Duration of PD y mdash 96 (65) 110 (74) F134 = 04 p = 054
Levodopa dose in 24 h mdash 6735 (4282) 7173 (4217) F134 = 01 p = 076
ChEI n () mdash 2 (11) 8 (47) FET p = 0025
Antipsychotics n () mdash 0 (0) 3 (18) FET p = 0095
UPDRS-III total score 22 (25) 347 (188) 559 (193) F253 = 581 p lt 0001abc
CAMCOG total score 952 (70) 845 (114) 746 (153) F253 = 147 p lt 0001abc
NPI total (A times B) hallucinations mdash 01 (02) 35 (24) F134 = 396 p lt 0001
NPI total score mdash 75 (66) 225 (178) F131 = 105 p = 0003
CAF total mdash 21 (27) 58 (38) F133 = 118 p = 0002
MAYO total mdash 14 (11) 24 (15) F133 = 60 p = 0020
Abbreviations CAF = Clinician Assessment of Fluctuation scale CAMCOG = Cambridge Cognition Examination ChEI = cholinesterase inhibitor FET = Fisherexact test MAYO = Mayo Fluctuations Composite Score MMSE = Mini-Mental State Examination NPI = Neuropsychiatric Inventory PD = Parkinson diseaseUPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VH = visual hallucinationThe pareidolia task result is the number of pareidolias seenSignificant (p lt 005) Tukey post hoc testsa Control vs PD-nonVHbControl vs PD-VHcPD-VH vs PD-nonVH
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e679
groups There was a significant group difference in theGABA+Cr ratio (figure 1) with post hoc Tukey testfinding GABA+Cr reduced in PD-VH relative to PD-nonVH The group difference remained significant afterincluding CAMCOG as a measure of cognitive ability in thelinear model (F248 = 327 p = 0047) There were no sig-nificant differences in gray or white matter proportionwithin the voxel between groups (table 2) Within the PDgroup GABA+Cr correlated with visual acuity (Spearmanρ = 04 p = 0025) MMSE (ρ = 035 p = 0047) UPDRS-III(ρ = minus0345 p = 0049) and MAYO total score (ρ = minus0627p lt 0001) There were no significant correlations (p gt 01)with disease duration CAMCOG angle or motion test orthe pareidolia test There were no significant correlationsbetween GABA+Cr and NPI hallucination score aftercontrolling for VH group
There was no significant difference in GABA+Cr levels be-tween participants with PD taking cholinesterase inhibitors vsthose not taking them (0098 SD 0012 vs 0093 SD 0008 t31= 116 p = 026) between those taking antipsychotic agents(quetiapine) vs those not (0088 SD 0007 vs 0097 SD 0011t31 = 142 p = 017) and there was no significant correlationbetween levodopa dose and GABA+Cr (ρ = 0018 p = 09)
The fMRI scans were not acquired on one PD participant whodid not tolerate the full scanning session and 4 PD participantsrsquoscans were excluded because of excessive motion leaving 17PD-VH 14 PD-nonVH and 20 controls with usable fMRI dataAll groups showed a typical activation pattern to the checker-board (data available fromNewcastle University e-prints [figure2] eprintnclacuk247552) but there were no significant dif-ferences in activation between any groups In the ROI analysis
e680 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
spectra28 (2) subtraction of aligned spectra to produceGABA+ spectra followed by averaging across acquisitions(3) fitting a gaussian to the 3-ppm GABA+ peak to quantifyGABA+ based on the area under the curve For a typicaledited spectrum and Gaussian fit see data available fromNewcastle University e-prints (figure 1) eprintnclacuk247552
Choline creatine and NAA (N-acetylaspartate) amplitudeswere quantified from nonedited spectra only using theAMARES (Advanced Method for Accurate Robust and Ef-ficient Spectral fitting of MRS data with use of prior knowl-edge) algorithm from jMRUI (java-based magnetic resonanceuser interface)29 GABA+ and NAA were expressed as ratiosand normalized to creatine MRS fit quality was assessed by anexperienced physicist as described previously21
fMRI analysisImaging data were processed with Statistical ParametricMapping (SPM)12 (filionuclacukspm) similar to ourprevious work25 For each participant the T1 anatomicalimage was segmented and spatially normalized in SPM usingthe default parameters The fMRI data for each stimuluscondition were slice timing corrected motion corrected byaligning all functional images together and then coregisteredwith the T1 anatomical image The spatial normalizationparameters from the T1 segmentation were used to transformthe fMRI data to standard space with a voxel size of 3 times 3 times3 mm The normalized images were then smoothed with a 6 times6 times 6 mm full width at half maximum gaussian kernel A high-pass filter of 128 seconds was used and serial correlationswere removed with SPMrsquos AR(1) model
The general linear model in SPM was used to conducta whole-brain analysis of the fMRI data We created a designmatrix by convolving the time course of the checkerboardblock with the canonical hemodynamic response function andits first derivative The 6 parameters from the motion cor-rection were included in the design matrix as covariates of nointerest Individual participant and second-level (random-effects) group analyses were conducted Contrast images weregenerated from β estimates for the comparison of checker-board vs baseline Results are shown with a voxelwisethreshold of p lt 0001 (uncorrected) followed by clusterwisethreshold of p lt 005 family-wise error (FWE)-corrected formultiple comparisons
We also used a region of interest (ROI) analysis focusing onthe visual areas Five ROIs in MNI (Montreal NeurologicalInstitute) space were defined averaging across left and righthemispheres V1 V2 V3 V4 and V5 were taken from theSPM Anatomy toolbox (fz-juelichdeinminm-1DEFor-schung_docsSPMAnatomyToolboxSPMAnatomyTool-box_nodehtml) We also defined 3 ROIs from the overallactivation across all participantsmdashall voxels with activation inthe occipital lobe all voxels with associated deactivation (bothwith voxelwise threshold at p lt 005 FWE-corrected) and
a bilateral LGN (lateral geniculate nucleus) region (voxelwisep lt 0001 uncorrected)
Structural MRI analysis
Gray matterFor analysis of gray matter atrophy the T1-weighted struc-tural images were segmented with the SPM12 segment tooland then processed using the DARTEL (Diffeomorphic An-atomical Registration Through Exponentiated Lie Algebra)Toolbox to create a group-specific template to which theindividual images were spatially normalized Images weremodulated to preserve the total tissue amount during nor-malization and smoothed with an 8-mm gaussian filter Weused the SPM Anatomy toolbox30 to identify location ofsignificant clusters For each participant we also extracted thefraction of gray matter white matter and CSF within theindividual spectroscopy voxel location
Diffusion white matterDTI data were processed using FSL (fslfmriboxacukfslfslwiki) using the topup program to correct susceptibility-induced distortions using the 2 b = 0 smiddotmmminus2 images withopposite phase encoding The eddy package was then used tocorrect images for eddy current distortion movement andmotion-induced signal dropout Fractional anisotropy (FA)andmean diffusivity (MD) were then calculated with the dtifitsoftware and the TBSS (tract-based spatial statistics) pack-age31 used to align the FA images all together create a whitematter skeleton of major tracts and extract FA andMD valuesfor each participant on the white matter skeleton The imageswere visually inspected at each stage
Statistical analysisROI data and clinical variables were analyzed with the Sta-tistical Package for Social Sciences (SPSS version 19 IBMCorp Armonk NY) Independent t tests or analysis of vari-ance was used to compare groups for continuous variablesSpearman rank correlation coefficient was used to comparecontinuous variables
For the fMRI voxelwise data a 3-group analysis of variancewas performed using SPM to determine overall activationpatterns and investigate group differences Voxel-based mor-phometry (VBM) was done using SPM on the smoothedmodulated gray matter images using a 3-group analysis ofcovariance with age and intracranial volume (sum of CSFgray matter and white matter) as covariates and also with theaddition of CAMCOG as a measure of cognitive function Weinvestigated with SPM the relationship of GABA+creatine(Cr) with voxelwise gray matter volume with covariates ofage intracranial volume and group (PD-nonVH PD-VH andcontrols)
For the diffusion analysis voxelwise differences inMD and FAbetween groups (with age as a covariate) were estimated usingthe FSL randomise package We also used this to look atvoxelwise correlations with occipital GABA+Cr controlling
e678 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
for age and group (PD-nonVH PD-VH and control) Thiswas repeated with the addition of CAMCOG as a measure ofcognitive function
Data availabilityAnonymized data on which this article is based will be sharedon request with any appropriately qualified investigator
ResultsMRI scans were obtained on 20 of the controls and 36 par-ticipants with PD of whom 15 had mild cognitive impairmentand 21 had PD dementia Table 1 shows the demographics forthese participants There were no significant differences in ageor sex between hallucination groups and there were no dif-ferences in duration of PD or levodopa dose between the PD-VH and PD-nonVH groups However the PD-VH group hadworse motor function according to the UPDRS-III score (p lt0001) worse cognition on the CAMCOG scale (p lt 0001)
and were more likely to be taking cholinesterase inhibitors (p= 0025) As expected the PD-VH group had significantlyhigher hallucination scores and were more likely to havemisperceptions on the pareidolia test compared to the PD-nonVH group Data available from Newcastle Universitye-prints (table 1 eprintnclacuk247552) compare thepatients with PD with vs without dementia There were nosignificant differences in age sex years of education durationof PD or levodopa dose The participants with dementia hada higher UPDRS-III score poorer vision and a greater ten-dency to hallucinate as indicated by the NeuropsychiatricInventory (NPI) and pareidolia test
We excluded 4 participants (1 PD-VH 2 PD-nonVH 1control) from the spectroscopy analysis because they didnot meet MRS quality-assurance criteria21 There were nosignificant differences in age sex or diagnosis of dementiabetween the excluded and nonexcluded participantsTable 2 shows the ratio of GABA+ and NAA to Cr for the
Table 1 Demographics
Control (n = 20) PD-nonVH (n = 19) PD-VH (n = 17)
Age y 754 (50) 723 (51) 755 (45) F253 = 26 p = 0083
Female n () 6 (30) 2 (11) 4 (24) χ2 = 18 p = 041
Dementia n () 0 7 (37) 15 (83) FET p = 0007
Education y 135 (21) 111 (15) 116 (22) F253 = 87 p = 0001ab
Duration of PD y mdash 96 (65) 110 (74) F134 = 04 p = 054
Levodopa dose in 24 h mdash 6735 (4282) 7173 (4217) F134 = 01 p = 076
ChEI n () mdash 2 (11) 8 (47) FET p = 0025
Antipsychotics n () mdash 0 (0) 3 (18) FET p = 0095
UPDRS-III total score 22 (25) 347 (188) 559 (193) F253 = 581 p lt 0001abc
CAMCOG total score 952 (70) 845 (114) 746 (153) F253 = 147 p lt 0001abc
NPI total (A times B) hallucinations mdash 01 (02) 35 (24) F134 = 396 p lt 0001
NPI total score mdash 75 (66) 225 (178) F131 = 105 p = 0003
CAF total mdash 21 (27) 58 (38) F133 = 118 p = 0002
MAYO total mdash 14 (11) 24 (15) F133 = 60 p = 0020
Abbreviations CAF = Clinician Assessment of Fluctuation scale CAMCOG = Cambridge Cognition Examination ChEI = cholinesterase inhibitor FET = Fisherexact test MAYO = Mayo Fluctuations Composite Score MMSE = Mini-Mental State Examination NPI = Neuropsychiatric Inventory PD = Parkinson diseaseUPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VH = visual hallucinationThe pareidolia task result is the number of pareidolias seenSignificant (p lt 005) Tukey post hoc testsa Control vs PD-nonVHbControl vs PD-VHcPD-VH vs PD-nonVH
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e679
groups There was a significant group difference in theGABA+Cr ratio (figure 1) with post hoc Tukey testfinding GABA+Cr reduced in PD-VH relative to PD-nonVH The group difference remained significant afterincluding CAMCOG as a measure of cognitive ability in thelinear model (F248 = 327 p = 0047) There were no sig-nificant differences in gray or white matter proportionwithin the voxel between groups (table 2) Within the PDgroup GABA+Cr correlated with visual acuity (Spearmanρ = 04 p = 0025) MMSE (ρ = 035 p = 0047) UPDRS-III(ρ = minus0345 p = 0049) and MAYO total score (ρ = minus0627p lt 0001) There were no significant correlations (p gt 01)with disease duration CAMCOG angle or motion test orthe pareidolia test There were no significant correlationsbetween GABA+Cr and NPI hallucination score aftercontrolling for VH group
There was no significant difference in GABA+Cr levels be-tween participants with PD taking cholinesterase inhibitors vsthose not taking them (0098 SD 0012 vs 0093 SD 0008 t31= 116 p = 026) between those taking antipsychotic agents(quetiapine) vs those not (0088 SD 0007 vs 0097 SD 0011t31 = 142 p = 017) and there was no significant correlationbetween levodopa dose and GABA+Cr (ρ = 0018 p = 09)
The fMRI scans were not acquired on one PD participant whodid not tolerate the full scanning session and 4 PD participantsrsquoscans were excluded because of excessive motion leaving 17PD-VH 14 PD-nonVH and 20 controls with usable fMRI dataAll groups showed a typical activation pattern to the checker-board (data available fromNewcastle University e-prints [figure2] eprintnclacuk247552) but there were no significant dif-ferences in activation between any groups In the ROI analysis
e680 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
for age and group (PD-nonVH PD-VH and control) Thiswas repeated with the addition of CAMCOG as a measure ofcognitive function
Data availabilityAnonymized data on which this article is based will be sharedon request with any appropriately qualified investigator
ResultsMRI scans were obtained on 20 of the controls and 36 par-ticipants with PD of whom 15 had mild cognitive impairmentand 21 had PD dementia Table 1 shows the demographics forthese participants There were no significant differences in ageor sex between hallucination groups and there were no dif-ferences in duration of PD or levodopa dose between the PD-VH and PD-nonVH groups However the PD-VH group hadworse motor function according to the UPDRS-III score (p lt0001) worse cognition on the CAMCOG scale (p lt 0001)
and were more likely to be taking cholinesterase inhibitors (p= 0025) As expected the PD-VH group had significantlyhigher hallucination scores and were more likely to havemisperceptions on the pareidolia test compared to the PD-nonVH group Data available from Newcastle Universitye-prints (table 1 eprintnclacuk247552) compare thepatients with PD with vs without dementia There were nosignificant differences in age sex years of education durationof PD or levodopa dose The participants with dementia hada higher UPDRS-III score poorer vision and a greater ten-dency to hallucinate as indicated by the NeuropsychiatricInventory (NPI) and pareidolia test
We excluded 4 participants (1 PD-VH 2 PD-nonVH 1control) from the spectroscopy analysis because they didnot meet MRS quality-assurance criteria21 There were nosignificant differences in age sex or diagnosis of dementiabetween the excluded and nonexcluded participantsTable 2 shows the ratio of GABA+ and NAA to Cr for the
Table 1 Demographics
Control (n = 20) PD-nonVH (n = 19) PD-VH (n = 17)
Age y 754 (50) 723 (51) 755 (45) F253 = 26 p = 0083
Female n () 6 (30) 2 (11) 4 (24) χ2 = 18 p = 041
Dementia n () 0 7 (37) 15 (83) FET p = 0007
Education y 135 (21) 111 (15) 116 (22) F253 = 87 p = 0001ab
Duration of PD y mdash 96 (65) 110 (74) F134 = 04 p = 054
Levodopa dose in 24 h mdash 6735 (4282) 7173 (4217) F134 = 01 p = 076
ChEI n () mdash 2 (11) 8 (47) FET p = 0025
Antipsychotics n () mdash 0 (0) 3 (18) FET p = 0095
UPDRS-III total score 22 (25) 347 (188) 559 (193) F253 = 581 p lt 0001abc
CAMCOG total score 952 (70) 845 (114) 746 (153) F253 = 147 p lt 0001abc
NPI total (A times B) hallucinations mdash 01 (02) 35 (24) F134 = 396 p lt 0001
NPI total score mdash 75 (66) 225 (178) F131 = 105 p = 0003
CAF total mdash 21 (27) 58 (38) F133 = 118 p = 0002
MAYO total mdash 14 (11) 24 (15) F133 = 60 p = 0020
Abbreviations CAF = Clinician Assessment of Fluctuation scale CAMCOG = Cambridge Cognition Examination ChEI = cholinesterase inhibitor FET = Fisherexact test MAYO = Mayo Fluctuations Composite Score MMSE = Mini-Mental State Examination NPI = Neuropsychiatric Inventory PD = Parkinson diseaseUPDRS-III = Unified Parkinsonrsquos Disease Rating Scale motor subsection VH = visual hallucinationThe pareidolia task result is the number of pareidolias seenSignificant (p lt 005) Tukey post hoc testsa Control vs PD-nonVHbControl vs PD-VHcPD-VH vs PD-nonVH
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e679
groups There was a significant group difference in theGABA+Cr ratio (figure 1) with post hoc Tukey testfinding GABA+Cr reduced in PD-VH relative to PD-nonVH The group difference remained significant afterincluding CAMCOG as a measure of cognitive ability in thelinear model (F248 = 327 p = 0047) There were no sig-nificant differences in gray or white matter proportionwithin the voxel between groups (table 2) Within the PDgroup GABA+Cr correlated with visual acuity (Spearmanρ = 04 p = 0025) MMSE (ρ = 035 p = 0047) UPDRS-III(ρ = minus0345 p = 0049) and MAYO total score (ρ = minus0627p lt 0001) There were no significant correlations (p gt 01)with disease duration CAMCOG angle or motion test orthe pareidolia test There were no significant correlationsbetween GABA+Cr and NPI hallucination score aftercontrolling for VH group
There was no significant difference in GABA+Cr levels be-tween participants with PD taking cholinesterase inhibitors vsthose not taking them (0098 SD 0012 vs 0093 SD 0008 t31= 116 p = 026) between those taking antipsychotic agents(quetiapine) vs those not (0088 SD 0007 vs 0097 SD 0011t31 = 142 p = 017) and there was no significant correlationbetween levodopa dose and GABA+Cr (ρ = 0018 p = 09)
The fMRI scans were not acquired on one PD participant whodid not tolerate the full scanning session and 4 PD participantsrsquoscans were excluded because of excessive motion leaving 17PD-VH 14 PD-nonVH and 20 controls with usable fMRI dataAll groups showed a typical activation pattern to the checker-board (data available fromNewcastle University e-prints [figure2] eprintnclacuk247552) but there were no significant dif-ferences in activation between any groups In the ROI analysis
e680 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
groups There was a significant group difference in theGABA+Cr ratio (figure 1) with post hoc Tukey testfinding GABA+Cr reduced in PD-VH relative to PD-nonVH The group difference remained significant afterincluding CAMCOG as a measure of cognitive ability in thelinear model (F248 = 327 p = 0047) There were no sig-nificant differences in gray or white matter proportionwithin the voxel between groups (table 2) Within the PDgroup GABA+Cr correlated with visual acuity (Spearmanρ = 04 p = 0025) MMSE (ρ = 035 p = 0047) UPDRS-III(ρ = minus0345 p = 0049) and MAYO total score (ρ = minus0627p lt 0001) There were no significant correlations (p gt 01)with disease duration CAMCOG angle or motion test orthe pareidolia test There were no significant correlationsbetween GABA+Cr and NPI hallucination score aftercontrolling for VH group
There was no significant difference in GABA+Cr levels be-tween participants with PD taking cholinesterase inhibitors vsthose not taking them (0098 SD 0012 vs 0093 SD 0008 t31= 116 p = 026) between those taking antipsychotic agents(quetiapine) vs those not (0088 SD 0007 vs 0097 SD 0011t31 = 142 p = 017) and there was no significant correlationbetween levodopa dose and GABA+Cr (ρ = 0018 p = 09)
The fMRI scans were not acquired on one PD participant whodid not tolerate the full scanning session and 4 PD participantsrsquoscans were excluded because of excessive motion leaving 17PD-VH 14 PD-nonVH and 20 controls with usable fMRI dataAll groups showed a typical activation pattern to the checker-board (data available fromNewcastle University e-prints [figure2] eprintnclacuk247552) but there were no significant dif-ferences in activation between any groups In the ROI analysis
e680 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
(table 3) there were significant within-group activations in allregions apart from the V5 in the PD-VH group (1-sample testT16 = 064 p = 053) However there were no significant dif-ferences in activation between groups for any region In the PDgroup there was a significant positive correlation between theGABA+Cr ratio and activation in the V5 ROI (Pearsondegrees of freedom = 29 r = 0373 p = 0046) but not with theV1ndashV4 ROIs (Pearson r lt 033 p gt 008)
MRI diffusion data were obtained on 17 PD-VH 18 PD-nonVH and 20 controls The TBSS analysis found wide-spread differences between controls and PD-VH in both FAand MD (data available from Newcastle University e-prints[figure 3] eprintnclacuk247552) However after includingCAMCOG as a covariate in the analysis this obviated sig-nificant group differences For the voxelwise correlations be-tween GABA+Cr and both MD and FA controlling for ageand group there was only a very small cluster (24 voxels) inthe posterior corpus callosum This was still significant afterinclusion of CAMCOG in the model
The VBM analysis on the 17 PD-VH 19 PD-nonVH and 20controls found a significant cluster of reduced gray matter inthe right anterior temporal lobe of the PD-VH group com-pared to both the PD-nonVH and the control group (figure 2data available from Newcastle University e-prints [table 2]eprintnclacuk247552) This cluster extended to the hip-pocampus and amygdala in the control vs PD-VH compari-son and there was a nonsignificant cluster in the righthippocampus and amygdala There was also a cluster of re-duced gray matter in the PD-VH compared to the controlgroup in the V4 region (27 V4 26 fusiform gyrus FG116 V3v) With the addition of CAMCOG as a covariate tothe model there were still significant differences in the an-terior temporal lobe for the PD-nonVH vs PD-VH compari-son (figure 2 data available from Newcastle Universitye-prints [table 2] eprintnclacuk247552)
To investigate associations between GABA+ and atrophy weperformed a VBM analysis of gray matter against GABA+Crcontrolling for age and group There was an occipital cluster(66 in V1 and 22 in V2) where GABA+Cr positivelycorrelated with gray matter (data available from NewcastleUniversity e-prints [table 2 figure 4] eprintnclacuk247552) but this was not significant after correcting formultiple comparisons (cluster p = 008 FWE-corrected)
DiscussionWe found reduced levels of GABA+ in the PD-VH group andthere was evidence of gray matter loss in the anterior temporallobe as well as region V4 of the visual cortex There werehowever no alterations in functional activity in response tovisual excitation by the checkerboard stimulus or in whitematterdiffusion parameters once covariates were accounted for
As hypothesized GABA+ concentration was reduced in PD-VH compared to PD-nonVH This agrees with the neuro-pathologic study finding reduced GABAergic markers inDLB5 We found that the participants with hallucinations hadworse acuity which correlated with GABA+ levels in the PDgroup Combined with previous research that found that oc-cipital GABA levels decrease after eye occlusion32 our find-ings support the hypothesis that poor input to the visualcortex leads to levels of inhibitory GABA being reduced tooptimize visual processing at the price of increased mis-classifications of ambiguous stimuli33 The absence of asso-ciations between GABA+ and severity of visual hallucinationssuggests that low GABA levels may predispose people tohallucinate but the occurrence of visual hallucinations iscontrolled by other factors including attention and the visualenvironment If visual hallucinations are partly facilitated bydecreased levels of GABA in the occipital cortex causing hy-perexcitability one therapeutic strategy might be to utilize
Table 3 Checkerboard fMRI BOLD β value from predetermined regions of interest
Abbreviations ANOVA = analysis of variance BOLD = blood oxygen levelndashdependent LGN = lateral geniculate nucleus PD = Parkinson diseaseOne-sample t test for activation within region of interest for each groupa p lt 001b p lt 005
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e681
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
antiepileptic drugs In PD the 5-hydroxytryptamine type 3 (5-HT3) antagonist ondansetron and the 5-HT2 reverse agonistpimavanserin have been used to treat visual hallucinations34
Since 5-HT receptors can modulate release of GABA35 it maybe that the mode of action of these drugs in treating halluci-nations is partly through their effect on GABA
Figure 2 Voxelwise morphometry results showing regions of altered gray matter
Decreased graymatter for (A) PD-VH lt control (B) PD-VH lt PD-nonVH (C) PD-VH lt PD-nonVH controlling for CAMCOG score Voxelwise threshold = p lt 0001uncorrected for multiple comparisons (radiologic convention L = R) CAMCOG = Cambridge Cognition Examination PD-nonVH = Parkinson diseasendashnon-visual hallucinator PD-VH = Parkinson diseasendashvisual hallucinator
e682 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
There was no difference in functional activation between thePD-VH group and any other group This relative lack of dif-ference in functional activity fits with the suggestion that vi-sual hallucinations are a side effect of neural changes aimed atpreserving visual function in the face of worsening visual inputor connectivity36 It is also in agreement with the postmortemobservations5 of little Lewy body disease pathology in theprimary visual cortex but alterations of neurone function inthe fusiform gyrus37 The PD-VH group unlike the PD-nonVH group did not show significant activation in the V5region Duann et al38 reported that activation in the V5 regionto a flickering checkerboard was more variable within-subjectcompared to primary visual cortex and speculated that thismight be due to differing levels of top-down influence such aspaying attention to the motion aspect of the stimulus Wepreviously reported25 reduced activation in V5 in DLB toa motion stimulus and it could be that dysfunction of thisregion contributes to visual hallucinations as object motion isimproperly tracked leading to discrepancies between the in-ternal model of the world and reality
We found only a weak association in the PD group betweenGABA+ and blood oxygen levelndashdependent (BOLD) activa-tion in the V5 ROI This goes against our hypothesizednegative relationship between GABA and occipital BOLDactivations which was based on previous observations thatthese factors are related39 However some recent studies innormal participants have also failed to demonstrate a signifi-cant association between occipital GABA and BOLD40 Pos-sible explanations for the lack of an association include the factthat the BOLD signal is an indirect measure of neuronal ac-tivity and is dependent on blood flow and vascular reactivitywhich could be altered in our participants
We found widespread alterations in MD and FA in the PD-VH group in comparison to controls controlling for ageHowever after including CAMCOG score in the model therewere no significant group differences and there was onlya very small region where GABA+ correlated with FA andnone with MD Previous reports have found widespreadreductions in FA and increases in MD in PD dementia41
suggesting that the DTI group differences were driven byoverall disease severity rather than being specifically relatedto the presence of visual hallucinations Few studies haveinvestigated the relationship between DTI measures and vi-sual hallucinations in PD Lee et al42 found increased MD inthe parietotemporal region of PD-VH with more widespreadchanges in those with dementia Although we found thatdifferences in MD and FA were not specifically related tovisual hallucinations nevertheless it is possible that thepresence of the white matter alterations may have contributedto the formation of hallucinations in the group as suggestedby the disconnection models of visual hallucinations34
The ventral visual stream is likely to be involved in visualhallucination genesis since it is chiefly responsible for objectrepresentation and recognition43 The ventral stream includes
projections from the primary visual cortex to the temporallobe Previous MRI studies of gray matter atrophy in PD-VHhave found a number of regions involved including thetemporal lobe and lateral occipital lobe4445 Ventral streamtemporal areas contain relatively high numbers of Lewybodies4647 with a gradient of increasing density toward theanterior temporal lobe37 and it has been speculated that thesepathologic changes may contribute to visual hallucinations inDLB The midline occipital lobe is relatively spared inDLB4648 and as shown by our fMRI data is functionallyintact suggesting that the observed GABA reduction may bedriven by ventral stream pathology leading to altered con-nectivity between the primary visual cortex and higher visualareas
Our most significant structural finding was gray matter loss inthe temporal pole and amygdala along with reductions in PD-VH relative to controls in area V4 of the occipital lobe The V4area projects to the parahippocampal gyrus43 and is involvedin object recognition and coordinating signals between theearly and higher visual areas The combination of atrophy inventral stream structures and white matter changes includingto the temporal and frontal lobes is consistent with the hy-pothesis49 of disrupted communication between the ventralvisual stream and lateral frontal cortex as being mechanisti-cally involved in the generation of visual hallucinations
Although we used a well-established MRS technique for in-vestigating GABA there are some limitations to the studyThe magnetic resonance spectrum of GABA is complex andcoincides with that of other molecules To maximize thesignal-to-noise ratio of the MRS we did not use macromol-ecule suppression techniques23 and our measured signal thusrepresents a combination of GABA and macromoleculesOther limitations of the study include that because of timeconstraints we acquired a spectrum from only one locationand thus we are not able to say whether the GABA+ changesin PD-VH are specific to the occipital lobe Since visual hal-lucinations are more common in more severe disease and inthose with cognitive impairment it may be that GABA+ levelsrelated to disease severity rather than specifically hallucina-tions However our finding of increased GABA+ remainedsignificant after including a measure of global cognition in themodel suggesting that the changes were not purely driven bydisease stage
Finally an inherent difficulty in investigating visual halluci-nations is that the investigator must rely on subjectivereports from the participant thus risking misclassificationparticularly in individuals with cognitive impairment andmaking it more challenging to find correlates of hallucina-tion severity We cross-checked hallucination reports be-tween participants and their informants to increasereliability and used the previously validated pareidoliatest15 finding significantly increased rates of visual mis-perception in the PD-VH group providing confidence inour visual hallucinations group classification
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e683
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
We found alterations to GABA+ in the occipital cortex to-gether with structural changes in the ventral stream of patientswith PD who had visual hallucinations Further longitudinalstudies are required to elucidate the connection betweenthese changes and how they influence the development ofvisual hallucinations This may have important translationalimplications as remediation of GABAergic function or re-duction in visual cortical hyperexcitability may representa novel treatment approach for visual hallucinations in PD
Author contributionsMichael Firbank drafting and revising the manuscript analysisof data statistical analysis Jehill Parikh drafting and revisingthe manuscript analysis of data Nicholas Murphy revising themanuscript acquisition of data Alison Killen revisingthe manuscript acquisition of data Charlotte Allan revisingthe manuscript analysis of data Daniel Collerton revising themanuscript Andrew Blamire revising the manuscript studyconcept obtaining funding study supervision John-PaulTaylor revising the manuscript study concept obtainingfunding study supervision
AcknowledgmentThe authors are grateful to Professor Etsuro Mori andcolleagues at the Department of Behavioral Neurology andCognitive Neuroscience Tohoku University School ofMedicine Sendai Japan for providing a copy of the pareidoliatask
Study fundingThis research was supported by the National Institute forHealth Research (NIHR) Newcastle Biomedical ResearchCentre (BRC) based at Newcastle Hospitals NHS Founda-tion Trust and Newcastle University CLA is supported byNIHR Newcastle BRC and Local Clinical Research Network(Greenshoots funding)
DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures
Received March 1 2018 Accepted in final form May 23 2018
References1 Hely MA Reid WG Adena MA Halliday GM Morris JG The Sydney multicenter
study of Parkinsonrsquos disease the inevitability of dementia at 20 years Mov Disord200823837ndash844
2 Urwyler P Nef T Muri R et al Visual hallucinations in eye disease and Lewy bodydisease Am J Geriatr Psychiatry 201624350ndash358
3 Muller AJ Shine JM Halliday GM Lewis SJ Visual hallucinations in Parkinsonrsquosdisease theoretical models Mov Disord 2014291591ndash1598
4 Tsukada H Fujii H Aihara K Tsuda I Computational model of visual hallucination indementia with Lewy bodies Neural Netw 20156273ndash82
5 Khundakar AA Hanson PS Erskine D et al Analysis of primary visual cortex indementia with Lewy bodies indicates GABAergic involvement associated with re-current complex visual hallucinations Acta Neuropathol Commun 2016466
6 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184
7 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356
8 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707
9 OrsquoGorman RL Michels L Edden RA Murdoch JB Martin E In vivo detection ofGABA and glutamate with MEGA-PRESS reproducibility and gender effects J MagnReson Imaging 2011331262ndash1267
10 Wood JS FirbankMJMosimannUP et al Testing visual perception in dementia withLewy bodies and Alzheimer disease Am J Geriatr Psychiatry 201321501ndash508
11 Firbank M Kobeleva X Cherry G et al Neural correlates of attention-executivedysfunction in Lewy body dementia and Alzheimerrsquos disease Hum Brain Mapp 2016371254ndash1270
12 Taylor JP Firbank M Barnett N et al Visual hallucinations in dementia withLewy bodies transcranial magnetic stimulation study Br J Psychiatry 2011199492ndash500
13 Cummings JL The Neuropsychiatric Inventory assessing psychopathology in de-mentia patients Neurology 199748S10ndashS16
14 Mosimann UP Collerton D Dudley R et al A semi-structured interview toassess visual hallucinations in older people Int J Geriatr Psychiatry 200823712ndash718
15 Yokoi K Nishio Y Uchiyama M Shimomura T Iizuka O Mori E Hallucinators findmeaning in noises pareidolic illusions in dementia with Lewy bodies Neuro-psychologia 201456245ndash254
16 Ferman TJ Smith GE Boeve BF et al DLB fluctuations specific features that reliablydifferentiate DLB from AD and normal aging Neurology 200462181ndash187
17 Walker MP Ayre GA Cummings JL et al The Clinician Assessment of Fluctuationand the One Day Fluctuation Assessment Scaletwo methods to assess fluctuatingconfusion in dementia Br J Psychiatry 2000177252ndash256
18 Archibald NK Clarke MP Mosimann UP Burn DJ Visual symptoms in Parkinsonrsquosdisease and Parkinsonrsquos disease dementia Mov Disord 2011262387ndash2395
19 Ffytche DH Creese B Politis M et al The psychosis spectrum in Parkinson diseaseNat Rev Neurol 20171381ndash95
20 Mescher M Merkle H Kirsch J Garwood M Gruetter R Simultaneous in vivospectral editing and water suppression NMR Biomed 199811266ndash272
21 Sedley W Parikh J Edden RA Tait V Blamire A Griffiths TD Human auditorycortex neurochemistry reflects the presence and severity of tinnitus J Neurosci 20153514822ndash14828
22 Tkac I Starcuk Z Choi IY Gruetter R In vivo 1H NMR spectroscopy of rat brain at1ms echo time Magn Reson Med 199941649ndash656
23 Edden RA Puts NA Barker PB Macromolecule-suppressed GABA-edited magneticresonance spectroscopy at 3T Magn Reson Med 201268657ndash661
24 Dyke K Pepes SE Chen C et al Comparing GABA-dependent physiologicalmeasures of inhibition with proton magnetic resonance spectroscopy measurement ofGABA using ultra-high-field MRI Neuroimage 2017152360ndash370
25 Taylor JP Firbank MJ He J et al Visual cortex in dementia with Lewy bodiesmagnetic resonance imaging study Br J Psychiatry 2012200491ndash498
26 Brainard DH The Psychophysics Toolbox Spat Vis 199710433ndash43627 Edden RA Puts NA Harris AD Barker PB Evans CJ Gannet a batch-processing tool
for the quantitative analysis of gamma-aminobutyric acid-edited MR spectroscopyspectra J Magn Reson Imaging 2014401445ndash1452
28 Near J Edden R Evans CJ Paquin R Harris A Jezzard P Frequency and phase driftcorrection of magnetic resonance spectroscopy data by spectral registration in thetime domain Magn Reson Med 20147344ndash50
29 Naressi A Couturier C Castang I de Beer R Graveron-Demilly D Java basedgraphical user interface for MRUI a software package for quantitation of in vivomedical magnetic resonance spectroscopy signals Comput Biol Med 200131269ndash286
30 Eickhoff SB Stephan KE Mohlberg H et al A new SPM toolbox for combiningprobabilistic cytoarchitectonic maps and functional imaging data Neuroimage 2005251325ndash1335
31 Smith SM Jenkinson M Johansen-Berg H et al Tract-based spatial statistics vox-elwise analysis of multi-subject diffusion data Neuroimage 2006311487ndash1505
32 Lunghi C Emir UEMorroneMC Bridge H Short-termmonocular deprivation altersGABA in the adult human visual cortex Curr Biol 2015251496ndash1501
33 Bowman AR Bruce V Colbourn CJ Collerton D Compensatory shifts in visualperception are associated with hallucinations in Lewy body disorders Cogn Res PrincImplic 2017226
34 De Deurwaerdere P Di Giovanni G Serotonergic modulation of the activity ofmesencephalic dopaminergic systems therapeutic implications Prog Neurobiol2017151175ndash236
35 Ciranna L Serotonin as a modulator of glutamate- and GABA-mediated neuro-transmission implications in physiological functions and in pathology Curr Neuro-pharmacol 20064101ndash114
36 Collerton D Taylor JP Tsuda I et al How can we see things that are not thereCurrent insights into complex visual hallucinations J Conscious Stud 201623195ndash227
37 Dey M Erskine D Singh P et al Does abnormal ventral visual stream functionunderlie recurrent complex visual hallucinations in dementia with Lewy bodiesPresented at the International DLB Conference December 1ndash4 2015 FortLauderdale
38 Duann JR Jung TP Kuo WJ et al Single-trial variability in event-related BOLDsignals Neuroimage 200215823ndash835
39 Violante IR Ribeiro MJ Edden RA et al GABA deficit in the visual cortex of patientswith neurofibromatosis type 1 genotype-phenotype correlations and functional im-pact Brain 2013136918ndash925
e684 Neurology | Volume 91 Number 7 | August 14 2018 NeurologyorgN
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
40 Harris AD Puts NA Anderson BA et al Multi-regional investigation of the re-lationship between functional MRI blood oxygenation level dependent (BOLD) ac-tivation and GABA concentration PLoS One 201510e0117531
41 Hall JM Ehgoetz Martens KA Walton CC et al Diffusion alterations associated withParkinsonrsquos disease symptomatology a review of the literature Parkinsonism RelatDisord 20163312ndash26
42 LeeWW Yoon EJ Lee JY Park SW Kim YK Visual hallucination and pattern of braindegeneration in Parkinsonrsquos disease Neurodegener Dis 20171763ndash72
43 Kravitz DJ Saleem KS Baker CI Ungerleider LG Mishkin M The ventral visualpathway an expanded neural framework for the processing of object quality TrendsCogn Sci 20131726ndash49
44 Goldman JG Stebbins GT Dinh V et al Visuoperceptive region atrophy in-dependent of cognitive status in patients with Parkinsonrsquos disease with hallucinationsBrain 2014137849ndash859
45 Lenka A Jhunjhunwala RJ Saini J Pal PK Structural and functional neuroimaging inpatients with Parkinsonrsquos disease and visual hallucinations a critical review Parkin-sonism Relat Disord 201521683ndash691
46 Harding AJ Broe GA Halliday GM Visual hallucinations in Lewy body disease relateto Lewy bodies in the temporal lobe Brain 2002125391ndash403
47 Ferman TJ Arvanitakis Z Fujishiro H et al Pathology and temporal onset of visualhallucinations misperceptions and family misidentification distinguishes dementia withLewy bodies from Alzheimerrsquos disease Parkinsonism Relat Disord 201319227ndash231
48 Perry RH Irving D Blessed G Fairbairn A Perry EK Senile dementia of Lewy bodytype a clinically and neuropathologically distinct form of Lewy body dementia in theelderly J Neurol Sci 199095119ndash139
49 Collerton D Perry E McKeith I Why people see things that are not there a novelperception and attention deficit model for recurrent complex visual hallucinationsBehav Brain Sci 200528737ndash757
NeurologyorgN Neurology | Volume 91 Number 7 | August 14 2018 e685
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
DOI 101212WNL0000000000006007201891e675-e685 Published Online before print July 18 2018Neurology
Michael J Firbank Jehill Parikh Nicholas Murphy et al Reduced occipital GABA in Parkinson disease with visual hallucinations
This information is current as of July 18 2018
ServicesUpdated Information amp
httpnneurologyorgcontent917e675fullincluding high resolution figures can be found at
httpnneurologyorgcgicollectiondwiDWIfollowing collection(s) This article along with others on similar topics appears in the
Permissions amp Licensing
httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in
Reprints
httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online
ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by
reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology
![Page 1: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/1.jpg)
![Page 2: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/2.jpg)
![Page 3: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/3.jpg)
![Page 4: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/4.jpg)
![Page 5: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/5.jpg)
![Page 6: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/6.jpg)
![Page 7: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/7.jpg)
![Page 8: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/8.jpg)
![Page 9: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/9.jpg)
![Page 10: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/10.jpg)
![Page 11: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/11.jpg)
![Page 12: Reduced occipital GABA in Parkinson disease with …occipital lobe (data available from Newcastle University e-prints [figure 1]: eprint.ncl.ac.uk/247552). Sequence parameters were](https://reader036.documents.pub/reader036/viewer/2022062507/5fd7cf36f78f8445ba57ba06/html5/thumbnails/12.jpg)
