Received: 10 September 2021 Revised: 9 August 2022 Accepted: 16 August 2022 DOI: 10.1002/brb3.2755 ORIGINAL ARTICLE Shape analysis of subcortical structures in obsessive-compulsive disorder and the relationship with comorbid anxiety, depression, and medication use: A meta-analysis by the OCD Brain Imaging Consortium Jean-Paul Fouche1 Nynke A. Groenewold1 Tatum Sevenoaks1 Sarah Heany1 Christine Lochner2 Pino Alonso3,4,5 Marcelo C. Batistuzzo6,7 Narcis Cardoner4,8,9 Christopher R. K. Ching10 Stella J. de Wit11,12 Boris Gutman13 Marcelo Q. Hoexter6 Neda Jahanshad10 Minah Kim14,15 Jun Soo Kwon15,16,17 David Mataix-Cols18 Jose M. Menchon3,4,5 Euripedes C. Miguel6 Takashi Nakamae19 Mary L. Phillips20 Jesus Pujol21 Yuki Sakai19,22 Je-Yeon Yun23,24 Carles Soriano-Mas3,4,25 Paul M. Thompson10 Kei Yamada26 Dick J. Veltman11,12 Odile A. van den Heuvel11,12 Dan J. Stein1,2 1 Department of Psychiatry and Neuroscience Institute, University of Cape Town, Cape Town, South Africa 2 SAMRC Unit on Risk & Resilience in Mental Disorders, Department of Psychiatry, Stellenbosch University, Stellenbosch, South Africa 3 Department of Psychiatry, Bellvitge University Hospital, Bellvitge Biomedical Research Institute, IDIBELL, Barcelona, Spain 4 Carlos III Health Institute, Centro de Investigacion Biomedica en Red de Salud Mental (CIBERSAM), Madrid, Spain 5 Department of Clinical Sciences, School of Medicine, University of Barcelona, Barcelona, Spain 6 Department & Institute of Psychiatry, University of Sao Paulo Medical School, Sao Paulo, Brazil 7 Department of Methods and Techniques in Psychology, Pontifical Catholic University, Sao Paulo, SP, Brazil 8 Sant Pau Mental Health Group, Institut d’Investigacio Biomedica Sant Pau (IBB-Sant Pau), Hospital de la Sant Creu i Sant Pau, Barcelona, Spain 9 Department of Psychiatry and Forensic Medicine, Universitat Autonoma de Barcelona, Barcelona, Spain 10 Imaging Genetics Center, USC Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine of USC, Marina del Rey, California, USA 11 Department of Psychiatry, Amsterdam UMC, Vrije Universiteit, Amsterdam Neuroscience, Amsterdam, The Netherlands 12 Department of Anatomy & Neurosciences, Amsterdam UMC, Vrije Universiteit, Amsterdam Neuroscience, Amsterdam, The Netherlands 13 Department of Biomedical Engineering, Illinois Institute of Technology, Chicago, Illinois, USA 14 Department of Neuropsychiatry, Seoul National University Hospital, Seoul, Republic of Korea 15 Department of Psychiatry, Seoul National University College of Medicine, Seoul, Republic of Korea 16 Institute of Human Behavioral Medicine, SNU MRC, Seoul, Republic of Korea 17 Department of Brain and Cognitive Sciences, College of Natural Sciences, Seoul National University, Seoul, Republic of Korea 18 Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden 19 Department of Psychiatry, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan 20 Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, USA 21 MRI Research Unit, Radiology Department, Hospital del Mar, Barcelona, Spain 22 ATR Brain Information Communication Research Laboratory Group, Kyoto, Japan This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Brain and Behavior published by Wiley Periodicals LLC. Brain Behav. 2022;12:e2755. wileyonlinelibrary.com/journal/brb3 1 of 12 https://doi.org/10.1002/brb3.2755 2 of 12 FOUCHE ET AL . 23 Seoul National University Hospital, Seoul, Republic of Korea 24 Yeongeon Student Support Center, Seoul National University College of Medicine, Seoul, Republic of Korea 25 Department of Social Psychology and Quantitative Psychology, Universitat de Barcelona—UB, Barcelona, Spain 26 Department of Radiology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan Correspondence Jean-Paul Fouche, Department of Psychiatry Abstract and Neuroscience Institute, University of Cape Town, Cape Town, South Africa. Objective: Neuroimaging studies of obsessive-compulsive disorder (OCD) patients Email: jpfouche123@gmail.com have highlighted the important role of deep gray matter structures. Less work has however focused on subcortical shape in OCD patients. Funding information FIS PI18/00856 grant, NIA T32AG058507, Methods: Here we pooled brain MRI scans from 412 OCD patients and 368 controls to NIH grant U54 EB 020403, RO1MH117601, perform a meta-analysis utilizing the ENIGMA-Shape protocol. In addition, we inves- Basic Research Laboratory Program through the National Research Foundation of Korea tigated modulating effects of medication status, comorbid anxiety or depression, and (NRF) (grant no. 2018R1A4A1025891), NIMH: disease duration on subcortical shape. P50MH106435, R37MH100041, R01MH115466, R01MH059929, Results: There was no significant difference in shape thickness or surface area between R01MH060952 and the Pittsburgh OCD patients and healthy controls. For the subgroup analyses, OCD patients with Foundation, PI16/00889 and PI19/01171 grants from the Carlos III Heath Institute; comorbid depression or anxiety had lower thickness of the hippocampus and caudate PERIS SLT006/17/00249 grant from nucleus and higher thickness of the putamen and pallidum compared to controls. OCD Generalitat de Catalunya, and NIH grant U54 EB 020403 patients with comorbid depression had lower shape surface area in the thalamus, cau- date nucleus, putamen, hippocampus, and nucleus accumbens and higher shape surface area in the pallidum. OCD patients with comorbid anxiety had lower shape surface area in the putamen and the left caudate nucleus and higher shape surface area in the pallidum and the right caudate nucleus. Further, OCD patients on medication had lower shape thickness of the putamen, thalamus, and hippocampus and higher thick- ness of the pallidum and caudate nucleus, as well as lower shape surface area in the hippocampus and amygdala and higher surface area in the putamen, pallidum, and cau- date nucleus compared to controls. There were no significant differences between OCD patients without co-morbid anxiety and/or depression and healthy controls on shape measures. In addition, there were also no significant differences between OCD patients not using medication and healthy controls. Conclusions: The findings here are partly consistent with prior work on brain volumes in OCD, insofar as they emphasize that alterations in subcortical brain morphology are associated with comorbidity and medication status. Further work is needed to understand the biological processes contributing to subcortical shape. KEYWORDS anxiety, depression, gray matter, magnetic resonance imaging, neuroimaging, obsessive- compulsive disorder, subcortical 1 INTRODUCTION Fineberg, Marteau & Legault, 2010). Patients with OCD have been found to have functional and structural alterations in the parallel Obsessive-compulsive disorder (OCD) is a debilitating neuropsychi- cortico-striatal-thalamic-cortical circuits of the brain, as well as alter- atric disorder that affects 1–3% of the population (Ruscio, Stein, Chiu ations in the fronto-parietal, fronto-limbic, and cerebellar regions & Kessler, 2010), and that is associated with significant impairment, (Milad & Rauch, 2012; van den Heuvel et al., 2016). Accordingly, reduced quality of life, and socio-economic burden (Hollander, Stein, there has been a focus in OCD neuroimaging research on volumetric FOUCHE ET AL . 3 of 12 alterations in deep gray matter structures, including the basal ganglia accumbens; NAcc). The ventral putamen and nucleus accumbens have (Gilbert et al., 2008; Pujol et al., 2004; Szeszko et al., 2008; Zarei et al., a more prominent role in emotion processing in a motivational context, 2011). whereas the dorsal putamen, along with other nuclei in the basal gan- Reports on basal ganglia volumes in OCD have however been het- glia, is more strongly implicated in motor learning (Haber, 2016; Haber erogeneous (Peng et al., 2012; Radua, van den Heuvel, Surguladze & Knutson, 2010). & Mataix-Cols, 2010; Rotge et al., 2010). Prospective meta-analyses To our best knowledge, no multi-site studies have been conducted can be useful to examine robust effects in heterogeneous populations, that compared the shape of subcortical regions between OCD patients especially when based on collaborative and harmonized analyses of and controls. The ENIGMA consortium has developed a pipeline for neuroimaging datasets, because this approach is less susceptible to investigating the shape of subcortical brain regions in multi-site studies publication bias (Radua et al., 2010; Zugman et al., 2020). Using a (Gutman, Wang, Rajagopalan, Toga & Thompson, 2012; Gutman et al., similar approach, a study by the ENIGMA-OCD consortium analyzed 2013a, 2013b), which facilitates harmonization of brain atlases, pro- subcortical volumes of 1,830 OCD patients and 1,759 controls and cessing methods, and statistical models across sites and meta-analysis found, on average, larger pallidum, and smaller hippocampus volumes of the aggregated group-level results. Notably, shape analysis allows in OCD, with findings especially pronounced in patients on medica- for the exploration of subtle morphological changes with minimal tion (Boedhoe et al., 2017). Age and disease duration may also impact manipulation of the source image (for example, without spatial smooth- basal ganglia findings; Pujol et al. (2004) reported that larger volumes ing). It can help to refine and extend results obtained with volumetric of striatal structures in OCD were associated with older age and longer methods such as voxel-based morphometry by utilizing thickness and disease duration, a finding supported by the OCD Brain Imaging Con- surface area parameters that can precisely localize regional geometric sortium (OBIC) (De Wit et al., 2014) and the ENIGMA-OCD study shape deformations (Gutman et al., 2012, Gutman et al., 2013a, 2013b). (Boedhoe et al., 2017). Another caveat is that many OCD patients The present study analyzed data from OCD patients and healthy present with comorbid anxiety and depression (Rasmussen & Eisen, controls assessed at six academic centers across three continents (Asia, 1992), and these disorders may further impact volumetric findings. Europe, and South America). By pooling data and using standardized There are scarce data on shape alterations in OCD, despite the ENIGMA-Shape pipelines, we aimed to have enough statistical power availability of several new methods for providing shape information to compare subcortical shape between groups of OCD patients and (Shi et al., 2014; Wang et al., 2011), and growing evidence that shape healthy controls, as well as examine the associations of regional sub- variability of brain structures has an important heritable component cortical shape with factors such as medication status, disease duration, (Roshchupkin et al., 2016). Two studies examined the shape of the hip- and psychiatric comorbidity. A meta-analysis was performed as it can pocampus in OCD patients versus controls. The first study (Hong et al., optimally account for cohort-specific variations in sample characteris- 2007) identified surface deformities in the anterior part of the hip- tics by performing clustered analysis per cohort (Boedhoe et al., 2017). pocampus and around the border between the body and tail, based on Based on previous studies (Hong et al., 2007; Zhang et al., 2019a, manual segmentation of the hippocampus. The second study (Zhang 2019b) and our prior mega-analytic work in OCD (Boedhoe et al., 2017; et al., 2019) showed a lateral displacement in the middle and poste- Kong et al., 2020), we expected alterations to thickness and surface rior hippocampus in OCD patients using a vertex-wise shape analysis. area in the pallidum, hippocampus, and amygdala in patients with OCD. Moreover, one study (Zhang et al., 2019b) investigating the shape of six For completeness, other regions of interest included the thalamus, subcortical regions revealed right-sided expansions in the pallidum and caudate, putamen, and nucleus accumbens. Further, we expected that lateral amygdala in OCD patients using a vertex-wise shape analysis. alterations in subcortical shape would be more pronounced in OCD The subcortical structures that have been implicated in OCD con- patients on psychotropic medication and with longer disease duration. tain functionally distinct subregions. For example, the hippocampus is not a singular unit, but consists of several subregions, including the cornu ammonis subfields (CA) 1–4, dentate gyrus (DG), and the subicu- 2 METHODS lum (SUB). While the more dorsal regions are involved in memory formation and spatial cognition, the ventral regions subserve affective 2.1 Participants processing (Fanselow & Dong, 2010). Similarly, the amygdala contains functionally distinct subnuclei, most importantly the basolateral and Data were obtained from six research centers participating in the centromedial amygdala (BLA, CMA) (Mosher, Zimmerman & Gothard, OCD Brain Imaging Consortium OBIC. The combined cohort has been 2010; Sah, Faber, Lopez de Armentia & Power, 2003). The CMA, with its described previously (De Wit et al., 2014). More detailed information reciprocal connections with the basal ganglia, midbrain, and brain stem, regarding demographics, comorbidity, and symptom dimensions (mea- appears to be involved in allocating attention and generating auto- sured using the Yale-Brown Obsessive-Compulsive Scale [Y-BOCS] nomic responses to salient environmental cues. In contrast, the BLA, (Goodman et al., 1989)) is also available in Table 1. Participants were with its connections to the CMA and with extensive cortical regions, screened for DSM-IV axis I disorders, and exclusion criteria were ages is primarily involved in evaluating the emotional content of sensory under 18 and above 65, current psychotic disorder, recent history of inputs (Mosher et al., 2010). Furthermore, the striatum is comprised of a substance use disorder, intellectual disability, and severe organic or the dorsal striatum (caudate, putamen) and ventral striatum (nucleus neurological pathology. Psychiatric comorbidity was allowed if OCD 4 of 12 FOUCHE ET AL . TA B L E 1 Demographic and clinical characteristics of the healthy controls (N = 368) and OCD patient (N = 412) group (as shown in De Wit et al. (2014)) Characteristic OCD patients Healthy controls Statistics Mean SD Mean SD t p Age (years) (range) 32.1 (29.4) 9.6 30.2 (28.5) 9.3 2.9 0.004 Education level (years (range) 13.7 (10.8) 2.8 14.6 (8.7) 3.1 −4.0 <0.001 YBOCS score (IQR) 24.9 (5.9) 6.2 Age at onset of clinical symptoms 20.1 (7.8) 8.7 (years, range) N % N % χ2 p Male 202 49.0 195 53.0 1.2 0.28 Right-handed 354 85.9 330 89.7 1.0 0.65 Ethnicity 2.7 0.26 Caucasian 195 47.3 192 52.2 Asian 171 41.5 146 39.7 Other 6 1.5 11 3.0 Medication use at time of scan 176 42.7 0 0.0 210.1 <0.001 Current comorbidity 149 36.2 0 0.0 210.1 <0.001 Lifetime comorbidity 213 51.7 7 1.9 253.7 <0.001 Prepubertal OCD onset 51 13.0 OCD symptom dimensions Mean SD (YBOCS) Aggression/checking (IQR) 236 (13.5) 57.2 Contamination/cleaning (IQR) 202 (28.7) 49.0 Symmetry/ordering (IQR) 168 (34.2) 40.8 Sexual/religious (IQR) 130 (21.5) 31.6 Hoarding (IQR) 87 (13.6) 21.1 was the primary diagnosis. The complete sample consisted of 412 this, a medial model of each individual surface model is fitted as demon- patients with primary OCD and 368 healthy controls after quality strated by Gutman et al. (2012). The medial and intrinsic features of checking of data and removal of problematic scans (see De Wit et al., the subcortical shape drive registration to a template parametrically 2014). on the sphere. To minimize metric distortion, the registration was per- formed in the fast spherical demons framework (Gutman et al., 2013a, 2013b). The templates and mean medial curves are distributed as 2.2 Image processing part of the ENIGMA-Shape package (http://enigma.usc.edu/ongoing/ enigma-shape-analysis). The 780 1.5 tesla T1 -weighted brain MRI scans were processed with The resulting meshes for the 14 structures consisted of a total of FreeSurfer V5.3 (Fischl et al., 2002) on the Nehalem cluster at the Cen- 27,120 vertices. For these vertices, two measures were used to quan- tre for High Performance Computing (CHPC), Rosebank, Cape Town, tify shape, namely the radial distance and the natural logarithm of the South Africa. This was to obtain segmentations of seven subcorti- Jacobian determinant. The radial distance represented the distance of cal structures in both hemispheres: pallidum, hippocampus, amygdala, the vertex from the medial curve of the structure and is referred to as thalamus, caudate nucleus, putamen, and nucleus accumbens. the thickness of the structure. The Jacobian determinant captures the Segmentations were then processed using the ENIGMA shape anal- deformation required to map the subject-specific vertex to a template ysis pipeline (Gutman et al., 2012, Gutman et al., 2013a, 2013b). and indicates surface dilation due to subregional volume change, In short, a mesh model was created for the boundary of each offering a local measure of the surface area of the structure. The structure. Subcortical shapes were registered using the “Medial segmentations were inspected visually for any artifacts, as well as for Demons” framework, which matches shape curvatures and medial fea- accuracy of anatomical structure according to standardized ENIGMA tures to a pre-computed template (Gutman et al., 2013a, 2013b). To do protocols (http://enigma.ini.usc.edu/protocols/imaging-protocols/) FOUCHE ET AL . 5 of 12 2.3 Statistical analysis 24.9(SD = 6.2) and a mean age of clinical onset of 20.1 years (SD = 8.7). See De Wit et al. (2014) and Fouche et al. (2017), as well For both thickness and surface area of subcortical regions, group as Tables 1 and 2 for further details on the demographic and clinical differences between OCD patients and controls were investigated characteristics of the sample. by using a general linear model (GLM) with age, sex, and intracranial volume (ICV) as covariates within each site. The resulting group-level effect sizes and regression parameters were aggregated across sites 3.2 Group differences in subcortical shape in a random-effects mass univariate meta-analysis. To investigate age-by-diagnosis interactions, both linear and quadratic age terms 3.2.1 Subcortical shape in OCD patients compared were used as moderators in the main group analysis. to healthy controls Next, group differences between OCD patients and controls were examined for several subsamples selected on comorbidity and medica- Shape surface area tion use. OCD patients with and without lifetime depression comorbid- There was no significant difference (p < 0.05 searchlight FDR- ity were each contrasted with controls, and with each other, in separate corrected) in shape surface area for any of the subcortical structures comparisons. Hereafter, OCD patients with and without lifetime anx- when comparing the OCD patients (n = 412) to healthy controls iety comorbidity were each contrasted with controls, and with each (n = 368). There was also no significant age-by-diagnosis interaction or other. Furthermore, OCD patients with and without psychotropic med- age2 -by-diagnosis interaction for shape surface area. ication use were each contrasted with controls, and with each other. The statistical approach for these comparisons was the same as in the Shape thickness main analysis. There was no significant difference (p < 0.05 searchlight FDR- Finally, the shape correlates of clinical variability within the patient corrected) in shape thickness for any of the subcortical structures sample were assessed. Multiple regression analyses were performed when comparing the OCD patients (n = 412) to healthy controls to investigate the effects of OCD severity, duration of disease, age (n = 368). There was also no significant age-by-diagnosis interaction or of OCD onset, and symptom dimensions (see supplementary file for age2 -by-diagnosis interaction for shape thickness. detail of these covariates) within the OCD group. Sample selection was based on the availability of clinical information for each analysis. Symptom dimensions were coded as follow: 0 for lifetime absence of 3.2.2 Group differences in subcortical shape: a dimension and 1 for lifetime presence. These dichotomous variables Comorbid depressive disorder were then included in the multiple regression analysis. To correct for multiple comparisons per subcortical structure, a Shape surface area searchlight false discovery rate (FDR) threshold (Langers, Jansen & OCD patients with comorbid depression (n = 101) showed lower sur- Backes, 2007) was used at q = 0.05. This correction was applied face area in the right hippocampus (Cohen’s d = −0.173, 2.3% change) locally to individual structures as we were interested in specific sub- and higher surface area in the right pallidum (Cohen’s d = 0.174, 4.5 cortical structures, namely the pallidum, hippocampus, and amygdala % change) compared to controls (n = 386). Other regions that demon- as our ROIs, in addition to other subcortical structures that were strated lower surface area were the left thalamus (Cohen’s d = −0.184, explored, namely the thalamus, caudate nucleus, putamen, and nucleus 4.7% change), left caudate nucleus (Cohen’s d = −0.150, 7.5% change), accumbens. The distance in the searchlight procedure was defined as left nucleus accumbens (Cohen’s d = −0.162, 1.3% change) as well the Euclidean distance between atlas vertices within each subcortical as right putamen (Cohen’s d = −0.139, 6.7% change). Significant structure. differences in shape surface area are presented in Figure 1 and Table 3. There was no significant difference in shape surface area between OCD patients without comorbid depression (n = 287) and controls 3 RESULTS (n = 368). There was also no significant difference between OCD patients with (n = 101) and without (n = 287) comorbid depression. 3.1 Sample characteristics Shape thickness Patients with primary OCD (n = 412) and healthy controls (n = 368) OCD patients with comorbid depression (n = 101) had lower thick- were matched on sex, handedness, and ethnicity in individual cohorts. ness in the left hippocampus (Cohen’s d = −0.135, 5.4% change) and However, patients were significantly older (OCD group: 32.1 years higher thickness in the right pallidum (Cohen’s d = 0.159, 7.8% change) [SD = 9.6]; control group: 30.2 years [SD = 9.3], t = 2.9, p = 4 × 10−3 ) compared to controls. For the other regions, there was lower thick- and had a lower level of education (OCD group: 13.7 years [SD = 2.8]; ness in the left caudate nucleus (Cohen’s d = −0.132, 2.8% change) and control group: 14.6 years [SD 3.1], t = −4.0, p < 1 × 10−3 ) than the con- higher thickness in the right putamen (Cohen’s d = 0.167, 6.5% change) trol group in the aggregated cohort. The OCD group had a mean of OCD patients when compared to controls. There was no significant Yale-Brown Obsessive Compulsive Scale (Y-BOCS) score of difference in shape thickness or surface area between OCD patients TA B L E 2 Demographic characteristics for OCD subgroups and healthy controls 6 of 12 OCD patients with anxiety OCD patients without OCD patients with OCD patients without OCD patients on OCD patients not Characteristic (n = 83) anxiety (n = 190) depression (n = 101) depression (n = 287) med (n = 176) on med (n = 222) Statistics Age (years, range) 33.8, 25.6 (8.5) 32.6, 23.9 (7.9) 31.5, 28.5 (8.2) 30.8, 29.2 (8.4) 31.2, 26.9 (6.9) 32.2, 28.1 (9.1) F = 2.86; (SD) p = 0.89 Education level 12.8, 10.1 (2.5) 13.5, 9.8 (3.1) 14.7, 10.7 (2.8) 13.2, 10.5 (2.9) 13.8, 9.9 (3.1) 14.1, 10.3 (3.0) F = 3.01; (years) p = 0.67 Age of onset of 19.1, 7.5 (9.2) 21.5, 8.1 (8.7) 22.1, 7.7 (9.3) 20.5, 8.2 (9.8) 21.3, 7.9 (8.9) 20.5, 8.3 (8.3) F = 2.97; p = 0.65 OCD (years, range) (SD) Male 41 (49%) 93 (49%) 48 (48%) 140 (49%) 89 (51%) 108 (49%) Χ2 = 3.8; p = 0.87 Right-handed 68 (82%) 168 (89%) 85 (84%) 248 (86%) 159 (90%) 195 (88%) χ2 = 2.9;p = 0.79 Ethnicity Caucasian 41 (49%) 101 (53%) 53 (52%) 132 (46%) 93 (53%) 123 (55%) χ2 = 2.6;p = 0.58 Asian 42 (51%) 87 (46%) 47 (47%) 154 (54%) 81 (46%) 99 (45%) χ2 = 3.7p 0.89 Other 0 2 (1%) 1 (1%) 1 (< 1%) 2 (1%) 0 χ2 = 3.2p 0.79 depression; OCDanx: OCD patients with comorbid anxiety. thickness (Radial distance); TBM: Shape surface area (Jacobian of the different subcortical regions for clarification. RAD: Shape OCD patients on medication; OCDdep: OCD patients with major structures between OCD cohorts and controls. The legend next to each figure shows the extent of the shape change (B-value). Results determinant); LH: left hemisphere; RH: right hemisphere; OCDmed: differences in shape thickness are presented in Figure 1 and Table 4. are shown at p < 0.05, FDR corrected. The legend at the top is a label F I G U R E 1 Maps showing shape differences for all the subcortical OCD patients with comorbid depression (n = 287) and controls (n = There was no significant difference in shape thickness between with (n = 101) and without (n = 287) comorbid depression. Significant FOUCHE ET AL . TA B L E 3 A summary of all the significant group comparisons of shape surface area between OCD patients and healthy controls (p < 0.05, corrected for multiple comparisons with FDR FOUCHE ET AL . Searchlight) Shape surface area effect sizes (Cohen’s d) and percentage change of the structure Nucleus Group comparison Amygdala Hippocampus Putamen Caudate nucleus accumbens Thalamus Pallidum OCDdep (n = 101) vs. Ns RH: d = −0.173 (2,3%) RH: d = −0.139 LH: d = −0.150 LH: d = −0.162 LH: d = −0.184 RH: d = 0.174 (4.5%) controls (n = 368) (6.7%) (7.5%) (1.3%) (4.7%) OCDanx (n = 190) vs. Ns Ns RH: d = −0.212 RH: d = 0.200 Ns Ns RH: d = 0.178 (4.6%) controls (n = 368) (6.1%) (5.8%) LH: d = −0.115 (4.3%) OCDmed (n = 176) vs. RH: d = −0.137 RH: d = −0.153 (5.6%) RH: d = 0.132 RH: d = 0.123 Ns Ns RH: d = 0.144 (5.7%) controls (n = 368) (6.8%) (5.5%) (6.2%) Abbreviations: Ns: not significant.;OCD: obsessive-compulsive disorder; OCDanx: OCD patients with comorbid anxiety disorder; OCDdep: OCD patients with comorbid major depression; OCDmed: OCD patients on medication. TA B L E 4 A summary of all the significant group comparisons of shape thickness between OCD patients and healthy controls (p < 0.05, corrected for multiple comparisons with FDR Searchlight) Shape thickness effect size (Cohen’s d) and percentage change in the structure Nucleus Group comparison Amygdala Hippocampus Putamen Caudate nucleus accumbens Thalamus Pallidum OCDdep (n = 101) vs. controls Ns LH: d = −0.135 RH: d = 0.167 (6.5%) LH: d = −0.132 Ns Ns RH: d = 0.159 (n = 368) (5.4%) (2.8%) (7.8%) OCDanx (n = 190) vs. controls Ns LH: d = −0.111 RH: d = 0.205 (7.8%) LH: d = −0.139 Ns Ns RH: d = 0.202 (n = 368) (2.4%) (5.6%) (4.3%) OCDmed (n = 176) vs. Ns RH: d = −0.146 LH: d = −0.115 RH: d = 0.175 (4.4%) Ns RH: d = −0.126 RH: d = 0.177 controls (n = 368) (4.6%) (7.1%) (6.4%) (5.9%) LH: d = −0.161 (5.8%) Abbreviations: Ns: not significant; OCD: obsessive-compulsive disorder; OCDanx: OCD patients with comorbid anxiety disorder; OCDdep: OCD patients with comorbid major depression; OCDmed: OCD patients on medication. 7 of 12 8 of 12 FOUCHE ET AL . 368), and also not between OCD patients with (n = 101) and without Cohen’s d = −0.161, 5.8% change, RH: Cohen’s d = −0.146, 4.6% (n = 287) comorbid depression. change), and higher thickness in the right pallidum (Cohen’s d = 0.177, 5.9% change) compared to controls. Other regions that demonstrated lower thickness were the left putamen (Cohen’s d = −0.115, 7.1% 3.2.3 Group differences in subcortical shape: change) and right thalamus (Cohen’s d = −0.126, 6.4% change) com- Comorbid anxiety disorder pared to controls. There was also higher thickness of the right caudate nucleus (Cohen’s d = 0.175, 4.4% change) in the OCD patients. Shape surface area There was no significant difference in shape thickness between OCD patients with comorbid anxiety (n = 83) demonstrated higher sur- OCD patients without medication use (n = 222) and controls (n = face area in the right pallidum (Cohen’s d = 0.178, 4.6% change) than 368), and also not between OCD patients with (n = 176) and without controls (n = 368). For the other regions, there was lower surface area medication use (n = 222). in the left caudate nucleus (Cohen’s d = −0.115, 3.4% change) and right putamen (Cohen’s d = −0.212, 6.1% change), as well as higher surface area in the right caudate nucleus (Cohen’s d = 0.200, 5.8% change), 3.2.5 Associations of other clinical variables with compared to controls. subcortical shape in OCD patients There was no significant difference in shape surface area between OCD patients without comorbid anxiety (n = 190) and controls (n = Regression analyses indicated that there was no significant association 368). There was also no significant difference between OCD patients of symptom dimensions (n = 331), total YBOCS scores (n = 331), dura- with (n = 83) and without (n = 190) comorbid anxiety. tion of illness (n = 412), and age of onset (n = 412) with shape thickness or surface area for any of the subcortical regions. Shape thickness OCD patients with comorbid anxiety (n = 83) had lower thickness in the left hippocampus (Cohen’s d = −0.111, 2.4% change) and higher 4 DISCUSSION thickness in the right pallidum (Cohen’s d = 0.202, 4.3% change) com- pared to controls. The exploration of other subcortical regions showed Comparison of OCD patients and controls revealed no statistically sig- lower thickness in the left caudate nucleus (Cohen’s d = −0.139, 5.6% nificant differences in shape measures. However, alterations in shape change) and higher thickness in the right putamen (Cohen’s d = 0.205, were detected in multiple subcortical regions within OCD patients with 7.8% change). comorbid depression and anxiety and OCD patients on medication. In There was no significant difference in shape thickness between line with our hypotheses, the identified alterations involved the hip- OCD patients without comorbid anxiety (n = 190) and controls (n = pocampus and pallidum for all three subgroups of OCD patients. How- 368), and also not between OCD patients with (n = 83) and without (n ever, the shape of the amygdala was only found to be altered in OCD = 190) comorbid anxiety. patients that used medication. In addition, the caudate nucleus, puta- men, and thalamus were most consistently found to have an altered 3.2.4 Group differences in subcortical shape: shape in OCD patients with comorbidity and medication use compared Psychotropic medication use to controls. Thus, this high-powered study using high-dimensional shape analysis identified novel shape alterations in several subcortical Shape surface area regions that had not been identified in previous studies with smaller OCD patients with medication use (n = 176) had lower surface area samples (Hong et al., 2007; Zhang, Hu, Lu, et al., 2019; Zhang, Hu, Li, in the right hippocampus (Cohen’s d = −0.153, 5.6% change) and right et al., 2019) in OCD patients. Moreover, age, symptom dimensions, amygdala (Cohen’s d = −0.137, 6.8% change) and higher surface area in and disease duration were not associated with shape alterations the right pallidum (Cohen’s d = 0.144, 5.7% change) compared to con- in OCD. trols (n = 368). Other regions that demonstrated higher surface area One of the primary regions of interest for this study was the pal- in OCD patients with medication use were the right putamen (Cohen’s lidum. In OCD patients with depression or anxiety comorbidity or d = 0.132, 5.5% change), and right caudate nucleus (Cohen’s d = 0.123, on medication, the external part of the pallidum consistently demon- 6.2% change). strated higher surface area and thickness. Previous meta-analyses There was no significant difference in shape surface area between have reported greater pallidum volumes in OCD (Peng et al., 2012; OCD patients that did not use medication (n = 222) and controls (n = Radua et al., 2010; Rotge et al., 2010; Radua & Mataix-Cols, 2009), 368). There was also no significant difference between OCD patients and lower volume of this structure in anxiety disorders (Radua et al., with (n = 176) and without medication use (n = 222). 2010). As shape thickness and surface area are positively associated with volume (Roshchupkin et al., 2016), the findings here are consistent Shape thickness with this literature. One previous study in 83 OCD patients without OCD patients on psychotropic medication (n = 176) compared to con- medication use and without depression or anxiety comorbidity found trols (n = 368) had lower thickness in the bilateral hippocampus (LH: that expansion of the right pallidum was associated with illness dura- FOUCHE ET AL . 9 of 12 tion and symptom severity (Zhang, Hu, Li, et al., 2019). Furthermore, There are limited longitudinal studies that suggest that SRI treat- a smaller study in 22 OCD patients (11 with medication use and 5 ment can normalize volume of the thalamus and putamen in OCD with depression or anxiety comorbidity) reported trend-level expan- patients (Gilbert et al., 2000; Hoexter et al., 2012). However, from these sion of surface area in the pallidum (Shaw et al., 2015). The association studies, it is unclear whether medication use altered brain structure observed between surface area expansion of the right pallidum and or if the changes merely reflect symptom improvement. The ENIGMA- several clinical parameters could perhaps be explained by an asso- OCD consortium performed a large machine learning classification ciation with global clinical severity of OCD, associated with greater analysis of 2304 OCD patients and 2068 healthy controls. The analysis comorbidity. identified significant features that differentiated between medicated Other primary regions of interest included the hippocampus and versus unmedicated OCD patients in multivariate models, including amygdala. In OCD patients with depression or anxiety comorbidity, or the left thalamus and pallidum (Bruin et al., 2020). These findings on medication, there was also lower surface area and thickness of the are in line with our shape results for the thalamus and pallidum and hippocampal complex. The lateral ventral hippocampus has reciprocal support the interpretation that OCD-related alterations in brain mor- connections with the medial prefrontal cortex and together with the phometry are most pronounced in medicated patients. Future research amygdala is important in memory formation and emotional regulation should examine the association between subcortical shape and treat- (Izquierdo & Medina, 1997). Previous studies, including ENIGMA-OCD ment response in OCD patients, to understand whether shape features studies, have reported volume abnormalities in this region in OCD are impacted by disease severity and can signal treatment resistance or (Boedhoe et al., 2017; De Wit et al., 2014; Peng et al., 2012; Radua conversely treatment responsiveness. et al., 2010; Rotge et al., 2010, 2009). It is notable that the hippocampal The putamen was another region that was investigated without CA1-3 complex is susceptible to stress-related toxicity (Kassem et al., hypothesis. In OCD patients with comorbid depression and anxiety, 2013); however, this is not necessarily specific to only OCD but can also there was bilateral lower surface area and higher thickness in the puta- relate to comorbid depression and anxiety. Other ENIGMA working men. Volume changes in the putamen have been associated with OCD groups focusing on major depression (Schmaal et al., 2016), schizophre- characteristics previously (Kubota et al., 2016). Other studies have nia (van Erp et al., 2016), bipolar disorder (Hibar et al., 2018), and PTSD also found reduced volume and shape deformations of the bilateral (Logue et al., 2018) have found lower hippocampal volume in patients putamen in patients with major depressive disorder and social anxi- compared to controls. ety disorder (Lu et al., 2016; Zhao et al., 2017). However, there are In studies on Tourette syndrome and OCD, there is some evidence few other studies that support this finding and therefore it should be that a relative lower right amygdala volume is associated with higher interpreted with caution. Previous VBM meta-analyses have reported scores on the aggression/checking symptom dimension (Pujol et al., larger volumes of the putamen (Radua et al., 2010; Rotge et al., 2010) 2004). Amygdala alterations have also been demonstrated in OCD and in OCD patients, both adults and children. However, there are some Tourette syndrome (Ludolph et al., 2008; Peterson et al., 2007; Werner studies or meta-analyses that have shown no change of the puta- et al., 2010). In addition, one pediatric study found significant amyg- men and surrounding structures (Rotge et al., 2009). These discrepant dala volume reduction after treatment with paroxetine (Szeszko et al., VBM findings might be related to the shape changes of the putamen 2004). The association between amygdala shape and clinical charac- demonstrated here, with larger observed volume driven by putamen teristics of OCD deserves further investigation in samples where more thickness. The simultaneous higher thickness and lower surface area detailed clinical information is available may effectively mask the findings of deep gray matter volume in other Other regions that were examined without hypothesis included studies. the thalamus. In adult OCD patients on medication, there was lower Other significant findings in this study indicate lower surface thickness in the ventral lateral nucleus of the thalamus. This nucleus area in the left nucleus accumbens in OCD patients with comorbid is responsible for relaying information from the cerebellum and pal- depression, lower surface area in the right amygdala in OCD patients lidum to the primary motor regions, which could be associated with the on medication, and higher surface area in the right caudate nucleus as motor tics seen in some patients. Although not all data are consistent well as lower surface area in the left caudate nucleus in OCD patients (Boedhoe et al., 2017), a previous meta-analysis that included children with comorbid anxiety. Contrary to our other findings discussed, there with OCD demonstrated a larger thalamus in the unmedicated sub- is limited literature to support these findings. Future studies would sample of children with OCD (Rotge et al., 2009; Rotge et al., 2010). benefit in investigating volume and shape changes in these subcortical Prior work has suggested that thalamic volume in pediatric OCD is nor- regions in order to support and replicate these findings. malized (i.e. becomes smaller) after treatment with paroxetine (Gilbert While patients in this study were recruited with a primary diagnosis et al., 2000). A review by Piras et al. (2015) also highlighted that the of OCD, it is important to exclude the possibility that changes in shape thalamus is a key brain region associated with OCD in adults, with the of subcortical structures found here are not due to depression or anx- majority of studies indicating larger volumes in this region. It is possible iety comorbidity or medication use. Notably, most of the subcortical that thalamic enlargement is an early marker of OCD, and that medica- regions identified as differing in OCD are not prominently featured in tion in conjunction with brain maturation throughout adolescence acts studies investigating shape and volume alterations of subcortical struc- to counteract enlargement in this region, as reflected by the smaller tures in anxiety and depression (Bas-Hoogendam et al., 2017; Schmaal thickness of the ventral lateral nuclei here. et al., 2016; ; Zhao et al., 2017). Comorbidity and medication use can 10 of 12 FOUCHE ET AL . be markers of disease severity, and this may explain some of the con- Basic Research Laboratory Program through the National Research sistency in findings across these relevant subsamples. In addition, the Foundation of Korea (NRF) (grant no. 2018R1A4A1025891), NIMH: possibility that the comorbidity of OCD with depression or anxiety can P50MH106435, R37MH100041, R01MH115466, R01MH059929, possibly amplify regional brain alterations related to OCD deserves R01MH060952 and the Pittsburgh Foundation, PI16/00889 and further investigation. This argues against the exclusion of these comor- PI19/01171 grants from the Carlos III Heath Institute; PERIS bidities in neuroimaging studies of OCD to allow for an exploration of SLT006/17/00249 grant from Generalitat de Catalunya, NIH grant the full clinical spectrum. U54 EB 020403 This study is the largest shape analysis of subcortical structures in OCD to date. Strengths of this work include (1) the large sample CONFLICT OF INTEREST size, which increases the power to detect smaller statistical effects; (2) Christopher R. K. Ching received partial grant support from Biogen, methodological homogeneity, as all data were processed and analyzed Inc. (Boston, USA) for research unrelated to this manuscript. Boris at one site and on one processing platform; (3) the high-dimensionality Gutman receives financial support for consulting at Natera, Inc. (San of shape analysis, allowing us to investigate local differences in subcor- Carlos, CA, USA) in research unrelated to the manuscript. Neda Jahan- tical structures, rather than global volume; (4) in comparison to voxel shad received partial grant support from Biogen, Inc. (Boston, USA) based morphometry (VBM), shape analysis can distinguish between for research unrelated to this manuscript. Paul M. Thompson received thickness and surface area; all analyses are performed in native space partial grant support from Biogen, Inc. (Boston, USA) for research and abnormalities can be localized at finer scale, especially in the unrelated to this manuscript. Jose M. Menchon has received research boundaries between structures. funding, consultation or lecture fees from Janssen, AbBiotics, Exeltis, Several limitations should however be emphasized. First, clinical and Medtronic in the last 24 months unrelated to this manuscript. information on comorbid depression or anxiety was not available for some of the sites; therefore, the sample selection for the relevant DATA AVAILABILITY STATEMENT subanalyses differed from the whole group comparison. Second, the The data that support the findings of this study are available on request imaging data that were utilized here were all from 1.5T scanners, which from the corresponding author. The data are not publicly available due may have lower signal-to-noise ratio in certain cases and thus lower to privacy or ethical restrictions. sensitivity to detect group differences compared to 3T data. Third, OCD is a heterogenous disorder, and this study was not sufficiently ORCID powered to explore variations across different subtypes of OCD Jean-Paul Fouche https://orcid.org/0000-0002-0830-2324 (Mataix-Cols, do Rosario-Campos & Leckman, 2005). Fourth, the use Christine Lochner https://orcid.org/0000-0002-4766-3704 of cross-sectional comparisons limits our ability to infer causality Christopher R. K. Ching https://orcid.org/0000-0003-2921-3408 regarding the associations found here or to detect age-by-diagnosis Jun Soo Kwon https://orcid.org/0000-0002-1060-1462 findings. Notably, heritability estimates of shape of subcortical struc- Takashi Nakamae https://orcid.org/0000-0003-4265-198X tures are lower in older populations compared to younger populations Jesus Pujol https://orcid.org/0000-0002-9946-4547 (Roshchupkin et al., 2016), suggesting accumulative effects of environ- Carles Soriano-Mas https://orcid.org/0000-0003-4574-6597 mental factors on brain shape over time. We found no evidence that shape alterations were more pronounced in younger OCD patients, PEER REVIEW although the scope of this study was limited to adults. Future longi- The peer review history for this article is available at https://publons. tudinal studies may help to illuminate the neurodevelopmental and com/publon/10.1002/brb3.2755. neurodegenerative aspects of alterations in subcortical shape in OCD, including changes in shape with age across the lifespan. REFERENCES In summary, we found that comorbidity and medication use were Bas-Hoogendam, J. M., van Steenbergen, H., Pannekoek, J. N., Fouche, J. P., associated with alterations in shape thickness and surface area in sub- Lochner, C., Hattingh, C. J., & van der Wee, N. J. (2017). Voxel-based mor- phometry multi-center mega-analysis of brain structure in social anxiety cortical structures such as the hippocampus, pallidum, and thalamus. disorder. NeuroImage: Clinical, 16, 678–688. https://doi.org/10.1016/j. This is partly consistent with previous work on subcortical shape and ebiom.2018.08.048 volumes in OCD. Furthermore, our findings advance previous work by Boedhoe, P. S., Schmaal, L., Abe, Y., Ameis, S. H., Arnold, P. D., Batistuzzo, suggesting that differences in the shape of subcortical regions may M. C., & van den Heuvel, O. A. (2017). Distinct subcortical volume alter- ations in pediatric and adult OCD: A worldwide meta-and mega-analysis. occur mainly in OCD patients with comorbid depression, anxiety, and American Journal of Psychiatry, 174(1), 60–69. https://doi.org/10.1176/ medication use. Further work is needed to delineate the exact mech- appi.ajp.2016.16020201 anisms contributing to subcortical shape across the spectrum of OCD Bruin, W. B., Taylor, L., Thomas, R. M., Shock, J. P., Zhutovsky, P., Abe, patients. Y., & van Wingen, G. A. (2020). Structural neuroimaging biomarkers for obsessive-compulsive disorder in the ENIGMA-OCD consortium: Medication matters. Translational Psychiatry, 10(1), 1–12. ACKNOWLEDGMENTS De Wit, S. J., Alonso, P., Schweren, L., Mataix-Cols, D., Lochner, C., Menchón, The authors of this study were supported by FIS PI18/00856 grant, J. M., & Heuvel, O. A. (2014). Multicenter voxel-based morphome- NIA T32AG058507, NIH grant U54 EB 020403, RO1MH117601, try mega-analysis of structural brain scans in obsessive-compulsive FOUCHE ET AL . 11 of 12 disorder. American Journal of Psychiatry, 171(3), 340–349. https://doi. disorder. European Archives of Psychiatry and Clinical Neuroscience, 257(4), org/10.1176/appi.ajp.2013.13040574 185–190. https://doi.org/10.1007/s00406-006-0655-5 Fanselow, M. S., & Dong, H. W. (2010). Are the dorsal and ventral hippocam- Izquierdo, I., & Medina, J. H. (1997). Memory formation: The sequence of pus functionally distinct structures? Neuron, 65(1), 7–19. https://doi.org/ biochemical events in the hippocampus and its connection to activity in 10.1016/j.neuron.2009.11.031 other brain structures. Neurobiology of Learning and Memory, 68(3), 285– Fischl, B., Salat, D. H., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., & 316. https://doi.org/10.1006/nlme.1997.3799 Dale, A. M. (2002). Whole brain segmentation: Automated labeling of Kassem, M. S., Lagopoulos, J., Stait-Gardner, T., Price, W. S., Chohan, T. neuroanatomical structures in the human brain. Neuron, 33(3), 341–355. W., Arnold, J. C., & Bennett, M. R. (2013). Stress-induced grey mat- https://doi.org/10.1016/S0896-6273(02)00569-X ter loss determined by MRI is primarily due to loss of dendrites and Fouche, J. P., du Plessis, S., Hattingh, C., Roos, A., Lochner, C., Soriano-Mas, their synapses. Molecular neurobiology, 47(2), 645–661. https://doi.org/ C., & van den Heuvel, O. A. (2017). Cortical thickness in obsessive– 10.1007/s12035-012-8365-7 compulsive disorder: Multisite mega-analysis of 780 brain scans from six Kong, X. Z., Boedhoe, P. S., Abe, Y., Alonso, P., Ameis, S. H., Arnold, P. centres. The British Journal of Psychiatry, 210(1), 67–74. https://doi.org/ D., & Francks, C. (2020). Mapping cortical and subcortical asymmetry 10.1192/bjp.bp.115.164020 in obsessive-compulsive disorder: Findings from the ENIGMA Consor- Gilbert, A. R., Moore, G. J., Keshavan, M. S., Paulson, L. A. D., Narula, V., Mac tium. Biological Psychiatry, 87(12), 1022–1034. https://doi.org/10.1016/ Master, F. P., & Rosenberg, D. R. (2000). Decrease in thalamic volumes j.biopsych.2019.04.022 of pediatric patients with obsessive-compulsive disorder who are tak- Kubota, Y., Sato, W., Kochiyama, T., Uono, S., Yoshimura, S., Sawada, R., & ing paroxetine. Archives of General Psychiatry, 57(5), 449–456. https://doi. Toichi, M. (2016). Putamen volume correlates with obsessive compulsive org/10.1001/archpsyc.57.5.449 characteristics in healthy population. Psychiatry Research: Neuroimaging, Gilbert, A. R., Keshavan, M. S., Diwadkar, V., Nutche, J., MacMaster, F., 249, 97–104. https://doi.org/10.1016/j.pscychresns.2016.01.014 Easter, P. C., & Rosenberg, D. R. (2008). Gray matter differences between Langers, D. R., Jansen, J. F., & Backes, W. H. (2007). Enhanced signal pediatric obsessive-compulsive disorder patients and high-risk siblings: detection in neuroimaging by means of regional control of the global A preliminary voxel-based morphometry study. Neuroscience Letters, false discovery rate. Neuroimage, 38(1), 43–56. https://doi.org/10.1016/ 435(1), 45–50. https://doi.org/10.1016/j.neulet.2008.02.011 j.neuroimage.2007.07.031 Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. Logue, M. W., van Rooij, S. J., Dennis, E. L., Davis, S. L., Hayes, J. P., L., Hill, C. L., & Charney, D. S. (1989). The Yale-Brown obsessive com- Stevens, J. S., & Morey, R. A. (2018). Smaller hippocampal volume in post- pulsive scale: I. Development, use, and reliability. Archives of General traumatic stress disorder: A multisite ENIGMA-PGC study: Subcortical Psychiatry, 46(11), 1006–1011. https://doi.org/10.1001/archpsyc.1989. volumetry results from posttraumatic stress disorder consortia. Biologi- 01810110048007 cal psychiatry, 83(3), 244–253. https://doi.org/10.1016/j.biopsych.2017. Gutman, B. A., Wang, Y., Rajagopalan, P., Toga, A. W., & Thompson, P. M. 09.006 (2012)May. Shape matching with medial curves and 1-D group-wise reg- Ludolph, A. G., Pinkhardt, E. H., Tebartz van Elst, L., Libal, G., Ludolph, A. C., istration. In 2012 9th IEEE International Symposium on Biomedical Imaging Fegert, J. M., & Kassubek, J. (2008). Are amygdalar volume alterations (ISBI) (716–719) IEEE. https://doi.org/10.1109/ISBI.2012.6235648 in children with Tourette syndrome due to ADHD comorbidity? Devel- Gutman, B. A., Madsen, S. K., Toga, A. W., & Thompson, P. M. (2013a). A opmental Medicine & Child Neurology, 50(7), 524–529. https://doi.org/10. family of fast spherical registration algorithms for cortical shapes. In 1111/j.1469-8749.2008.03014.x International Workshop on Multimodal Brain Image Analysis (pp. 246–257). Lu, Y., Liang, H., Han, D., Mo, Y., Li, Z., Cheng, Y., & Sun, X. (2016). The volu- Springer. https://doi.org/10.1007/978-3-319-02126-3_24 metric and shape changes of the putamen and thalamus in first episode, Gutman, B. A., Hua, X., Rajagopalan, P., Chou, Y. Y., Wang, Y., Yanovsky, untreated major depressive disorder. NeuroImage: Clinical, 11, 658–666. I., Thompson, P. M., & Alzheimer’s Disease Neuroimaging Initiative. https://doi.org/10.1016/j.nicl.2016.04.008 (2013b). ). Maximizing power to track Alzheimer’s disease and MCI pro- Mataix-Cols, D., do Rosario-Campos, M. C., & Leckman, J. F. (2005). A multi- gression by LDA-based weighting of longitudinal ventricular surface fea- dimensional model of obsessive-compulsive disorder. American Journal of tures. Neuroimage, 70, 386–401. https://doi.org/10.1016/j.neuroimage. Psychiatry, 162(2), 228–238. https://doi.org/10.1176/appi.ajp.162.2.228 2012.12.052 Milad, M. R., & Rauch, S. L. (2012). Obsessive-compulsive disorder: Beyond Haber, S. N. (2016). Corticostriatal circuitry. Dialogues in Clinical Neuro- segregated cortico-striatal pathways. Trends in cognitive sciences, 16(1), science, 18(1), 7–21. 10.31887/DCNS.2016.18.1/shaber 43–51. https://doi.org/10.1016/j.tics.2011.11.003 Haber, S. N., & Knutson, B. (2010). The reward circuit: Linking primate Mosher, C. P., Zimmerman, P. E., & Gothard, K. M. (2010). Response char- anatomy and human imaging. Neuropsychopharmacology, 35(1), 4–26. acteristics of basolateral and centromedial neurons in the primate https://doi.org/10.1038/npp.2009.129 amygdala. Journal of Neuroscience, 30(48), 16197–16207. https://doi.org/ Hibar, D. P., Westlye, L. T., Doan, N. T., Jahanshad, N., Cheung, J. W., Ching, C. 10.1523/JNEUROSCI.3225-10.201 R., & Andreassen, O. A. (2018). Cortical abnormalities in bipolar disorder: Peng, Z., Lui, S. S., Cheung, E. F., Jin, Z., Miao, G., Jing, J., & Chan, R. C. (2012). An MRI analysis of 6503 individuals from the ENIGMA Bipolar Disorder Brain structural abnormalities in obsessive-compulsive disorder: Con- Working Group. Molecular psychiatry, 23(4), 932–942. https://doi.org/10. verging evidence from white matter and grey matter. Asian Journal of 1038/mp.2017.73 Psychiatry, 5(4), 290–296. https://doi.org/10.1016/j.ajp.2012.07.004 Hoexter, M. Q., Luis de Souza Duran, F., Chaubet D’Alcante, C., Dougherty, Peterson, B. S., Choi, H. A., Hao, X., Amat, J. A., Zhu, H., Whiteman, R., & D. D., Shavitt, R. G., Lopes, A. C., & Busatto, G. F. (2012). Gray matter Bansal, R. (2007). Morphologic features of the amygdala and hippocam- volumes in obsessive-compulsive disorder before and after fluoxetine pus in children and adults with Tourette syndrome. Archives of General or cognitive-behavior therapy: A randomized clinical trial. Neuropsy- Psychiatry, 64(11), 1281–1291. https://doi.org/10.1001/archpsyc.64.11. chopharmacology, 37, 734–745. 1281 Hollander, E., Stein, D. J., Fineberg, N. A., Marteau, F., & Legault, M. (2010). Piras, F., Piras, F., Chiapponi, C., Girardi, P., Caltagirone, C., & Spalletta, Quality of life outcomes in patients with obsessive-compulsive disorder: G. (2015). Widespread structural brain changes in OCD: A systematic Relationship to treatment response and symptom relapse. The Jour- review of voxel-based morphometry studies. Cortex; A Journal Devoted to nal of Clinical Psychiatry, 71(6), 784–792. https://doi.org/10.4088/jcp. the Study of the Nervous System and Behavior, 62, 89–108. https://doi.org/ 09m05911blu 10.1016/j.cortex.2013.01.016 Hong, S. B., Shin, Y. W., Kim, S. H., Yoo, S. Y., Lee, J. M., Kim, I. Y., & Kwon, J. S. Pujol, J., Soriano-Mas, C., Alonso, P., Cardoner, N., Menchón, J. M., (2007). Hippocampal shape deformity analysis in obsessive–compulsive Deus, J., & Vallejo, J. (2004). Mapping structural brain alterations in 12 of 12 FOUCHE ET AL . obsessive-compulsive disorder. Archives of General Psychiatry, 61(7), van den Heuvel, O. A., van Wingen, G., Soriano-Mas, C., Alonso, P., 720–730. https://doi.org/10.1176/appi.ajp.2008.08010033 Chamberlain, S. R., Nakamae, T., & Veltman, D. J. (2016). Brain circuitry Radua, J., & Mataix-Cols, D. (2009). Voxel-wise meta-analysis of grey matter of compulsivity. European Neuropsychopharmacology, 26(5), 810–827. changes in obsessive–compulsive disorder. The British Journal of Psychia- https://doi.org/10.1016/j.euroneuro.2015.12.005 try, 195(5), 393–402. https://doi.org/10.1192/bjp.bp.108.055046 van Erp, T. G., Hibar, D. P., Rasmussen, J. M., Glahn, D. C., Pearlson, G. Radua, J., van den Heuvel, O. A., Surguladze, S., & Mataix-Cols, D. D., Andreassen, O. A., & Turner, J. A. (2016). Subcortical brain vol- (2010). Meta-analytical comparison of voxel-based morphometry ume abnormalities in 2028 individuals with schizophrenia and 2540 studies in obsessive-compulsive disorder vs other anxiety disor- healthy controls via the ENIGMA consortium. Molecular psychiatry, 21(4), ders. Archives of General Psychiatry, 67(7), 701–711. https://doi.org/ 547–553. https://doi.org/10.1038/mp.2015.63 10.1001/archgenpsychiatry.2010.70 Wang, Y., Song, Y., Rajagopalan, P., An, T., Liu, K., Chou, Y. Y., Thompson, P. M., Rasmussen, S. A., & Eisen, J. L. (1992). The epidemiology and differen- . . . Alzheimer’s Disease Neuroimaging Initiative. (2011). Surface-based tial diagnosis of obsessive-compulsive disorder. In Zwangsstörungen/ TBM boosts power to detect disease effects on the brain: An N = 804 obsessive-compulsive disorders (pp. 1–14). Springer. https://doi.org/10. ADNI study. Neuroimage, 56(4), 1993–2010. https://doi.org/10.1016/j. 1007/978-3-642-77608-3_1 neuroimage.2011.03.040 Roshchupkin, G. V., Gutman, B. A., Vernooij, M. W., Jahanshad, N., Martin, Werner, C. J., Stöcker, T., Kellermann, T., Wegener, H. P., Schneider, F., Shah, N. G., Hofman, A., & Adams, H. H. H. (2016). Heritability of the shape of N. J., & Neuner, I. (2010). Altered amygdala functional connectivity in subcortical brain structures in the general population. Nature Communi- adult Tourette’s syndrome. European Archives of Psychiatry and Clinical cations, 7(1), 1–8. https://doi.org/10.1038/ncomms13738 Neuroscience, 260(2), 95–99. Rotge, J. Y., Guehl, D., Dilharreguy, B., Tignol, J., Bioulac, B., Allard, Zarei, M., Mataix-Cols, D., Heyman, I., Hough, M., Doherty, J., Burge, M., & Aouizerate, B. (2009). Meta-analysis of brain volume changes L., & James, A. (2011). Changes in gray matter volume and white in obsessive-compulsive disorder. Biological Psychiatry, 65(1), 75–83. matter microstructure in adolescents with obsessive-compulsive disor- https://doi.org/10.1016/j.biopsych.2008.06.019 der. Biological psychiatry, 70(11), 1083–1090. https://doi.org/10.1016/j. Rotge, J. Y., Langbour, N., Guehl, D., Bioulac, B., Jaafari, N., Allard, M., & biopsych.2011.06.032 Burbaud, P. (2010). Gray matter alterations in obsessive–compulsive Zhang, L., Hu, X., Lu, L., Li, B., Hu, X., Bu, X., Li, H., Tang, S., Yang, Y., disorder: An anatomic likelihood estimation meta-analysis. Neuropsy- Roberts, N., Sweeney, J. A., Gon, Q., & Huang, X. (2019a). Abnormal- chopharmacology, 35(3), 686–691. https://doi.org/10.1038/npp.2009. ities of hippocampal shape and subfield volumes in medication-free 175 patients with obsessive–compulsive disorder. Human Brain Mapping, Ruscio, A. M., Stein, D. J., Chiu, W. T., & Kessler, R. C. (2010). The epidemi- 40(14), 4105–4113. https://doi.org/10.1002/hbm.24688 ology of obsessive-compulsive disorder in the National Comorbidity Zhang, L., Hu, X., Li, H., Lu, L., Li, B., Hu, X., & Huang, X. (2019b). Charac- Survey Replication. Molecular Psychiatry, 15(1), 53–63. https://doi.org/ teristic alteration of subcortical nuclei shape in medication-free patients 10.1038/mp.2008.94 with obsessive-compulsive disorder. NeuroImage: Clinical, 24, 102040. Sah, P., Faber, E. L., Lopez de Armentia, M., & Power, J. M. J. P. R. (2003). https://doi.org/10.1016/j.nicl.2019.102040 The amygdaloid complex: Anatomy and physiology. Physiological Reviews, Zhao, K., Liu, H., Yan, R., Hua, L., Chen, Y., Shi, J., & Yao, Z. (2017). Corti- 83(3), 803–834. https://doi.org/10.1152/physrev.00002.2003 cal thickness and subcortical structure volume abnormalities in patients Schmaal, L., Veltman, D. J., van Erp, T. G., Sämann, P. G., Frodl, T., Jahanshad, with major depression with and without anxious symptoms. Brain and N., & Hibar, D. P. (2016). Subcortical brain alterations in major depressive Behavior, 7(8), e00754. https://doi.org/10.1002/brb3.754 disorder: Findings from the ENIGMA Major Depressive Disorder work- Zugman, A., Harrewijn, A., Cardinale, E., Zwiebel, H., Freitag, G. F., Werwath, ing group. Molecular Psychiatry, 21(6), 806–812. https://doi.org/10.1038/ K. E., & Winkler, A. M. (2020). Mega-analysis methods in ENIGMA: The mp.2015.69 experience of the generalized anxiety disorder working group. https:// Shaw, P., Sharp, W., Sudre, G., Wharton, A., Greenstein, D., Raznahan, A., doi.org/10.1002/hbm.25096 & Rapoport, J. (2015). Subcortical and cortical morphological anoma- lies as an endophenotype in obsessive-compulsive disorder. Molecular Psychiatry, 20(2), 224–231. https://doi.org/10.1038/mp.2014.3 Shi, Y., Lai, R., Wang, D. J., Pelletier, D., Mohr, D., Sicotte, N., & Toga, How to cite this article: Fouche, J.-P., Groenewold, N. A., A. W. (2014). Metric optimization for surface analysis in the Laplace- Sevenoaks, T., Heany, S., Lochner, C., Alonso, P., Batistuzzo, M. Beltrami embedding space. IEEE Transactions on Medical Imaging, 33(7), 1447–1463. https://doi.org/10.1109/TMI.2014.2313812 C., Cardoner, N., Ching, C. R. K., de Wit, S. J., Gutman, B., Szeszko, P. R., MacMillan, S., McMeniman, M., Lorch, E., Madden, R., Ivey, Hoexter, M. Q., Jahanshad, N., Kim, M., Kwon, J. S., J., & Rosenberg, D. R. (2004). Amygdala volume reductions in pediatric Mataix-Cols, D., Menchon, J. M., Miguel, E. C., Nakamae, T., . . . patients with obsessive–compulsive disorder treated with paroxetine: Stein, D. J. (2022). Shape analysis of subcortical structures in Preliminary findings. Neuropsychopharmacology, 29(4), 826–832. Szeszko, P. R., Christian, C., MacMaster, F., Lencz, T., Mirza, Y., Taormina, S. obsessive-compulsive disorder and the relationship with P., & Rosenberg, D. R. (2008). Gray matter structural alterations in psy- comorbid anxiety, depression, and medication use: A chotropic drug-naive pediatric obsessive-compulsive disorder: An opti- meta-analysis by the OCD Brain Imaging Consortium. Brain mized voxel-based morphometry study. American Journal of Psychiatry, and Behavior, 12, e2755. https://doi.org/10.1002/brb3.2755 165(10), 1299–1307.