Thompson et al. Translational Psychiatry (2020)10:100 https://doi.org/10.1038/s41398-020-0705-1 Translational Psychiatry REVIEW ARTICLE Open Access ENIGMA and global neuroscience: A decade of large-scale studies of the brain in health and disease across more than 40 countries Abstract This review summarizes the last decade of work by the ENIGMA (Enhancing NeuroImaging Genetics through Meta Analysis) Consortium, a global alliance of over 1400 scientists across 43 countries, studying the human brain in health and disease. Building on large-scale genetic studies that discovered the first robustly replicated genetic loci associated with brain metrics, ENIGMA has diversified into over 50 working groups (WGs), pooling worldwide data and expertise to answer fundamental questions in neuroscience, psychiatry, neurology, and genetics. Most ENIGMA WGs focus on specific psychiatric and neurological conditions, other WGs study normal variation due to sex and gender differences, or development and aging; still other WGs develop methodological pipelines and tools to facilitate harmonized analyses of “big data” (i.e., genetic and epigenetic data, multimodal MRI, and electroencephalography data). These international efforts have yielded the largest neuroimaging studies to date in schizophrenia, bipolar disorder, major depressive disorder, post-traumatic stress disorder, substance use disorders, obsessive-compulsive disorder, attention- deficit/hyperactivity disorder, autism spectrum disorders, epilepsy, and 22q11.2 deletion syndrome. More recent 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; ENIGMA WGs have formed to study anxiety disorders, suicidal thoughts and behavior, sleep and insomnia, eating disorders, irritability, brain injury, antisocial personality and conduct disorder, and dissociative identity disorder. Here, we summarize the first decade of ENIGMA’s activities and ongoing projects, and describe the successes and challenges encountered along the way. We highlight the advantages of collaborative large-scale coordinated data analyses for testing reproducibility and robustness of findings, offering the opportunity to identify brain systems involved in clinical syndromes across diverse samples and associated genetic, environmental, demographic, cognitive, and psychosocial factors. Introduction shows a world map of participating sites, broken down by The ENIGMA (Enhancing NeuroImaging Genetics working group). Thirty of the ENIGMA WGs focus on through Meta Analysis) Consortium is a collaboration of specific psychiatric and neurologic conditions. Four study more than 1400 scientists from 43 countries studying the different aspects of development and aging. Others study human brain. ENIGMA started 10 years ago, in 2009, with key transdiagnostic constructs, such as irritability, and the the initial aim of performing a large-scale neuroimaging importance of evolutionarily interesting genomic regions genetic study, and has since diversified into 50 working in shaping human brain structure and function. Central to groups (WGs), pooling worldwide data, resources and the success of these WGs are the efforts of dedicated expertise to answer fundamental questions in neu- methods development groups within ENIGMA. There are roscience, psychiatry, neurology, and genetics (Fig. 1 currently 12 WGs that develop and disseminate multi- scale and ‘big data’ analysis pipelines to facilitate harmo- nized analyses using genetic and epigenetic data, Correspondence: Paul M. Thompson (pthomp@usc.edu) multimodal (anatomical, diffusion, functional) magnetic Full list of author information is available at the end of the article. © The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Thompson et al. Translational Psychiatry (2020)10:100 Page 2 of 28 Fig. 1 World Map of ENIGMA’s Working Groups. The ENIGMA Consortium has grown to include over 1400 participating scientists from over 200 institutions, across 43 countries worldwide. ENIGMA is organized as a set of 50 WGs, studying 26 major brain diseases (see color key). Each group works closely with the others and consists of worldwide teams of experts in each brain disorder as well as experts in the major methods used to study each disorder. The diseases studied include major depressive disorder, bipolar disorder, schizophrenia, substance use disorder, post-traumatic stress disorder, attention-deficit/hyperactivity disorder, obsessive-compulsive disorder, and autism spectrum disorder, and several neurological disorders, including Parkinson’s disease, epilepsy, ataxia, and stroke. In recent years, new WGs were created that grew into worldwide consortia on epilepsy (Whelan et al.9), eating disorders (King et al.104), anxiety disorders (Groenewold et al.107), antisocial behavior, and infant neuroimaging. resonance imaging (MRI) and spectroscopy (MRS) mea- studies are being completed, ENIGMA is beginning to sures, in combination with genetic and epigenetic data, identify shared and distinct neuroimaging patterns in and data from electroencephalography (EEG). brain disorders with known genetic or clinical over- The Consortium has been a formidable force for dis- lap11,12, and to delineate the role of transdiagnostic risk covery and innovation in human brain imaging, sup- factors (e.g., childhood trauma) and clinical phenomena porting more than 200 active studies. The disorder- (e.g., suicidal thoughts and behaviors). In addition, specific WGs have published the largest neuroimaging ENIGMA’s genetic studies are now analyzing imaging and studies to date in schizophrenia (SCZ; total N = 9572; genetics data from more than 50,000 people to uncover 4474 cases)1, bipolar disorder (BD; total N = 6503; 2447 genetic markers that most robustly associated with brain cases)2, major depressive disorder (MDD; total N = structure and function, or imaging derived neurobiologi- 10,105; 2148 cases)3, post-traumatic stress disorder cal traits related to various disease conditions13–16. (PTSD; total N = 1868; 794 cases)4, substance use dis- As we detail in this review, the ENIGMA Consortium orders (SUD; total N = 3240; 2140 cases)5, obsessive- has made multiple, seminal contributions to neuroscience compulsive disorder (OCD; total N = 3665; 1905 cases)6, and psychiatry, including (a) characterization of robust attention-deficit/hyperactivity disorder (ADHD; total N = neuroimaging profiles for various brain disorders, (b) 4180; 2246 cases)7, autism spectrum disorders (ASD; total standardization of metrics used to assess clinical symp- N = 3222; 1571 cases)8, epilepsy (N = total 3876; 2149 toms of patients across multiple research sites, and (c) use cases)9, and 22q11.2 deletion syndrome (22q11DS; total of dimensional approaches that go beyond the N = 944; 474 cases)10. Key results of these studies are case–control comparisons of individuals with categorical summarized in Table 1. Building on this work, the focus diagnoses, and further enable the investigation of specific of the ENIGMA disorder-specific WGs now goes beyond genetic, and environmental features or neurobiological traditional diagnostic boundaries. As these first large-scale markers associated with disorder risk and treatment Thompson et al. Translational Psychiatry (2020)10:100 Page 3 of 28 Table 1 A Selection of key findings from ENIGMA’s Working Groups, along with key papers and current sample sizes. Working Number of Total N Age range Relevant publications Main findings group datasets (patient N) (in years) Clinical 22Q11DS 14 863 (533) 6–56 Villalón-Reina et al.17; Sun et al.10 Widespread reductions in diffusivity, pronounced in regions with major cortico- cortical and cortico-thalamic fibers; thicker cortical gray matter overall, but focal thickness reduction in temporal and cingulate cortex; cortical surface area showed pervasive reductions; lower cortical surface area in individuals with larger microdeletion; 22q- related psychosis associated with lower cortical thickness and significantly overlapped with findings from ENIGMA-SCZ group. 5,84 86 Addiction/ 118 18,823 (6,592) 7–68 Mackey et al. ; Conrod et al. Common neural substrate shared in SUDs dependence; differential patterns of regional volume as biomarkers of dependence on alcohol and nicotine; lower volume or thickness observed, with greatest effects associated with alcohol use disorder; insula and medial orbitofrontal cortex affected, regardless of dependence. 7,91 47 ADHD 37 4180 (2246) 4–63 Hoogman et al. ; Klein et al. ; Reduction in bilateral amygdala, striatal, and Zhang-James94; Hess et al.92 hippocampal volumes in the ADHD population, especially in children; lower cortical surface area values found in children with ADHD, but not in adolescents or adults; lower surface area associated with ADHD symptoms in the general population in childhood; genetic association studies suggest that genes involved in neurite outgrowth play a role in findings of reduced volume in ADHD; gene-expression studies imply that structural brain alterations in ADHD can also be explained in part by the differential vulnerability of these regions to mechanisms mediating apoptosis, oxidative stress, and autophagy. 97 8 ASD 54 3583 (1774) 2–64 Postema et al. ; van Rooij et al. Altered morphometry in the cognitive and affective parts of the striatum, frontal cortex and temporal cortex in ASD. 69 23 BD 44 11,100 (3100) 8–86 Favre et al. ; Nunes et al. ; Volumetric reductions in hippocampus and Hibar et al.2,68 thalamus and enlarged lateral ventricles in patients; thinner cortical gray matter in bilateral frontal, temporal and parietal regions; strongest effects on left pars opercularis, fusiform gyrus and rostral middle frontal cortex in BD. Thompson et al. Translational Psychiatry (2020)10:100 Page 4 of 28 Table 1 continued Working Number of Total N Age range Relevant publications Main findings group datasets (patient N) (in years) Eating 28 anorexia 2531 (897 AN; 10–50 AN; Walton et al.48 Signs of inverse concordance between greater Disorders nervosa (AN); 307 BN) 12–46 BN thalamus volume and risk for anorexia nervosa 12 bulimia (AN); variation in gene DRD2 significantly nervosa (BN) associated with AN only after conditioning on its association with caudate volume; genetic variant linked to LRRC4C reached significance after conditioning on hippocampal volume. Epilepsy 24 3876 (2149) 18–55 Whelan et al.9 Patients with IGE showed volume reductions in the right thalamus and lower thickness in the bilateral precentral gyri; both MTLE subgroups showed volume reductions in the ipsilateral hippocampus, and lower thickness in extrahippocampal cortical regions, including the precentral and paracentral gyri; lower subcortical volume and cortical thickness were associated with a longer duration of epilepsy in the all-epilepsies and right MTLE groups. 124,169,170 171 HIV 12 1044 (all 22–81 Nir et al. ; Fouche et al. In the full group, subcortical volume patients) associations implicated the limbic system: lower current CD4+ counts were associated with smaller hippocampal and thalamic volumes; a detectable viral load was associated with smaller hippocampal and amygdala volumes; limbic effects were largely driven by participants on cART; in subset of participants not on cART, smaller putamen volumes were associated with lower CD4+ count. 67 75 MDD 38 14,249 (4379) 10–89 van Velzen et al. ; Tozzi et al. ; Han Significantly lower hippocampal volumes; et al.72; Frodl et al.74; Renteria et al.172; thinner orbitofrontal cortex, anterior and Schmaal et al.3,70; Ho et al.137; posterior cingulate, insula and temporal lobes Saemann et al.83 cortex in adult MDD patients; lower total surface area and regional reductions in frontal regions and primary and higher-order visual, somatosensory and motor areas in adoloescent MDD patients; greater exposure to childhood adversity associated with smaller caudate volumes in females, independent of MDD; patients reporting suicidal plans or attempts showed a smaller ICV volume compared to controls. 6,88,167 45 OCD 38 3665 (1905) 5–65 Boedhoe et al. ; Hibar et al. Subcortical abnormalities in pediatric and adult patients; pallidum (bigger) and hippocampus (smaller) key in adults, and thalamus (bigger) key in (unmedicated) pediatric group; parietal cortex consistently Thompson et al. Translational Psychiatry (2020)10:100 Page 5 of 28 Table 1 continued Working Number of Total N Age range Relevant publications Main findings group datasets (patient N) (in years) implicated both in children and adults; more widespread cortical thickness abnormalities in medicated adults, and more pronounced surface area deficits (mainly in frontal regions) in medicated pediatric OCD patients. PTSD 16 3118 (1288) 17–85 Dennis et al.76; Salminen et al.80; Significantly smaller hippocampi, on average, Logue et al.4; O’Leary et al.78 in individuals with current PTSD compared with trauma-exposed control subjects, and smaller amygdalae. Schizophrenia 39 9572 (4474) 18–77 Holleran et al.57; van Erp et al.1,54,55; Positive symptom severity was negatively Kelly et al.56; Walton et al.62,63; related to bilateral STG thickness; widespread Kochunov et al.66 thinner cortex and smaller surface area, largest effect sizes in frontal and temporal lobe regions; smaller hippocampus, amygdala, thalamus, accumbens and intracranial volumes; larger pallidum and lateral ventricle volumes; widespread reductions in FA, esp. in anterior corona radiata and corpus callosum; higher mean and radial diffusivity; left MOFC thickness significantly associated with negative symptom severity; link between prefrontal thinning and negative symptom severity in schizophrenia. 100 53 CNV 37 16,889 (24 3–90 van der Meer et al. ; Sonderby 16p11.2 distal CNV: Negative dose-response 16p11.2 distal associations with copy number on intracranial and 125 15q11.2 volume and regional caudate, pallidum and CNV carriers) putamen volumes. 15q11.2 CNV: Decrease in accumbens and cortical surface area in deletion carriers and negative dose response on cortical thickness. Non-clinical EEG 5 8425 5–73 Smit et al.40 Identified several novel genetic variants associated with oscillatory brain activity; replicated and advanced understanding of previously known genes associated with psychopathology (i.e., schizophrenia and alcohol use disorders); these psychopathological liability genes affect brain functioning, linking the genes’ expression to specific cortical/subcortical brain regions. GWAS 34 22,456 3–91 Satizabal et al.14; Grasby et al.13; Over 200 genetic loci where common Hibar et al.25,173; Adams et al.169 variation is associated with cortical thickness or surface area; over 40 common genetic variants associated with subcortical volumes. Laterality 99 17,141 3–90 de Kovel et al.71; Kong et al.90,154; Average patterns of left-right anatomical Postema et al.97; Guadalupe et al.174 asymmetry of the healthy brain were mapped, Thompson et al. Translational Psychiatry (2020)10:100 Page 6 of 28 Table 1 continued Working Number of Total N Age range Relevant publications Main findings group datasets (patient N) (in years) as regards cortical regional surface areas, thicknesses, and subcortical volumes; fronto- occipital gradient in cortical thickness asymmetry was found, with frontal regions generally thicker on the left, and occipital regions on the right; asymmetries of various structural measures were significantly heritable, indicating genetic effects that differ between the two sides; age, sex and intracranial volume affected some asymmetries, but handedness did not; disorder case–control analyses revealed subtle reductions of regional cortical thickness asymmetries in ASD, as well as altered orbitofrontal surface area asymmetry; little evidence for altered anatomical asymmetry was found in MDD; pediatric patients with OCD showed evidence for altered asymmetry of the thalamus and pallidum. Lifespan 91 14,904 healthy 2–92 Dima et al.175; Frangou et al.176 Thickness in almost all cortical regions individuals decreased prominently in the first two to three decades of life, with an attenuated or plateaued slope afterwards; exceptions to this pattern were entorhinal and temporopolar cortices whose thickness showed an attenuated inverse U-shaped relation with age, and anterior cingulate cortex, which showed a U-shaped association with age; age at peak cortical thickness was 6–7 years for most brain regions. Plasticity 36 10,199 (2242) 6–97 Brouwer et al.38,39 Heritability estimates of change rates were generally higher in adults than in children suggesting an increasing influence of genetic factors explaining individual differences in brain structural changes with age; for some structures, the genetic factors influencing change were different from those influencing the volume itself, suggesting the existence of genetic variants specific for brain plasticity. outcome. The large scale and inclusivity of these analyses ability to identify the most robust pattern of non- —in terms of populations, sample sizes, numbers of invasively measured neurobiological features involved in coordinating centers, and diversity of imaging and genetic clinical syndromes across multiple samples that are more data—has been instrumental for demonstrating robust representative of the global population. This also results associations between clinical factors and brain alterations, in robust effect size estimates, without the confounds of and for stratifying patients with the same diagnosis literature-based meta-analyses based on published data according to differential treatment outcomes10,17. Thus, a with possible publication bias (as noted in Kong et al.)18. valuable aspect of the existing ENIGMA studies is the These data also provide a unique opportunity to assess Thompson et al. Translational Psychiatry (2020)10:100 Page 7 of 28 important sources of disease heterogeneity, including key genetic, environmental, demographic, and psychosocial factors. Here, we provide a synopsis of the first decade of ENIGMA’s activities and highlight the successes and challenges encountered along the way. History ENIGMA was launched in December 2009 to help ‘break the logjam’ in genetic studies of the brain. At the time, most neuroimaging genetics studies were assessing historically candidate genetic variations, mostly in very small samples of a few tens to hundreds of participants (e.g., COMT, 5-HTTLPR, BDNF). These studies typically reported ‘candidate gene’ effects that did not replicate when tested in independent cohorts19–21. It became apparent that very large numbers of genetic loci con- tributed to variation in complex neurological or psychia- tric traits, including imaging-derived brain measures— each with a very small effect size—and only a few genetic loci accounted for more than 1% of the variance in any complex brain condition or measure22. Thus, scientists began to recognize the need to pool multiple datasets worldwide to perform better-powered studies of these traits. In response, the ENIGMA Consortium’s initial plan was to merge two ‘big data’ sources—neuroimaging and genetics—with the aim of discovering the impact of genetic factors on brain systems, to determine whether these genetic factors underlie manifestation of disorders within the brain, and to identify diagnostic and prognostic neuroimaging biomarkers. A further goal was to improve on previous literature-based meta-analyses by using har- monized processing and analysis protocols on an unpre- cedented scale. This was the impetus that launched ENIGMA’s early studies. In 2014, the NIH Big Data to Knowledge (BD2K) pro- gram awarded a consortium grant to ENIGMA with seed funding for WGs on nine disorders: SCZ, BD, MDD, Fig. 2 ENIGMA’s Working Group Flowchart. ENIGMA’s working OCD, ADHD, ASD, SUD, 22q11DS, and the effects of the groups are divided into technical groups that work on testing human immunodeficiency virus (HIV) on the brain. This harmonized methods, and clinical groups that study different support led to the largest neuroimaging studies for the disorders and conditions across psychiatry and neurology, as well as some behaviors (e.g., schizotypy and antisocial behaviors). The use of nine targeted disorders, with results reported in over 50 harmonized analysis methods across all the working groups has manuscripts. These initial successes provided the driving enabled cross-disorder comparisons (e.g., in the affective/psychosis force to establish an additional 21 disease WGs (see spectrum of depression to bipolar disorder to schizophrenia), and Working Group chart, Fig. 2). transdiagnostic analyses of risk factors such as childhood trauma Following the model established by the Psychiatric across a number of disorders (such as major depressive disorder (MDD) and post-traumatic stress disorder (PTSD)). Several working Genomics Consortium (PGC), which emphasized har- groups, such as brain trauma and anxiety, consist of several subgroups monization of genomic analysis protocols across sites, the examining subtypes (e.g., panic disorder or social anxiety), and allow ENIGMA Consortium created harmonized protocols to analyses of overlap and differences (e.g., between military and civilian analyze brain structure and function, along with genetic, brain trauma). and clinical data across its WGs. Instead of centralizing data, ENIGMA opted to work as a ‘distributed con- sortium’, asking groups to run standardized protocols was important for the rapid acceptance of the consortium themselves, rather than the approach used in the PGC, in the field, as it made contribution very easy; further, the where data are centralized. At the time, ENIGMA design memoranda of understanding provided the basic Thompson et al. Translational Psychiatry (2020)10:100 Page 8 of 28 guidelines for the trusted collaborative networks to usually include unidentifiable derived imaging metrics develop. In the meantime—with views on data sharing (e.g., hippocampal volume) and demographic or clinical having changed quite considerably—many ENIGMA information (age at scan, sex, diagnostic status, etc.); WGs now also share (derived) individual data, allowing however, this form of analysis may limit participation and for more in-depth analyses. requires individual data transfer agreements with partici- In ENIGMA’s genetic studies, many participating cen- pating sites. We note, because of these required agree- ters use different genotyping chips, so data were first ments with potentially clinically sensitive patient imputed to common genomic references (such as the information, and the project-specific design of the ‘cen- 1000 Genomes reference panel), allowing each partici- tralized’ approaches, ENIGMA does not curate a database pating site to perform the same association tests between for repeated or open access, and each cohort PI approves brain measures and genetic variation at over 10 million of each project for which they contribute data. loci across the genome. Furthermore, the ENIGMA Consortium standardized procedures for the extraction ENIGMA’s genetic studies of brain metrics (such as cortical thickness, cortical sur- Uncovering the genetic basis of brain morphometric face area, and subcortical volume) from raw neuroimaging variation data, implemented consensus protocols for data quality The first demonstration of the value of the ENIGMA control and outlier handling, and pioneered new approach was the identification of genetic loci associated meta-analytic methods for the analysis of aggregated sta- with variation in subcortical volumes including the cau- tistical information (http://enigma.ini.usc.edu/protocols/). date, putamen, and hippocampus (see Fig. 3)14,24,25. These ENIGMA’s meta-analyses estimated the size and precision genome-wide association studies (GWAS) yielded intri- of the effects after pooling evidence from multiple cohorts, guing new leads regarding the genetic architecture of the and they also ranked the neuroimaging effect sizes of human brain that were only possible because ENIGMA findings emerging from case–control comparisons, afforded increased power to detect subtle effects. More thereby setting the stage for deeper, secondary analyses recently, ENIGMA identified more than 200 individual aiming to explore potential moderators of psychiatric and loci that significantly contribute to variation in brain neurological disease. More recently, many ENIGMA measures, with p-values reaching 10−180; each single locus groups have moved beyond cohort level meta-analyses to accounted for only 0.1–1% of phenotypic variance, but up pooled, or ‘mega’-analyses (Using brain volumetric data to 20% of the variance in aggregate. For this effort from ENIGMA’s OCD, ADHD, and ASD working groups, ENIGMA had partnered with the CHARGE Consortium Boedhoe et al.12 compared meta-analysis to mega-analyses and UK Biobank on a series of studies of 70 cortical that model site or cohort effects as random effects, measures, including regional cortical thickness and sur- showing broad agreement. Mega-analyses allow more face area13. These discoveries resulted in an annotated sophisticated statistical adjustments as they pool more atlas of common genetic variants that contribute to information across cohorts; meta-analyses tend to be more shaping the human cerebral cortex. Of particular interest, efficient when ethical, legal or logistic constraints govern we found that genetic loci affecting brain morphology or restrict individual-level data transfer (e.g., genome-wide show enrichment for developmentally regulated genes13 genetic data).), where anonymized and unidentifiable and human-specific regulatory elements26,27. Ongoing individual-level data are aggregated in a central location, efforts are beginning to map these genetic effects at a allowing more flexible statistical designs, such as machine finer-grained spatial resolution using shape analysis, sur- learning analyses23, reliable estimation of interaction face- and voxel-based analyses28–31. Moving beyond the effects, and examination of polygenic risk scores. The type mass univariate methods, which analyze each brain and amount of data transferred for each analysis is chosen measure separately, ENIGMA has begun to use multi- pragmatically for each study. Distributed analyses promote variate methods to meet the challenge of quantifying the scientific engagement from many groups worldwide and complex relationships between brain networks—or ‘con- take advantage of distributed computing resources that nectomes’—and the genome32–34. scale up as the network grows; here the data transferred is Current ENIGMA sample sizes (which now exceed mainly aggregate measures such as quality control metrics 50,000) are sufficiently large to identify genetic associa- and the statistical metrics derived from agreed-upon tions at a pace comparable to that of GWAS for other analytical tests. On the other hand, the centralized ana- phenotypes. In a recent analysis, Holland35 contrasted lyses are preferable when a variable of interest is sparsely rates of discovery of genetic loci by ENIGMA and the distributed across sites, (e.g., individuals with 22q11DS PGC and noted the distribution of effect sizes for some exhibiting psychotic symptoms) or when a specific method brain measures (e.g., putamen volume) may indeed be is being developed, and computational power or expertise enriched for slightly larger effects compared to behavioral is available at only a few sites; here the data transferred traits (see also Le and Stein36 and Franke et al.37). Still, a Thompson et al. Translational Psychiatry (2020)10:100 Page 9 of 28 Fig. 3 Genetic Influences on brain structure: effects of common and rare genetic variants. ENIGMA’s large-scale genetic analyses study the effects of both common and rare genetic variants on brain measures. a A series of progressively larger genome-wide association studies have revealed over 45 genetic loci associated with subcortical structure volumes (Hibar et al.25, Satizabal et al.14) and over 200 genetic loci associated with cortical thickness and surface area Grasby et al.13. The Manhattan plots here (adapted from Hibar et al.25, show the genome (on the x-axis) and the evidence for association (as a logarithm of the p-value, on the y-axis) for each common genetic variant (or SNP) with the volume of each brain structure shown. b Genetics of Hippocampal Volume. A subsequent genome-wide association study (GWAS) of 33,536 individuals discovered six independent loci significantly associated with hippocampal volume, four of them novel. Of the novel loci, two lie within key genes involved in neuronal migration and microtubule assembly (ASTN2 and MAST4) (Hibar et al.173). An interactive browser, ENIGMA-Vis—http://enigma-brain.org/ enigmavis—can be used to navigate ENIGMA’s genomic data. Initially started as a web page to plot ENIGMA summary statistics data for a specific genomic region, ENIGMA-Vis grew over the years into a portal with tools to query, visualize, and navigate the effects, and relate them to other GWAS. c In complementary work on rare variants by the ENIGMA-CNV Working Group, Sønderby and colleagues (2018) examined effects of the 16p11.2 distal CNV that predisposes to psychiatric conditions including autism spectrum disorder and schizophrenia. ENIGMA (including the 16p11.2 European Consortium) and deCODE datasets were combined to discover negative dose-response associations with copy number on intracranial volume and regional caudate, pallidum and putamen volumes—suggesting a neuropathological pattern that may underlie the neurodevelopmental syndromes. The agreement across datasets is apparent in the Forest plots for each brain region. [Data adapted, with permission from the authors and publishers]. central understanding gained from the ENIGMA asso- overlapping genetic drivers, perhaps boosting the power ciation screens is that neuroimaging genetics studies—just to discover genetic loci, by aggregating regions based on like analyses of behavioral measures, require tens (perhaps their genetic correlation. hundreds) of thousands of participants to obtain robust and reproducible effects of common polymorphisms. Uncovering the genetic basis of brain change Most individual effect sizes are very small explaining The quest to discover genetic loci that modulate brain <0.2% of variance, as for other complex human traits. development and aging led to the launch of the ENIGMA- GWAS of multiple imaging measures may offer a way to Plasticity WG38, which uses longitudinal brain imaging parcellate the brain into clusters or sectors with data from 36 cohorts worldwide to estimate rates of brain Thompson et al. Translational Psychiatry (2020)10:100 Page 10 of 28 growth or atrophy, and performs GWAS to find genetic GWAS results. Anatomical or genetic clustering is yet markers that may influence these rates of change. The another methodological improvement implemented by ENIGMA-Plasticity WG has established the heritability of ENIGMA, that can be used widely to enhance detection of brain changes over time and has shown that distinct genetic associations in multiple brain disorders (see Lor- genetic factors influence regional brain volumes and their enzi, Couvy-Duchesne for other multivariate imaging rate of change, implying the existence of genetic variants GWAS approaches50,51). specifically associated with change39. The WG is further investigating how closely developmental and aging-related Uncovering the epigenetic basis of brain morphometric genes overlap, and how they overlap with genetic loci that variation are associated with risk for development of psychiatric Inspired by these successes, ENIGMA widened the and neurological disease throughout life. Overall, the high scope of its WGs to embrace the study of epigenetic rate of discovery driven by ENIGMA is offering initial variations. ENIGMA’s Epigenetics group has already glimpses of the overlap among genetic drivers of brain identified two sites in the genome where methylation change throughout life with specific markers of brain relates to hippocampal volume (N = 3337)52. Ongoing structure and function. studies focus on brain measures sensitive to epigenetic age, an index of biological as opposed to chronological Uncovering the genetic basis of brain functional variation aging, in both health and disease. The ENIGMA Consortium has also carried out genetic association studies of EEG-derived phenotypes. The first From common nucleotide variations to rare copy number study40 of the EEG WG performed the largest GWAS to variants (CNV) date of oscillatory power across a range of frequencies The ENIGMA-CNV WG was launched to study the (delta 1–3.75 Hz, theta 4–7.75 Hz, alpha 8–12.75 Hz, and effects of CNVs, relatively rare genetic variants predis- beta 13–30 Hz) in 8425 healthy subjects. They identified posing individuals to various neuropsychiatric disorders. several novel genetic variants associated with alpha The ENIGMA collaborative approach is ideal for studying oscillatory brain activity that were previously linked to low-frequency variants, as such efforts require large psychiatric disorders. samples that are usually beyond the scope of a single study. Their first reports were on the 16p11.2 distal53 and Characterizing the association between brain morphology 15q11.295 CNVs (Fig. 3) and additional studies on other and disease-risk genes CNVs are underway. In an early ENIGMA study, minimal overlap was detected between schizophrenia-related and brain-related ENIGMA disorder-based neuroimaging studies genetic loci37. These questions were revisited with Baye- ENIGMA-schizophrenia sian models41 and LD-score regression methods42 which The Schizophrenia WG was formed in 2012, and has identified stronger overlap between genetic loci involved since analyzed data from 39 cohorts worldwide and has in cortical structure and loci implicated in insomnia, identified case–control differences in brain morpho- major depression, Parkinson’s disease, and general cog- metry1,54,55 and WM microstructure56,57, on an unpre- nitive ability or IQ13. Despite initial negative results37, cedented scale. ENIGMA-Schizophrenia was the first ENIGMA’s growing sample size led to more powerful working group to publish large-scale analyses of disease, results, allowing for the recent successes in the discovery in two seminal papers on case–control differences in of brain-related genetic variants that also affect risk for brain morphometry based on the largest samples to date. schizophrenia43,44, OCD45, anxiety disorders46, PTSD46, Van Erp and ENIGMA colleagues54 first reported that ADHD47, anorexia nervosa48, Tourette syndrome49, and patients with SCZ (N = 2028 patients) had smaller hip- insomnia13. pocampus (Cohen’s d = −0.46), amygdala (d = −0.31), As the sample size of brain scans in the ENIGMA thalamus (d = −0.31), nucleus accumbens (d = −0.25), Consortium increased beyond 50,000 MRI scans, it total intracranial volumes (d = −0.12), and larger palli- became possible to discover further genetic loci associated dum (d = 0.21) and lateral ventricle volumes (d = 0.37) with multiple brain traits implicated in brain disorders. A compared to healthy controls (N = 2540). In a subsequent recent example is an ENIGMA-CHARGE GWAS of white study, the team expanded their sample to include 4474 matter (WM) hyperintensities, a sign of vascular brain individuals with SCZ and 5098 controls to study cortical disease, by Mather et al. (in prep), which found hetero- structures1. Compared to healthy controls, patients with geneous effects for variants associated with lesions near SCZ had globally thinner cortices (left/right hemisphere: the ventricles versus lesions elsewhere in the brain. An d = −0.53/−0.52) and smaller overall cortical surface area innovative feature of this analysis was the use of anato- (left/right hemisphere: d = −0.25/−0.25), with greatest mical clustering of traits to yield more powerful brain effect sizes in frontal and temporal regions. Thompson et al. Translational Psychiatry (2020)10:100 Page 11 of 28 Figures 4 and 5 present these cortical and subcortical standard anatomical MRI. ENIGMA’s Diffusion MRI findings alongside data from several other disorders. It is working group, launched in 2012 with protocols for dif- notable that these findings from ENIGMA13,54 were fusion tensor imaging (DTI), published a series of papers replicated in a large independent study by the Japanese on the heritability and reproducibility of DTI measures COCORO Consortium58, and a recent Norwegian study derived with a protocol based on tract-based spatial sta- of 16 cohorts by Alnæs et al.59. The convergence of all tistics64–66. Over ten of ENIGMA’s working groups have three studies, reviewed in Kochunov et al.60, represents a since used this protocol to rank effect sizes for DTI new level of rigor and reproducibility in a field where the metrics across key WM tracts. existence of morphometric correlates of schizophrenia Kelly et al. reported on widespread WM abnormalities was once hotly debated61. in schizophrenia, pooling data from 2359 healthy controls Brain alterations were also discovered in relation to and 1963 patients with SCZ from 29 independent inter- clinical features of the disease. In follow-up analyses, national studies56. Significant reductions in fractional Walton et al. found that positive symptom severity was anisotropy (FA) in patients with SCZ were widespread negatively related to the thickness of the superior tem- across major WM fasciculi. While effect sizes varied by poral gyrus bilaterally62, while the severity of negative tract and included significant reductions in the anterior symptoms was negatively related to the cortical thickness corona radiata (d = 0.40) and corpus callosum (d = 0.39, of several prefrontal regions and particularly the left specifically its body (d = 0.39) and genu (d = 0.37)), medial orbitofrontal cortex (MOFC)63. effects were observed throughout the brain, with peak At this point it is worth considering the added value of reductions observed for the entire WM skeleton (d = other data modalities, such as diffusion MRI, which offers 0.42). Figure 6 shows these findings alongside data from complementary information on microstructural abnorm- two other disorders for which ENIGMA published large- alities, especially in the WM, that are not detectable on scale DTI analyses, MDD67, and 22q11DS17. Fig. 4 ENIGMA’s large-scale studies of nine brain disorders. Cortical gray matter thickness abnormalities as Cohen’s d, are mapped for nine different disorders, for which worldwide data were analyzed with the same harmonized methods. Although the cohorts included in the studies differed, as did the scanning sites and age ranges studied, some common and distinct patterns are apparent. Cortical maps for major depressive disorder (MDD), bipolar disorder (BD) and schizophrenia show gradually more extensive profiles of deficits. Across all disorders, the less prevalent disorders tend to show greater effects in the brain: the relatively subtle pattern of hippocampal-limbic deficits in MDD broadens to include frontal deficits in bipolar disorder (consistent with frontal lobe dysfunction and impaired self-control). In schizophrenia, deficits widen to include almost the entire cortex—only the primary visual cortex (specifically the calcarine cortex) failed to show thickness alterations in patients, after meta-analysis. Autism spectrum disorder (ASD) and the 22q deletion syndrome (22q11DS)—a risk condition for ASD—are associated with hypertrophy in frontal brain regions, while patients with obsessive-compulsive disorder (OCD) and alcohol use disorder tend to show deficits in frontal brain regions involved in self-control and inhibition. More refined analyses are now relating symptom domains to these and other brain metrics, within and across these and other disorders. Thompson et al. Translational Psychiatry (2020)10:100 Page 12 of 28 Fig. 6 White matter microstructure in schizophrenia, major depressive disorder, and 22q11.2 deletion syndrome. a White matter microstructural abnormalities are shown, by tract, based on the largest-ever diffusion MRI studies of these three disorders. In Fig. 5 Subcortical abnormalities in schizophrenia, bipolar schizophrenia (SCZ), fractional anisotropy, a measure of white matter disorder, major depressive disorder, and ADHD. a ENIGMA’s microstructure, is lower in almost all individual regions, and in the full publications of the three largest neuroimaging papers on skeleton. In major depressive disorder (MDD), a weak pattern of effects schizophrenia (SCZ), bipolar disorder (BD), and major depressive is observed, again with MDD patients showing on average lower FA disorder (MDD), suggested widespread cross-disorder differences in across the full white matter skeleton, when compared to controls. In effects (van Erp et al.54, Hibar et al.68). By processing 21,199 people’s comparisons between 22q11.2 deletion syndrome (22q11DS) and brain MRI scans consistently, we found greater brain structural matched controls, by contrast, the average FA along the full white abnormalities in SCZ and BD versus MDD, and a very different pattern matter skeleton does not show systematic differences; instead, while in attention-deficit/hyperactivity disorder (ADHD; Hoogman et al.7). some regions do show on average lower FA in affected individuals Subcortically, all three disorders involve hippocampal volume deficits compared with controls, several white matter regions show higher FA. —greatest in SCZ, least in MDD, and intermediate in BD. As a slightly b Relative to appropriately matched groups of healthy controls (HC), simplified ‘rule of thumb’, the hippocampus, ventricles, thalamus, group differences in fractional anisotropy are shown for ENIGMA’s amygdala and nucleus accumbens show volume reductions in MDD studies of SCZ, MDD (both in adults), and 22q11.2 deletion syndrome. that are around half the magnitude of those seen in BD, which in turn [Data adapted, with permission of the authors and publishers, from are about half the magnitude of those seen in SCZ. The basal ganglia Kelly et al.56, van Velzen et al.67, and Villalón-Reina et al.17; a key to the are an exception to this rule—perhaps because some antipsychotic tract names appears in the original papers; some tracts (i.e. the treatments have hypertrophic effects on the basal ganglia, leading to hippocampal portion of the cingulum) were omitted from the volume excesses in medicated patients. In ADHD, however, the 22q11DS analysis as they were not consistently in the field of view for amygdala, caudate and putamen, and nucleus accumbens all show some cohorts of the working group]. deficits, as does ICV (ventricular data is not included here for ADHD, as it was not measured in the ADHD study). A web portal, the ENIGMA Viewer, provides access to these summary statistics from ENIGMA’s published studies of psychiatric and neurological disorders (http:// gyrus (d = −0.29), and left rostral middle frontal cortex enigma-viewer.org/About_the_projects.html). b Independent work by (d = −0.28). Interestingly, lithium use was associated with the Japanese Consortium, COCORO, found a very similar set of effect thicker cortex in several areas. The WG also examined sizes for group differences in subcortical volumes between case–control differences in subcortical volumes in 1710 schizophrenia patients and matched controls. patients with BD and 2594 healthy controls; they found that BD was associated with reductions in the volume of ENIGMA-BD the hippocampus (d = −0.23) and the thalamus (d = Formed shortly after the Schizophrenia WG, and fol- −0.15), and with enlarged lateral ventricular volume (d = lowing similar protocols, the ENIGMA’s BD WG reported 0.26). A follow-up study, showed that when applied to on cortical thickness and surface area measures using regional cortical thickness, surface area, and subcortical anatomical MRI data from 1837 adults with BD and 2582 volumes, machine learning methods (based on support healthy controls, from 28 international groups68. BD was vector machines) differentiated BD participants from associated with reduced cortical thickness in bilateral controls with above chance accuracy even in a large and frontal, temporal and parietal regions, and particularly in heterogeneous sample of 3020 participants from 13 the left pars opercularis (d = −0.29), the left fusiform ENIGMA cohorts worldwide23. Aggregate analyses of Thompson et al. Translational Psychiatry (2020)10:100 Page 13 of 28 individual subject data yielded better performance than trained to predict chronological age from morphometric meta-analysis of site-level results. Age and exposure to data from 2188 controls across 19 cohorts and subse- anticonvulsants were associated with greater odds of quently applied to hold-out data from 2126 healthy con- correct classification. Although short of the 80% clinically trols and 2675 people with MDD. The largest brain aging relevant threshold, the 65.2% accuracy (0.71 ROC-AUC) effects were observed in antidepressant users (+1.4 years; is promising, as the study focused on a difficult to diag- d = 0.15), currently depressed (+1.5 years; d = 0.18), and nose, highly heterogeneous condition and used only remitted patients (+2.2 years; d = 0.18), compared to engineered features, not raw brain imaging data. ENIG- controls. Within ENIGMA-MDD, Opel et al. also studied MA’s multi-site design may also offer a more realistic the effects of obesity on structural brain metrics of assessment of “real-world” accuracy, by repeatedly leaving patients and controls (N = 6420)73. Obesity effects were out different sites’ data for cross-validation. Future mul- not different between patients and controls, but there was tisite brain-imaging machine learning studies will begin to a significant obesity by age interaction in relation to move towards sharing of more detailed individual subject cortical thickness, with thinner cortices in older obese data, not only a selection of discrete features or site-level individuals. Cortical thickness deficits related to obesity results derived from a single modality; unsupervised were strongest in the temporal and frontal cortical machine learning techniques may offer potential to better regions, and overlapped with patterns observed in several understand the heterogeneity in the disorder. The neuropsychiatric disorders, but exceeded those found in ENIGMA-BD DTI WG conducted both a mega- and MDD without regard for BMI—in terms of the effect sizes meta-analysis of 3033 subjects (1482 BD and 1551 con- and range of structures affected. The magnitude of these trols)69. Both analyses found lower FA in patients with BD effects suggests a need to better understand the connec- compared with healthy controls in most brain regions, tions between BMI, brain aging and mental health. with the highest effect sizes in the corpus callosum and Capitalizing on the statistical power of ENIGMA to cingulum. examine the role of risk factors, Frodl74 and Tozzi75 examined the association between retrospectively assessed ENIGMA-MDD childhood maltreatment (including emotional, physical and Brain morphometric analyses conducted by the sexual abuse, or emotional and physical neglect), and brain ENIGMA-MDD WG were based on MRI data from 1728 morphometry in 3036 and 3872 individuals (aged 13–89) patients with MDD and 7199 controls for subcortical with and without MDD, respectively. Greater exposure to volumes70 and from 2148 patients with MDD and 7957 childhood maltreatment was associated with lower cortical controls for cortical measures3. These studies found that thickness of the banks of the superior temporal sulcus and patients with MDD had lower hippocampal volumes (d = supramarginal gyrus, and with lower surface area across the −0.14), an effect driven by patients with recurrent illness whole brain and in the middle temporal gyrus. Sex differ- (d = −0.17) and by patients with an adolescent (≤21 ences were also observed: in females, greater maltreatment years) age of onset (d = −0.20). First-episode patients severity was associated with overall lower gray matter showed no subcortical volume differences compared to thickness and smaller caudate volumes, whereas in males, controls. Adult patients (>21 years) had reduced cortical greater maltreatment severity was associated with lower thickness in bilateral orbitofrontal cortex (OFC), anterior thickness of the rostral anterior cingulate cortex. and posterior cingulate cortex, insula, and temporal lobe In addition to these investigations of gray matter in regions (d’s: −0.10 to −0.14). In contrast, adolescent MDD, a large-scale analysis of WM microstructure with patients showed no differences in cortical thickness but DTI has also been completed, comparing 1305 adults and showed lower total surface area, which seemed to be adolescents with MDD to 1602 healthy controls from especially driven by lower surface area in frontal (medial 20 samples worldwide67. In adults with MDD, widespread OFC and superior frontal gyrus), visual, somatosensory, lower FA values were found in 16 out of 25 WM tracts of and motor areas (d = −0.26 to −0.57). Moreover, these interest (d’s = 0.12–0.26), with the largest differences in differences in gray matter morphometry observed in the corpus callosum and corona radiata. Widespread MDD do not involve abnormal asymmetry, as shown in a increased radial diffusivity (RD) was also observed (d’s = joint study by the Laterality and the MDD WGs involving 0.12–0.18) and was driven by patients with recurrent 2540 MDD individuals and 4230 controls, from 32 MDD and an adult-onset of depression. datasets71. A follow-up analysis on a subset of these aforemen- ENIGMA-PGC Post-Traumatic Stress Disorder tioned data found that the brain MRIs of adult patients In partnership with the PGC, ENIGMA launched a WG with MDD (18–75 years old) appeared, on average, 1.08 on PTSD that has analyzed neuroimaging and clinical years older than those of controls (d = 0.14)72. This ‘brain data from 1868 individuals (including 794 patients with age’ estimate was based on a machine learning algorithm PTSD) from 16 cohorts. In this first ENIGMA-PTSD Thompson et al. Translational Psychiatry (2020)10:100 Page 14 of 28 study, Logue and colleagues found that patients with to date of brain structure in adult and pediatric OCD, current PTSD had smaller hippocampal volumes (d = using both meta- and mega-analytic approaches6,88. The −0.17) compared to trauma-exposed controls4. Child- first study analyzed MRI scans from 1830 patients diag- hood trauma predicted smaller hippocampal volume (d = nosed with OCD and 1759 controls across 35 cohorts −0.17) independent of diagnosis. In a subsequent study, from 26 sites worldwide88. Unmedicated pediatric OCD the WG found that cortical thickness in 3378 individuals patients demonstrated larger thalamic volumes, while the (including 1309 patients with PTSD) was lower in PTSD pallidum was enlarged in adult OCD patients with disease in the orbitofrontal cortex, cingulate cortex, precuneus, onset at childhood. Adult OCD patients also had sig- insula, and lateral parietal cortices. In addition, a DTI nificantly smaller hippocampal volumes (d = −0.13), with meta-analysis of 3057 individuals (including 1405 patients stronger effects in medicated patients with adult-onset with PTSD) from 25 cohorts found alterations in WM OCD compared to healthy controls (d = −0.29). A cor- organization in the tapetum, a structure that connects the tical study included data from 1905 patients diagnosed left and right hippocampus76. Structural covariance net- with OCD and 1760 healthy controls across 38 cohorts work analysis applied to data from 3505 individuals from 27 sites worldwide. In adult patients diagnosed with (including 1344 patients with PTSD), which examined OCD versus controls, significantly smaller surface area of correlated patterns of cortical thickness and surface area, the transverse temporal cortex (d = −0.16) and a thinner found that PTSD is associated with network centrality inferior parietal cortex (d = −0.14) were found. Medi- features of the insula and visual association areas77. To cated adult patients with OCD also showed thinner cor- extend these findings, ongoing studies are assessing cor- tices throughout the brain (Cohen’s d effect sizes varied tical structure78,79 and hippocampal subfields in PTSD between −0.10 and −0.26). Pediatric patients with OCD and MDD80–83, to better understand the pattern and showed significantly thinner inferior and superior parietal regional specificity of hippocampal deficits in the two cortices (d’s = −0.24 to −0.31), but none of the regions disorders, and whether these patterns coincide. analyzed showed significant differences in cortical surface area. However, medicated pediatric patients with OCD ENIGMA-Addictions/SUD had smaller surface area in frontal regions (d’s = −0.27 to The ENIGMA-Addictions/SUDs WG has 33 partici- −0.33), that may indicate a delayed cortical maturation. pating sites, contributing MRI data from 12,347 indivi- The absence of cortical surface area abnormalities in adult duals of whom 2277 are adult patients with SUD relating patients with a childhood onset of OCD could indicate a to one of five substances (alcohol, nicotine, cocaine, normalization of these abnormalities—a hypothesis that is methamphetamine, or cannabis)5,84,85. In these data, now being explored with longitudinal data collection. Mackey5 observed lower cortical thickness/subcortical To assess whether the anatomical differences could be volume in cases relative to controls in regions that play used to create a neuroimaging biomarker for OCD, a key roles in evaluating reward (MOFC, amygdala), task machine learning analysis of the cortical and subcortical monitoring (superior frontal cortex), attention (superior data was performed with 2304 OCD patients and 2068 parietal cortex, posterior cingulate) and perception/reg- controls. Classification performance across ten different ulation of internal body states (insula). While the most machine and deep learning approaches was poor. With pervasive case–control differences appeared to be related site-stratified cross-validation, the ROC-AUC ranged to alcohol dependence, some effects were observed for between 0.57 and 0.62. The performance dropped to substance dependence generally (e.g., the insula and chance level when leave-one-site-out cross-validation was MOFC). A support vector machine trained on cortical used, with classification performance between 0.51 and thickness and subcortical volume successfully classified 0.54. This indicates that these anatomical brain features set-aside test sets for both alcohol (ROC-AUC: 0.74–0.78; do not provide a biomarker for OCD. But when patients p < 0.0001) and nicotine dependence (ROC-AUC: were stratified according to whether they had used med- 0.60–0.64; p < 0.0001), relative to non-dependent con- ication, classification performance improved remarkably. trols5. A separate meta-analysis also compared the effect Medicated OCD patients and controls could then be size of addiction-related brain impairment to that of other distinguished with 0.73, unmedicated OCD and controls psychiatric disorders: effect sizes of alcohol-related brain with 0.61, and medicated and unmedicated OCD patients differences in subcortical brain regions were equivalent to with 0.86 ROC-AUC. These multivariate results therefore those reported for schizophrenia86. mirror the univariate results, and highlight that medica- tion use is associated with large differences in brain ENIGMA-Obsessive-Compulsive Disorder anatomy89. The ENIGMA’s OCD WG grew out of a previously The OCD WG, in conjunction with the Laterality WG, established consortium (the OCD Brain Imaging Con- studied brain asymmetry in OCD using 16 pediatric sortium, or OBIC)87, and has published the largest studies datasets (501 patients with OCD and 439 healthy Thompson et al. Translational Psychiatry (2020)10:100 Page 15 of 28 controls), and 30 adult datasets (1777 patients and 1654 even when the disease is not, and be found in non-affected controls)90. In the pediatric datasets, the largest family members at a higher rate than in the general case–control differences were observed for volume population95,96), as they were also seen in unaffected asymmetry of the thalamus (more leftward in patients siblings of people with ADHD in a subsample analysis of compared to controls; d = 0.19) and the pallidum (less the cortical features. The stronger effects in children may leftward in patients compared to controls; d = −0.21). No reflect a developmental delay, perhaps due in part to asymmetry differences were found in the adult datasets. genetic risk factors, given recent findings of overlap These findings may reflect altered neurodevelopmental between the genetic contributions to ADHD and to sub- processes in OCD, affecting cortico-striato-thalamo- cortical volumes13,47. cortical circuitry, which is involved in a wide range of cognitive, motivational and emotional processes. ENIGMA-Autism Spectrum Disorders The ENIGMA-ASD WG published the largest neuroi- ENIGMA-Attention-Deficit/Hyperactivity Disorder maging study of autism analyzing data from 1571 partici- ENIGMA’s ADHD WG has analyzed data from up to pants with ASD and 1651 controls, from 49 sites worldwide 2264 participants with ADHD and 1934 controls from up (ages 2–64 years)8. Unlike most of the disorders discussed to 36 sites (age range: 4–63 years; 66% males)91. Volumes so far, the direction of effects seen in ASD varied by brain of the nucleus accumbens (d = −0.15), amygdala (d = region, and did so across the age span analyzed. ASD was −0.19), caudate (d = −0.11), hippocampus (d = −0.11), associated with larger lateral ventricle and intracranial putamen (d = −0.14), and ICV (d = −0.10) were smaller volumes, greater frontal cortical thickness and lower tem- in cases relative to controls. Effect sizes were highest in poral cortical thickness (d = −0.21 to 0.20). Participants children. No statistically significant univariate case/con- with ASD also had, on average, lower subcortical volumes trol differences were detected in adults. Volume differ- for the pallidum, putamen, amygdala, and nucleus accum- ences were found to have similar effect sizes in those bens. Post hoc fractional polynomial analyses showed a treated with psychostimulant medication and those naïve sharp increase in volumes in the same regions in childhood, to psychostimulants. Bioinformatics analyses suggested peaking in adolescence and decreasing again in adulthood. that the selective subcortical brain region vulnerability Overall, patients with ASD showed altered morphometry in was associated with differential expression of oxidative the cognitive and affective associated-regions of the stria- stress, neurodevelopment and autophagy pathways92. tum, frontal cortex, and temporal cortex. The ENIGMA-ADHD WG was the first WG in The ASD group worked together with the Laterality ENIGMA to perform a detailed investigation of the case- group to produce the largest ever study of brain asym- control effects on the cerebellum. Differential age trajec- metry in ASD, involving 1774 patients and 1809 controls, tories were identified for children with ADHD when from 54 datasets97. Generally, subtle but widespread compared with typically developing children for the cor- reductions of cortical thickness asymmetries were present pus medullare93. in patients with ASD compared to controls, as well as In an analysis of the cerebral cortex, lower surface area volume asymmetry of the putamen, and surface area values were found, on average, in children with ADHD, asymmetry of the MOFC (the strongest effect had mainly in frontal, cingulate, and temporal regions; the Cohen’s d = −0.16). Altered lateralized neurodevelop- largest effect was for total surface area (d = −0.21). ment may, therefore, be a feature of ASD, affecting Fusiform gyrus and temporal pole cortical thickness were widespread cortical regions with diverse functions. also lower in children with ADHD. All effects were most pronounced in early childhood. Neither surface area nor Neurogenetic disorders, CNV, and rare thickness differences were found in the adolescent or neurodevelopmental conditions adult groups7, but machine learning analyses supported Several neurodevelopmental disorders arise due to the the hypothesis that the case–control differences observed abnormal duplication or deletion of segments of the gen- in childhood could be detected in adulthood94. Impor- ome. ENIGMA has dedicated WGs studying 22q11DS, tantly, many of the same surface area features were Gaucher’s disease, and Hepatic Glycogen storage dis- associated with subclinical ADHD symptoms in children ease98,99, along with a CNV WG meta-analyzing imaging from the general population that do not have a clinical data from carriers of several other CNVs53,100. Here, we psychiatric diagnosis. Several of the observed brain focus on the work of the two most established groups, that alterations fulfilled many of the criteria of ‘endopheno- examine carriers of 22q11.2 deletions and other CNVs. types’ (An endophenotype is a trait, such as brain struc- ture or function, related to the biological process of a ENIGMA-22q11.2 Deletion Syndrome disorder; to qualify as an endophenotype, the trait, should 22q11DS is associated with a 20-fold increased risk for be heritable, co-segregate with an illness, yet be present psychosis, and an elevated risk for developmental Thompson et al. Translational Psychiatry (2020)10:100 Page 16 of 28 neuropsychiatric disorders such as ASD. 22q11DS provides gyri. The pattern of results fits well with known molecular a ‘genetics-first’ framework to study the brain markers functions of the genes in the 15q11 region and suggests underlying complex psychiatric phenotypes. The ENIGMA- involvement of these genes in neuronal plasticity and 22q11DS working group analyzed the largest dataset to date cortical development. Thus, the results from ENIGMA- of brain images from patients with 22q11DS from 10 CNV have shown that several CNVs cause abnormal brain cohorts including 466 individuals with 22q11DS and 374 patterns and inform on genetically determined variation matched controls. Compared to controls, 22q11DS indivi- in brain development and their relation to neurodeve- duals showed overall thicker cortical gray matter (left/right lopmental disorders. Additional studies on other CNVs hemispheres: Cohen’s d = 0.61/0.65), but pervasive reduc- are in progress. tions in cortical area (left/right hemispheres: d = −1.01/ −1.02), with specific anatomic patterns. Machine learning Newly established working groups methods were applied to the cortical thickness and area In the last two years, seven additional ENIGMA WGs measures to achieve a high accuracy (sensitivity 94.2%; have formed to study specific disorders and important specificity 93.3%) in classifying 22q11DS cases and con- transdiagnostic conditions: anxiety disorders, suicidal trols10. ENIGMA subcortical shape analysis pipelines also thoughts and behavior, sleep and insomnia, eating dis- identified complex structural differences across many sub- orders (including bulimia and anorexia nervosa sub- cortical structures between individuals with 22q11DS and groups104), irritability, antisocial behavior, and controls101. Analysis of diffusion MRI from the same sub- dissociative identity disorder. The starting point of the jects (N = 594) revealed abnormalities in the corpus callo- anxiety group was an international voxel-based morpho- sum, superior longitudinal fasciculus, and corona radiata17. metry mega-analysis on social anxiety disorder105, sup- Ongoing work uses more advanced imaging protocols17— ported by findings demonstrating that structural brain including ‘multishell’ diffusion protocols that allow for the alterations related to social anxiety run in families106. At estimation of biophysical compartments in the tissue—to present, the anxiety WG has four subgroups including test hypotheses about specific cellular processes and specific over 5000 patients: besides social anxiety disorder (1250 fiber tracts that may be especially vulnerable in 22q11DS patients)107, there are groups devoted to generalized (e.g., the corpus callosum), as well as fiber tracts that appear anxiety disorder (1329 patients), panic disorder (1300 to be relatively spared (e.g., the cortico-fugal tracts17). patients), and specific phobia (1224 patients), allowing for disorder-specific and cross-disorder comparisons. The ENIGMA-Copy Number Variations antisocial behavior WG aims to clarify how conduct dis- This WG was set up to examine the effect of rare CNVs, order, psychopathy, and antisocial personality disorder risk factors for a variety of neuropsychiatric disorders, on relate to differences in brain structure, function, and brain structure. Due to their low prevalence102,103, their connectivity. Its goals include examination of different effects on the brain have been hard to establish. Sønderby phenotypes (e.g., reactive vs proactive aggression), and colleagues focused on the 16p11.2 distal CNV that population-based samples with dimensional measures of predisposes to psychiatric conditions including ASD and antisocial behavior, and genetic data from case–control schizophrenia. ENIGMA (including the 16p11.2 European and population-based studies. Consortium) and deCODE datasets were combined to Building on the promising findings from the psychiatric compare subcortical brain volumes of carriers of 15 WGs, ENIGMA established seven WGs studying specific 16p11.2 distal deletion and 18 duplication to 7714 non- conditions in neurology and cancer-related cognitive carriers which led to the discovery of negative dose- impairment: epilepsy, traumatic brain injury, Parkinson’s response associations with copy number on intracranial disease, neuro-HIV, ataxia, stroke recovery, and cancer/ volume and regional accumbens, caudate, pallidum and chemotherapy effects on the brain108,109. putamen volumes—suggesting a neuropathological pat- tern that may underlie the neurodevelopmental syn- ENIGMA-Epilepsy dromes53. A further study100 including the UK Biobank The ENIGMA-Epilepsy WG combined data from 24 assessed the association of the 15q11.2 CNV with cogni- centers across 14 countries to create the largest neuroi- tion and cortical and subcortical morphology in more maging study to date of epilepsy9. Data from 2149 indi- than 45,000 individuals from 38 datasets (203 individuals viduals with epilepsy were divided into four common with a 15q11.2 deletion, 45,247 non-carriers, and 306 epilepsy syndromes: idiopathic generalized epilepsies (N duplication carriers). The authors found a clear pattern of = 367), mesial temporal lobe epilepsies with hippocampal widespread poorer cognitive performance, smaller surface sclerosis (MTLE; left, N = 415; right, N = 339), and all area and thicker cortices for deletion carriers compared to other epilepsies in aggregate (N = 1026), compared to non-carriers and duplication carriers, particularly across 1727 matched healthy controls. Compared to controls, all the frontal lobe, anterior cingulate and pre/postcentral epilepsy groups showed lower volume in the right Thompson et al. Translational Psychiatry (2020)10:100 Page 17 of 28 thalamus (d = −0.24 to −0.73), and lower thickness in the FA in veterans and service members with blast-related precentral gyri bilaterally (d = −0.34 to −0.52). Both injuries, and altered subcortical volumes in the group with MTLE subgroups also showed profound volume reduc- military TBI overall117. Work is ongoing to study the tion in the ipsilateral hippocampus (d = −1.73 to −1.91), effects of injuries sustained during and outside deploy- and lower thickness in cortical regions, including the ment, and severity and mechanisms of injury. precentral and paracentral gyri (d = −0.36 to −0.52) compared to controls. Notably, the effect sizes for cortical ENIGMA-Parkinson’s Disease differences in this neurological disorder were much ENIGMA’s Parkinson’s disease WG has analyzed scans greater than those seen in all complex psychiatric dis- from 11 cohorts spanning 10 countries including 1288 orders. In an approach known as ‘virtual histology’, a patients with PD and 679 controls (age: 20–89 follow-up study110 overlaid the cortical deficit maps on years)122,123. A PD diagnosis was associated with moder- gene-expression data from the Allen Brain Atlas, and ately larger thalamic volumes (left: d = 0.29; right: d = detected enrichment for microglial markers in regions 0.17) and smaller pallidal volumes (left: d = −0.25; right: with greater deficits. The WG is currently combining DTI d = −0.21). There was also widespread and lower cortical data and exploring putative neuroanatomical biomarkers thickness in PD patients, while sparing the limbic and of medication treatment resistance and post-operative insular cortices. Ongoing work on a larger sample is outcomes. relating brain structure and WM microstructure to dis- ease severity, medication status and history and duration ENIGMA-Brain Injury of the illness as modifiers of these robust differences ENIGMA’s Brain Injury WG111 combines data from 72 between patients and controls. centers, and is organized into ten separate subgroups that focus on (1) acute mild traumatic brain injury (TBI), (2) adult moderate/severe TBI112, (3) pediatric moderate/ ENIGMA-Human Immunodeficiency Virus severe TBI113,114, (4) military-related brain injury115–118, The availability of combination antiretroviral therapy (5) sports-related concussion119, (6) intimate partner (cART) has now transformed HIV-infection from a pos- violence120, (7) MR spectroscopy121, (8) arterial spin sibly fatal diagnosis to a chronic condition, allowing for labeling, (9) resting state fMRI, and (10) cognitive end- viral suppression and stable immune function; however, points. These groups have recently started-up compared despite inconsistencies in neuroimaging studies, neuro- to other ENIGMA WGs, but are rapidly expanding in logical symptoms and consequences persist. This WG has membership and focus. In addition to meta- and mega- pooled data from 12 independent neuro-HIV studies from analyses of relevant existing datasets, the Brain Injury Africa, Asia, Australia, Europe, and North America; WGs endeavor to further extend efforts to promote volume estimates for eight subcortical brain regions were increased consistency in prospective data collection, both extracted from anatomical MRI from 1044 HIV + adults in terms of imaging data and associated cognitive out- (age: 22–81 years) to identify associations with plasma come data. Additionally, the WGs are engaged in the markers reflecting immunosuppression (CD4+ T-cell development of novel pipelines and analytic tools that count) or viral load124. Across participants, lower cur- address brain-injury specific issues or incorporate rent CD4+ count was associated with smaller hippo- sequences or techniques that are potentially useful in campal and thalamic volumes. A detectable viral load was addressing injury associated pathology. For example, also associated with smaller hippocampal (d = 0.24) and future planned studies will compute structural pathology amygdalar volumes (d = 0.18), supporting the importance profiles for individual TBI patients, including (i) mapping of achieving viral suppression and immune restoration. of the heterogeneous lesions using advanced lesion These limbic effects are in contrast to many of the early mapping methods (such as disconnectome symptom neuro-HIV findings that focused on basal ganglia struc- mapping), (ii) accurate quantification of brain atrophy (of tures, yet we found the limbic associations were largely the different brain regions) using tensor based morpho- driven by participants on cART, while basal ganglia effects metry, and (iii) identification of subject-specific epicenters (putamen) were detected in the subset of participants not best predictive of neurodegeneration using network on cART. These findings demonstrate the continuing spread models. Finally, the Brain Injury WGs will inter- effects of HIV on the brain in the current “cART era”. face with other disease-specific WGs where comorbidity Alterations in brain structures that are essential for with brain injury is high (e.g., substance use, PTSD, MDD, learning and memory has clinical significance given ADHD), as well as with methods-focused WGs (e.g., dif- mounting evidence of HIV-associated deficits in these fusion imaging, etc.). A preliminary report on 117 parti- cognitive domains among older HIV+ adults, and the cipants with military-relevant blast-related versus 227 possibility that HIV may contribute to abnormal brain participants with non-blast related injury revealed higher aging125. Thompson et al. Translational Psychiatry (2020)10:100 Page 18 of 28 ENIGMA-Ataxia Diffusion MRI WG on DTI is one of the most long- This WG includes 21 sites pooling data from more than standing. DTI offers information on microstructural 750 individuals with inherited ataxias, including Frie- abnormalities that are not detectable on standard anato- dreich Ataxia and Spinocerebellar Ataxia (SCA) 1, 2, 3, 6, mical MRI. As mentioned earlier, this WG has published a and 7 (the poly-glutamine SCAs), alongside over 800 series of papers on the heritability and reproducibility of controls. This group is undertaking optimization and DTI measures derived with a custom protocol based on standardization of protocols for cerebellar voxel-based tract-based spatial statistics64,134. Diagnosis-based WGs morphometry and parcellation, upper spinal cord cross- have used this protocol to rank effect sizes for DTI sectional area, and brainstem volume, in line with the key metrics as previously described or are undertaking similar regions of pathology in these diseases. Preliminary work studies including in 22q11DS17, epilepsy, PTSD76, military indicates that gray matter degeneration principally TBI113, HIV135, and OCD136. impacts the cerebellar anterior lobe in Friedreich ataxia, Other methodological WGs have focused on anatomical while all areas of the cerebellum are affected in the poly- shape analyses that enable a more precise characterization glutamine SCAs. However, both the magnitude and pat- of regional brain alterations thus resolving subregional tern of cerebellar gray matter degeneration are distinct effects in the basal ganglia, amygdala, and hippo- across these diseases and evolve with disease progression campus55,85,137–142. Other approaches currently used in and severity. ENIGMA include brain structural covariance analysis graph theory approach for intra-individual brain structural cov- ENIGMA-Stroke Recovery ariance networks in OCD77,143, sulcal morphometry144, The ENIGMA-Stroke recovery WG has addressed a hippocampal subfield analysis80–82,145 and disease effects on major gap in stroke research relating to the large-scale lateralization (in OCD, MDD, and ASD)71,90,97. More definition of lesion masks. Researchers in this WG have recently, ENIGMA’s Brain Age WG was formed to apply released a public archive of 304 T1-weighted MRIs with various algorithmic estimators of ‘brain age’ across several manually segmented stroke lesion masks126 (https://www. ENIGMA WGs72. From the ENIGMA-Brain Injury group, icpsr.umich.edu/icpsrweb/ADDEP/studies/36684), and the MR spectroscopy (MRS) WG has formed to focus on developed open-source software127 and analyses specific the harmonization of MRS data which could reach across for scalable128, reproducible lesion analyses (https:// other WGs in the future. github.com/npnl/PALS). In addition to this major meth- odological contribution they have analyzed data from 629 The impact of ENIGMA participants from 22 sites worldwide to identify reliable The ENIGMA Consortium has been a driving force in predictors of motor function after stroke129,130. They the field of neuroscience by making substantial con- found that motor-related subcortical volumes in the basal tributions to the science of brain variation and shaping ganglia and thalamus are positively associated with post- the working practices of the field at various levels. In stroke motor performance, and depend on impairment reflecting on the key achievements, three areas stand out: severity, time since stroke, and lesion laterality. In con- trast, enlarged lateral ventricles are associated with worse Promoting robustness and reproducibility post-stroke motor outcomes. The group now has data ENIGMA’s “big data” approach to neuroimaging from 1625 participants from 32 sites worldwide, and addresses directly the reproducibility challenges that pla- ongoing work in the group focuses on quantifying lesion gue many areas of biomedical science—including neu- overlap with major motor-related structures, such as the roscience146–148. Neuroimaging has received considerable corticospinal tracts and subcortical regions131,132, and scrutiny regarding the reliability of published findings, relating these measures with subcortical volumetric given the literature replete with studies based on small measures to motor outcomes133. samples and seemingly unlimited methodological free- dom149,150. Many other approaches also aim to tackle this ENIGMA-methods focused working groups reproducibility crisis, by building data repositories that can The ENIGMA Consortium functions as a driving force be accessed for replication151–153; yet ENIGMA offers an for the development, validation and implementation of opportunity to collaborate with teams of diverse experts novel methods to address the complexities of analyses of irrespective of whether or not any data is shared. In one large imaging datasets and to derive more mechanistic recent study by ENIGMA’s Laterality group, the authors insights into the processes that underpin variation in examined brain asymmetry in 99 MRI datasets worldwide brain organization in health and disease. To achieve this, (from N = 17,141 people) and found that, as expected, the ENIGMA has dedicated WGs focused on the develop- reproducibility of findings increased with the effect size ment of more innovative pipelines for data analyses to be and sample size, in a setting that was free from publication applied for various dataset worldwide. The ENIGMA bias (data available at: http://conxz.net/neurohemi/)18,154. Thompson et al. Translational Psychiatry (2020)10:100 Page 19 of 28 For example, for effect sizes of d ≥ 0.6, the reproducibility countries—some of which have minimal diplomatic ties— rate was higher than 90% even when including the datasets collaborations are not only constructive in terms of col- with sample sizes as low as 15, while it was impossible to lective problem solving, but they also build connections obtain 70% reproducibility for small effects of d < 0.2, even between high income countries and the poorest nations with a relatively large minimum sample size threshold of across the globe and to build capacity in the latter159. 500. The unprecedented size of the datasets analyzed across ENIGMA boosts statistical power to detect the Setting methodological standards effects of disease and their moderators73,155. Through data The ENIGMA Consortium has provided a blueprint for sharing, investigators can now identify patterns of brain multi-site standardization in terms of mining legacy abnormalities that consistently characterize disorders or neuroimaging and genetic data. The success of this clinical syndromes, while assessing their reproducibility approach is obvious when considering the volume of over across continents. This is exemplified by the close match 50 published works that has relied on the ENIGMA between the schizophrenia findings by ENIGMA54,60 and pipelines. Furthermore, funding bodies, such as the independent work by the Japanese Consortium, National Institutes of Health in the United States, have COCORO58 and a recent Norwegian study of 16 cohorts gained interest in such approaches; program announce- by Alnæs59. In all three studies, schizophrenia patients ments requesting applications on aggregating existing showed enlargement of the lateral ventricles, pallidum, biomedical data, or making use of existing resources, have putamen, and caudate, and volume reduction in the hip- become increasingly common. Moving forward, ENIGMA pocampus, amygdala, thalamus and accumbens, with a remains a test-bed of unprecedented scale and power for strong agreement in the magnitude and rank order of developing and benchmarking novel analytic methods. effects from highest to least group difference. Similarly, a This contribution is of paramount importance as recent GWAS study of the UK Biobank dataset156 was able advanced statistical modeling and bioinformatics become to replicate the majority of the genetic loci discovered by essential for analytic pipelines. In efforts complementary ENIGMA in two separate GWAS of subcortical to the technical advances made in multisite data collec- volumes24,25. Thus, the international, multi-site nature of tion initiatives such as the Alzheimer’s Disease Neuroi- ENIGMA studies likely promotes representative findings maging Initiative, the Human Connectome Project, and that are widely generalizable. Meanwhile, the larger and the UK Biobank, ENIGMA’s studies have required us to more diverse samples are valuable resources for under- continually develop new scientific approaches to analyze standing the heterogeneity across different studies, and data from diverse and independent populations. ENIGMA may provide new insights into the reproducibility issue has made several technical advances that may be adapted faced by the neuroimaging community. Moreover, to other domains, including creating, adapting, and ENIGMA offers a platform for investigators to converge extensively testing harmonized methods for distributed on methods for sharing and analyzing data acceptable to analysis, meta-analysis, and cross-site data integration (see the community. Supplementary Appendix A: Technical Contributions of ENIGMA also offers new opportunities to change the ENIGMA). landscape for how data can be used. In current research practices, a great resource of data remains largely Driving discovery untapped that is often known as “long-tail” data: data sets Neuroimaging and genetics are fields of both large and collected in individual laboratories that accumulate over small effects. For common, complex chronic diseases, many years and funding cycles157. Much valuable data effects on brain metrics can be very subtle, but for rare remains dormant (and unpublished) due to a lack of monogenic disorders and across the field of neurology— personnel and time to analyze it, and this is going to including epilepsy, brain injury, stroke and neuro-oncol- increase with studies including larger samples than before. ogy, disease effects can be relatively large (although not Efforts through ENIGMA to leverage ‘dormant’ data in exclusively). In the 10 years since ENIGMA was founded, labs throughout the world have at least three important the primary lesson has been on the power of worldwide advantages. First, data sharing increases the scope of the collaboration to discern subtle patterns in brain data, and science, enhancing opportunities for analyses not other- advance neuroscience beyond the capacity of any one wise possible with small sample sizes. Second, data shar- group of researchers collecting data on their own. New ing naturally engages scientists from distinct disciplines— discoveries regarding the factors that influence brain a crucial step for advancing the clinical neurosciences158. organization and its association with health and disease A final benefit that is sometimes overlooked in global are predicated on having adequate statistical power and scientific collaborations is their power to build and on developing new neuroimaging approaches aiming to enhance diplomatic relations and transcend political lead to more mechanistic explanations of the multi-scale conflicts between nations159. With representation from 43 organization of the brain. Thompson et al. Translational Psychiatry (2020)10:100 Page 20 of 28 Challenges and future directions Cross-disorder analyses Even given the advances made through ENIGMA dur- ENIGMA has recently created cross-disorder groups to ing its first decade, as a growing consortium, ENIGMA answer transdiagnostic questions that draw on data from faces important challenges. Thus far, ENIGMA has largely multiple WGs165. An exemplar of this approach is the relied on existing data, which implies a degree of het- newly formed ENIGMA-Relatives WG which examines erogeneity with respect to phenotyping—including clin- brain organization in the unaffected first-degree relatives ical assessments, scanners and imaging protocols. of patients with psychiatric disorders. The first study from Another limitation of this type of data is that the depth of this group focused on identifying common and distinct phenotyping varies across centers, which can lead to a anatomical patterns in patients with SCZ (N = 1016) or limited set of clinical and other scales shared by all cen- BD (N = 666) and their unaffected relatives (N for SCZ ters. As we discuss below, ENIGMA is now beginning to relatives = 1228 and for BD relatives = 852)11. A address these limitations with a series of newly funded remarkable finding from this study is that the first-degree and planned studies111,112,114,118–121. The paucity of relatives of BD patients had larger ICV compared to longitudinal data in the literature is also reflected within controls (d = 0.16) while first-degree relatives of SCZ ENIGMA, which includes a limited number of long- patients had smaller ICV and lower cortical thickness, and itudinal studies. Consequently, the data-driven approach when controlling for ICV, had regionally smaller thalamic used in ENIGMA is complementary, but not always volumes. Other newly formed groups aim to aggregate superior, to well-designed, hypothesis-driven, smaller- data across the spectrum of mental illness that may be scale prospective single-center or multi-center studies prone to similar symptoms and outcomes, such as suicidal with in-depth phenotyping. ideations and actions. Cross-disorder initiatives have formed within ENIGMA as partnerships between existing group members. The topology of collaboration (Fig. 7) Extending imaging modalities and computational includes working partnerships between group members approaches working on similar problems, providing natural connec- ENIGMA’s future developments will include the coor- tions between topics of study of the different groups. dinated analyses of new data modalities (such as resting state and task-related functional MRI160–163, as well as geostatistical and mobile sensor data), and deeper or more Sex differences refined analyses of current imaging modalities. Diffusion ENIGMA’s sex differences Initiative is probing disease MRI, in particular, is moving towards multi-shell proto- WG datasets to better understand sex disparities in risk cols that can better differentiate cellular and micro- factors, disease effects, or outcomes and their relationship structural sources of variance that may explain patterns with brain organization. In a new initiative, the ENIGMA- observed with DTI17. Multimodal projects that pool data Transgender WG is contributing additional insights with across imaging modalities are likely to boost the accuracy respect to the biological underpinnings of sex assigned at of machine learning methods for differential diagnosis, birth versus gender identity166. The first study from this outcome prediction, and subtyping. Unsupervised learn- group was based on more than 800 scans and pooled ing—applied to imaging and clinical data—may also help various MRI-based measures (cortical thickness, surface to identify homogeneous subgroups within and across area, and volume) across eight international sites. While disorders. Deep learning, for example, benefits from very effects varied depending on the morphometric measure large datasets, such as those analyzed in ENIGMA, and applied and the brain regions considered, a general pat- these and other artificial intelligence methods show pro- tern emerged: transgender men (assigned female at birth) mise in identifying unsuspected features and patterns in mostly resemble cis-gender women, whereas transgender images beyond those derived using traditional methods. women (assigned male at birth) range between cis-gender From its inception, ENIGMA has accommodated varying men and cis-gender women167. Ongoing initiatives data sharing practices across institutions and countries, examining sex-differences focus on sex-specific GWAS has used strategies (such as meta-analysis) to overcome studies, and developmental and aging trajectories. some of these, and is working with field experts on novel strategies (like COINSTAC or other distributed analysis Global health disparities approaches)164 to allow for more powerful analysis with- Health disparities, including those that exist in low and out sending data around the globe. On the ‘omics’ side, middle income countries, are also a topic of great interest whole genome sequencing promises to refine our under- for ENIGMA, as prevalence, treatments, and access to standing of causal loci across all phenotypes, from plasma healthcare varies within and across countries. While the markers and brain metrics to environmental exposure and analyses in ENIGMA so far tend to show cross-national clinical measures of disease burden. agreement in brain signatures and associated genetic loci Thompson et al. Translational Psychiatry (2020)10:100 Page 21 of 28 Fig. 7 Topology of large-scale scientific collaboration. a The topology of scientific collaboration in ENIGMA has some properties that resemble a modular hierarchical network (Ravasz and Barabasi170, Slaughter171). In this diagram (a), nodes represent individual scientists working on a project, and links denote active scientific collaborations (that might result in co-authored publications, like this review, for example). ENIGMA’s WGs resemble the yellow sets of nodes: guided by a small group of WG chairs, several clusters of scientists coordinate projects applying various methods to the same datasets (e.g., MRI and DTI meta-analysis, machine learning, and modeling of clinical outcomes). WGs study different disorders with the same harmonized methods, enabling to cross-disorder collaborations across WGs. The modular organization allows independent and coordinated projects to proceed in parallel, distributing work and coordination, without requiring a central hub for all communication. Real clusters may differ in their number of members and links [(b) shows a different graph with a similar hierarchical modular form], and may change dynamically over time as new groups and projects form and projects end. of various psychiatric diseases, more in-depth phenotyp- Publisher’s note ing may reveal circumstances where risk factors apply Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. more strongly to specific ethnic or sociodemographic groups, and means to remediate them, consistent with the Supplementary Information accompanies this paper at (https://doi.org/ concept of precision public health. 10.1038/s41398-020-0705-1). In closing, we reiterate ENIGMA’s mission statement, Received: 3 July 2019 Revised: 11 December 2019 Accepted: 20 December which reads: “Individually, we contribute little to the quest 2019 for truth, but working together, the whole vast world of science is within our reach.” (Aristotle, 350 BCE)168. References Acknowledgements 1. van Erp, T. G. M. et al. Cortical Brain Abnormalities in 4474 Individuals with The work reported here was supported in part by many public and private Schizophrenia and 5098 Control Subjects via the Enhancing Neuro Imaging agencies across the world. Individual authors’ funding is listed in Genetics through Meta Analysis (ENIGMA) Consortium. Biol. Psychiatry 84, Supplementary Appendix B. Core funding for ENIGMA was provided by the NIH 644–654 (2018). Big Data to Knowledge (BD2K) program under consortium grant U54 2. Hibar, D. P. et al. Cortical abnormalities in bipolar disorder: an MRI analysis of EB020403, by the ENIGMA World Aging Center (R56 AG058854), and by the 6503 individuals from the ENIGMA Bipolar Disorder Working Group. Mol. ENIGMA Sex Differences Initiative (R01 MH116147). Additional support was Psychiatry 23, 932–942 (2018). provided by grants to the ENIGMA-PGC PTSD Working Group (R01 MH111671; 3. Schmaal, L. et al. Cortical abnormalities in adults and adolescents with major PI: RAM), the ENIGMA-Addiction Working Group (R01 DA047119; to H.P.G. and depression based on brain scans from 20 cohorts worldwide in the ENIGMA P.J.C.), the ENIGMA Suicidal Thoughts and Behavior Working Group (R01 major depressive disorder working group. Mol. Psychiatry 22, 900–909 (2017). MH117601; to N.J. and L.S.), the ENIGMA Epilepsy Working Group (R01 4. Logue, M. W. et al. Smaller hippocampal volume in posttraumatic stress NS107739; to C.R.M.), a genotyping grant from the Australian NHMRC disorder: a multisite ENIGMA-PGC study: subcortical volumetry results from (APP1103623 and APP1158127; to SEM), a German federal grant to the ENIGMA posttraumatic stress disorder consortia. Biol. Psychiatry 83, 244–253 (2018). Task-Related fMRI Group (ER724/4-1 and WA1539/11-1; to H.W. and I.M.V.), a 5. Mackey, S. et al. Mega-analysis of gray matter volume in substance depen- Kavli Foundation Neuroscience without Borders seed grant (to N.J. and P.M.T.), dence: general and substance-specific regional effects. Am. J. Psychiatry 176, an NIH instrumentation grant (S10 OD023696 to P.K.), and K01 HD091283 (to S. 119–128 (2019). L.L.). We thank all scientists and participants in ENIGMA who made this work 6. Boedhoe, P. S. W. et al. Cortical abnormalities associated with pediatric and possible. A full list of ENIGMA Consortium current and past members can be adult obsessive-compulsive disorder: findings from the ENIGMA obsessive- found here http://enigma.ini.usc.edu/ongoing/members/. compulsive disorder working group. Am. J. Psychiatry 175, 453–462 (2018). 7. Hoogman, M. et al. Brain imaging of the cortex in ADHD: a coordinated analysis of large-scale clinical and population-based samples. Am. J. Psy- Conflict of interest chiatry. https://doi.org/10.1176/appi.ajp.2019.18091033 (2019). Individual authors’ disclosures and conflicts of interest are listed in 8. van Rooij, D. et al. Cortical and subcortical brain morphometry differences Supplementary Appendix C. between patients with autism spectrum disorder and healthy individuals Thompson et al. Translational Psychiatry (2020)10:100 Page 22 of 28 across the lifespan: results from the ENIGMA ASD Working Group. Am. J. 33. Thompson, P. M., Ge, T., Glahn, D. C., Jahanshad, N. & Nichols, T. E. Genetics of Psychiatry 175, 359–369 (2018). the connectome. Neuroimage 80, 475–488 (2013). 9. Whelan, C. D. et al. Structural brain abnormalities in the common epilepsies 34. Thompson, P. M., Hibar, D. P., Stein, J. L., Prasad, G. & Jahanshad, N. in Micro-, assessed in a worldwide ENIGMA study. Brain 141, 391–408 (2018). Meso- and Macro-Connectomics of the Brain (eds Kennedy, H., Van Essen, D. C. 10. Sun, D. et al. Large-scale mapping of cortical alterations in 22q11.2 deletion & Christen, Y. (Springer: Cham (CH), 2016). syndrome: convergence with idiopathic psychosis and effects of deletion 35. Holland, D. et al. Beyond SNP heritability: polygenicity and discoverability of size. Mol. Psychiatry https://doi.org/10.1038/s41380-018-0078-5 (2018). phenotypes estimated with a univariate gaussian mixture model. bioRxiv. 11. de Zwarte, S. M. C. et al. The association between familial risk and brain https://doi.org/10.1111/498550 (2018). abnormalities is disease-specific: an ENIGMA–relatives study of schizophrenia 36. Le, B. D. & Stein, J. L. Mapping causal pathways from genetics to neu- and bipolar disorder. Biol. Psychiatry 86, 545–556 (2019). ropsychiatric disorders using genome-wide imaging genetics: current status 12. Boedhoe, P. et al. Subcortical brain volume, regional cortical thickness and and future directions. Psychiatry Clin. Neurosci. https://doi.org/10.1111/ surface area variations across attention-deficit/hyperactivity disorder (ADHD), pcn.12839 (2019). autism spectrum disorder (ASD), and obsessive-compulsive disorder (OCD). 37. Franke, B. et al. Genetic influences on schizophrenia and subcortical brain Am. J. Psychiatry (2019) (submitted revision). Preprint at https://www.biorxiv. volumes: large-scale proof of concept. Nat. Neurosci. 19, 420–431 (2016). org/content/10.1101/673012v1. 38. Brouwer, R. M. et al. Genetic markers for brain plasticity. Presented at the 25th 13. Grasby, K. L. et al. The genetic architecture of the human cerebral cortex. Organization of Human Brain Mapping Annual Meeting Jun 9–13, abstract Science. (2020) (in press). Preprint at https://doi.org/10.1101/399402. T437 (Rome, Italy, 2019). 14. Satizabal, C. L. et al. Genetic architecture of subcortical brain structures in 39. Brouwer, R. M. et al. Genetic influences on individual differences in 38,854 individuals worldwide. Nat. Genet 51, 1624–1636 (2019). longitudinal changes in global and subcortical brain volumes: results of 15. Hofer, E. et al. Genetic determinants of cortical structure (thickness, surface the ENIGMA plasticity working group. Hum. Brain Mapp. 38, 4444–4458 area and volumes) among disease free adults in the CHARGE Consortium. (2017). bioRxiv. https://doi.org/10.1101/409649 (2019). 40. Smit, D. J. A. et al. Genome-wide association analysis links multiple psychiatric 16. Shin, J. et al. Planar cell polarity pathway and development of the human liability genes to oscillatory brain activity. Hum. Brain Mapp. 39, 4183–4195 visual cortex. bioRxiv. https://doi.org/10.1101/404558 (2018). (2018). 17. Villalón-Reina, J. E. et al. Altered white matter microstructure in 22q11.2 41. Smeland, O. B. et al. Genetic overlap between schizophrenia and volumes of deletion syndrome: A multi-site diffusion tensor imaging study. Mol. Psy- hippocampus, putamen, and intracranial volume indicates shared molecular chiatry. https://doi.org/10.1038/s41380-019-0450-0 (2019). genetic mechanisms. Schizophr. Bull. 44, 854–864 (2018). 18. Kong, X.-Z., ENIGMA Laterality Working Group, & Francks, F. An illustration of 42. Brainstorm Consortium, A. V. et al. Analysis of shared heritability in common reproducibility in human neuroscience in the absence of selective reporting. disorders of the brain. Science 360, https://doi.org/10.1126/science.aap8757. bioRxiv. Preprint at https://doi.org/10.1101/866301 (2019). (2018) 19. Jahanshad, N. et al. Do candidate genes affect the brain’s white matter 43. Lee, P. H. et al. Partitioning heritability analysis reveals a shared genetic microstructure? large-scale evaluation of 6,165 diffusion mri scans. bioRxiv. basis of brain anatomy and schizophrenia. Mol. Psychiatry 21, 1680–1689 https://doi.org/10.1101/107987 (2017). (2016). 20. Border, R. et al. No support for historical candidate gene or candidate gene- 44. Smeland, O. B. et al. Identification of genetic loci jointly influencing by-interaction hypotheses for major depression across multiple large sam- schizophrenia risk and the cognitive traits of verbal-numerical reasoning, ples. Am. J. Psychiatry 176, 376–387 (2019). reaction time, and general cognitive function. JAMA Psychiatry 74, 21. Farrell, M. S. et al. Evaluating historical candidate genes for schizophrenia. 1065–1075 (2017). Mol. Psychiatry 20, 555–562 (2015). 45. Hibar, D. P. et al. Significant concordance of genetic variation that 22. Kunkle, B. W. et al. Genetic meta-analysis of diagnosed Alzheimer’s disease increases both the risk for obsessive-compulsive disorder and the identifies new risk loci and implicates Aβ, tau, immunity and lipid processing. volumes of the nucleus accumbens and putamen. Br. J. Psychiatry 213, Nat. Genet 51, 414–430 (2019). 430–436 (2018). 23. Nunes, A. et al. Using structural MRI to identify bipolar disorders – 13 site 46. van der Merwe, C. et al. Concordance of genetic variation that increases risk machine learning study in 3020 individuals from the ENIGMA Bipolar Dis- for anxiety disorders and posttraumatic stress disorders and that influences orders Working Group. Mol. Psychiatry https://doi.org/10.1038/s41380-018- their underlying neurocircuitry. J. Affect Disord. 245, 885–896 (2019). 0228-9 (2018). 47. Klein, M. et al. Genetic markers of ADHD-related variations in intracranial 24. Stein, J. L. et al. Identification of common variants associated with human volume. Am. J. Psychiatry 176, 228–238 (2019). hippocampal and intracranial volumes. Nat. Genet. 44, 552–561 (2012). 48. Walton, E. et al. Exploration of shared genetic architecture between sub- 25. Hibar, D. P. et al. Common genetic variants influence human subcortical cortical brain volumes and anorexia nervosa. Mol. Neurobiol. https://doi.org/ brain structures. Nature 520, 224–229 (2015). 10.1007/s12035-018-1439-4 (2018). 26. Tilot, A. K. et al. Unearthing the evolutionary history of genetic variants 49. Mufford, M. et al. Concordance of genetic variation that increases risk for influencing human cortical surface area. biorXiv. https://doi.org/10.1101/ tourette syndrome and that influences its underlying neurocircuitry. Transl. 703793 (2019). Psychiatry 9, 120 (2019). 27. Tilot, A. K. et al. Unearthing the evolutionary history of genetic variants 50. Couvy-Duchesne, B. et al. A fast method for estimating statistical power of influencing human cortical surface area. in Abstract Presented at 49th multivariate GWAS in real case scenarios: examples from the field of imaging Annual Meeting of the Society for Neuroscience Nov 3–7 (San Diego, CA, genetics. Behav. Genet. 49, 112–121 (2019). 2018). 51. Lorenzi, M. et al. Susceptibility of brain atrophy to TRIB3 in Alzheimer’s 28. Medland, S. E., Jahanshad, N., Neale, B. M. & Thompson, P. M. Whole-genome disease, evidence from functional prioritization in imaging genetics. Proc. Natl analyses of whole-brain data: working within an expanded search space. Nat. Acad. Sci. USA 115, 3162–3167 (2018). Neurosci. 17, 791–800 (2014). 52. Jia, T. et al. Epigenome-wide meta-analysis of blood DNA methylation and its 29. Roshchupkin, G. V. et al. Heritability of the shape of subcortical brain struc- association with subcortical volumes: findings from the ENIGMA Epigenetics tures in the general population. Nat. Commun. 7, 13738 (2016). Working Group. Mol. Psychiatry https://doi.org/10.1038/s41380-019-0605-z 30. Adams, H. H. H. et al. Full exploitation of high dimensionality in brain ima- (2019). ging: The JPND working group statement and findings. Alzheimers Dement 53. Sønderby, I. E. et al. Dose response of the 16p11.2 distal copy number variant 11, 286–290 (2019). on intracranial volume and basal ganglia. Mol. Psychiatry https://doi.org/ 31. Roshchupkin, G. V. et al. One and a half million genome wide-association 10.1038/s41380-018-0118-1 (2018). studies of brain morphometry: a proof-of-concept study. Eur. Hum. Genet. 54. van Erp, T. G. M. et al. Subcortical brain volume abnormalities in 2028 indi- Conf. (ESHG) (2019) (submitted). viduals with schizophrenia and 2540 healthy controls via the ENIGMA con- 32. Jahanshad, N. et al. Genome-wide scan of healthy human connectome sortium. Mol. Psychiatry 21, 547–553 (2016). discovers SPON1 gene variant influencing dementia severity. Proc. Natl Acad. 55. Gutman, B. A. et al. A Meta-Analysis of deep brain structural shape Sci. USA 110, 4768–4773 (2013). abnormalities in 2,763 individuals with schizophrenia compared to 3,768 Thompson et al. Translational Psychiatry (2020)10:100 Page 23 of 28 healthy volunteers via the ENIGMA Consortium. Biol. Psychiatry (2020) 79. O’Leary, B. M., Hong, X., A MR, Israel, L. & Xin W. Development of cortical (submitted). vertex-based mega-analysis to study brain abnormalities in PTSD. Presented 56. Kelly, S. et al. Widespread white matter microstructural differences in schi- at Int. Soc. Traumatic Stress Studies Nov 8–10, abstract number #1–284 zophrenia across 4322 individuals: results from the ENIGMA Schizophrenia (Washington, DC, USA, 2018). DTI Working Group. Mol. Psychiatry 23, 1261–1269 (2018). 80. Salminen, L. et al. 108: Hippocampal subfield volumes relate to unique 57. Holleran, L. et al. The relationship between white matter microstructure and phenotypes of PTSD: International Analysis by the PGC-ENIGMA PTSD general cognitive ability in patients with schizophrenia and healthy partici- Working Group. Biol. Psychiatry, 85, S45 (2019). pants in the ENIGMA consortium. Am. J. Psychiatry (2019). 81. Salminen, L. et al. Hippocampal subregion abnormalities in current and 58. Okada, N. et al. Abnormal asymmetries in subcortical brain volume in schi- lifetime PTSD: International Analysis from the PGC-ENIGMA PTSD Working zophrenia. Mol. Psychiatry 21, 1460–1466 (2016). Group. in Abstract Presented at the 46th Annual Meeting of the International 59. Alnæs, D. et al. Brain heterogeneity in schizophrenia and its association Neuropsychological Society Feb 14–17 (Washington, DC, 2018). with polygenic risk. JAMA Psychiatry https://doi.org/10.1001/jamapsychiatry. 82. Saemann, P. et al. ENIGMA-MDD hippocampal subfield analysis of first epi- 2019.0257 (2019). sode and recurrent Major Depressive Disorder. Presented at the 23rd Orga- 60. Kochunov, P., Thompson, P. M. & Hong, L. E. Toward high reproducibility and nization of Human Brain Mapping Annual Meeting Jun 25–29, abstract 1254 accountable heterogeneity in schizophrenia research. JAMA Psychiatry (Vancouver, Canada, 2017). https://doi.org/10.1001/jamapsychiatry.2019.0208 (2019). 83. Saemann, P. G. et al. FreeSurfer based segmentation of hippocampal sub- 61. van Erp, T. G. M. et al. Reply to: New meta- and mega-analyses of magnetic fields: a review of methods and applications, with a novel quality control resonance imaging findings in schizophrenia: do they really increase our procedure and usefulness for ENIGMA studies and other collaborative efforts. knowledge about the nature of the disease process? Biol. Psychiatry 85, (2018) (submitted). e35–e39 (2019). 84. Mackey, S. et al. Genetic imaging consortium for addiction medicine: from 62. Walton, E. et al. Positive symptoms associate with cortical thinning in the neuroimaging to genes. Prog. Brain Res 224, 203–223 (2016). superior temporal gyrus via the ENIGMA Schizophrenia consortium. Acta 85. Chye, Y. et al. Subcortical surface morphometry in substance dependence: an Psychiatr. Scand. 135, 439–447 (2017). ENIGMA addiction working group study. Addiction Biol. e12830 (2019). 63. Walton, E. et al. Prefrontal cortical thinning links to negative symptoms 86. Conrod, P. et al. Cortical and subcortical differences between alcohol in schizophrenia via the ENIGMA consortium. Psychol. Med. 48, 82–94 dependent individuals and controls: meta analysis results from the ENIGMA- (2018). Addiction Working Group. Biol. Psychiatry 81, S41 (2017). 64. Smith, S. M. et al. Tract-based spatial statistics: voxelwise analysis of multi- 87. de Wit, S. J. et al. Multicenter voxel-based morphometry mega-analysis of subject diffusion data. Neuroimage 31, 1487–1505 (2006). structural brain scans in obsessive-compulsive disorder. Am. J. Psychiatry 171, 65. Jahanshad, N. et al. Multi-site genetic analysis of diffusion images and vox- 340–349 (2014). elwise heritability analysis: A pilot project of the ENIGMA-DTI working group. 88. Boedhoe, P. S. W. et al. Distinct subcortical volume alterations in pediatric Neuroimage 81, 455–469 (2013). and adult OCD: a worldwide meta- and mega-analysis. Am. J. Psychiatry 174, 66. Kochunov, P. et al. Integration of routine QA data into mega-analysis may 60–69 (2017). improve quality and sensitivity of multisite diffusion tensor imaging studies. 89. Bruin, W. B. et al. Structural neuroimaging biomarkers for obsessive- Hum. Brain Mapp. 39, 1015–1023 (2018). compulsive disorder in the ENIGMA-OCD consortium: medication matters. 67. van Velzen, L. S. et al. White matter disturbances in major depressive disorder: Preprint at https://doi.org/10.1101/19012567 (2019). a coordinated analysis across 20 international cohorts in the ENIGMA MDD 90. Kong, X.-Z. et al. Mapping cortical and subcortical asymmetry in obsessive- working group. Mol. Psychiatry https://doi.org/10.1038/s41380-019-0477-2 compulsive disorder: Findings from the ENIGMA Consortium. Biol. Psychiatry (2019). https://doi.org/10.1016/j.biopsych.2019.04.022 (2019). 68. Hibar, D. P. et al. Subcortical volumetric abnormalities in bipolar disorder. Mol. 91. Hoogman, M. et al. Subcortical brain volume differences in participants with Psychiatry 21, 1710–1716 (2016). attention deficit hyperactivity disorder in children and adults: a cross- 69. Favre, P. et al. Widespread white matter microstructural abnormalities in sectional mega-analysis. Lancet Psychiatry 4, 310–319 (2017). bipolar disorder: evidence from mega- and meta-analyses across 3033 92. Hess, J. L., Akutagava-Martins, G. C., Patak, J. D., Glatt, S. J. & Faraone, S. V. Why individuals. Neuropsychopharmacology 44, 2285–2293 (2019). is there selective subcortical vulnerability in ADHD? Clues from postmortem 70. Schmaal, L. et al. Subcortical brain alterations in major depressive disorder: brain gene expression data. Mol. Psychiatry 23, 1787–1793 (2018). findings from the ENIGMA Major Depressive Disorder working group. Mol. 93. Shaw, P. et al. A multicohort, longitudinal study of cerebellar development in Psychiatry 21, 806–812 (2016). attention deficit hyperactivity disorder. J. Child Psychol. Psychiatry 59, 71. de Kovel, C. G. F. et al. No alterations of brain structural asymmetry in Major 1114–1123 (2018). Depressive Disorder: An ENIGMA consortium analysis. Am. J. Psychiatry 176, 94. Zhang-James, Y. et al. Machine learning classification of attention-deficit/ 1039–1049 (2019). hyperactivity disorder using structural MRI data. bioRxiv. https://doi.org/ 72. Han, L. K. M. et al. Brain aging in major depressive disorder: results from the 10.1101/546671 (2019). ENIGMA Major Depressive Disorder working group. bioRxiv. https://doi.org/ 95. Gottesman, I. I. & Gould, T. D. The endophenotype concept in psy- 10.1101/560623 (2019). chiatry: etymology and strategic intentions. Am. J. Psychiatry 160, 73. Opel, N. et al. Brain structural abnormalities in obesity: Relation to age, 636–645 (2003). genetic risk, and common psychiatric disorders - Evidence through uni- 96. Flint, J. & Munafò, M. R. The endophenotype concept in psychiatric genetics. variate and multivariate mega-analysis including 6420 participants from the Psychol. Med. 37, 163–180 (2007). ENIGMA MDD working group. Mol. Psych (2020) (under review). 97. Postema, M. C. et al. Altered structural brain asymmetry in autism spectrum 74. Frodl, T. et al. Childhood adversity impacts on brain subcortical structures disorder in a study of 54 datasets. Nat. Commun. 10, 4658 (2019). relevant to depression. J. Psychiatr. Res. 86, 58–65 (2017). 98. Namazova-Baranova, L. et al. Cortical Morphometry in Gaucher Disease: 75. Tozzi, L. et al. Interactive impact of childhood maltreatment, depression, and Findings from the ENIGMA Storage Disease working group. Presented at the age on cortical brain structure: mega-analytic findings from a large multi-site 23rd Organization of Human Brain Mapping Annual Meeting Jun 25–29, cohort. Psychol. Med. 14, 1–12 (2019). abstract 1275 (Vancouver, Canada, 2017). 76. Dennis, E. L. et al. Altered white matter microstructural organization in 99. Namazova-Baranova, L. et al. Cortical Morphometry and White Matter posttraumatic stress disorder across 3,047 adults: results from the PGC- Integrity in Children with Hepatic Glycogen Storage Disease. Presented at the ENIGMA PTSD Consortium. Mol. J. Psychiatry https://doi.org/10.1038/s41380- 23rd Organization of Human Brain Mapping Annual Meeting Jun 25–29, 019-0631-x (2019). abstract 1277 (Vancouver, Canada, 2017). 77. Gopalkumar, R. A structural covariance network analyses of cortical thickness 100. van der Meer, D. et al. Association of copy number variation of the 15q11.2 and surface area in posttraumatic stress disorder (PTSD) from the ENIGMA region with cortical and subcortical morphology and cognition. JAMA Psy- Worldwide Consortium. (2020) (in preparation). chiatry https://doi.org/10.1001/jamapsychiatry.2019.3779 (2019). 78. O’Leary, B. M., Xie, C. G. & Angstadt, J. T. Development of cortical vertex-based 101. Ching, C. R. K., Thompson, P. M. & Bearden, C. E., for the ENIGMA 22q11.2 mega-analysis to study brain abnormalities in PTSD. Int. Soc. Traumatic Stress Deletion Syndrome Working Group. Convergent subcortical brain alterations Studies (2019) (submitted). in 22q11.2 deletion syndrome and schizophrenia. Presented at the 24th Thompson et al. Translational Psychiatry (2020)10:100 Page 24 of 28 Organization of Human Brain Mapping Annual Meeting Jun 17–21, abstract Working Group. Brain Imaging Behav. (Special Issue on ENIGMA Brain Injury). 1569 (Singapore, 2018). Preprint at https://doi.org/10.31234/osf.io/gesvh (2019) (submitted). 102. Stefansson, H. et al. CNVs conferring risk of autism or schizophrenia affect 122. van der Werf, Y. et al. International mega-analysis of cortical and subcortical cognition in controls. Nature 505, 361–366 (2014). morphometry in Parkinson’s Disease: ENIGMA-PD. Presented at the 14th 103. Crawford, K. et al. Medical consequences of pathogenic CNVs in adults: International Conference of Alzheimer’s and Parkinson’s Diseases Mar 26–31, analysis of the UK Biobank. J. Med. Genet. 56, 131–138 (2019). oral presentation at symposium 35 (Lisbon, Portugal, 2019). 104. King, J. A., Frank, G. K. W., Thompson, P. M. & Ehrlich, S. Structural neuroi- 123. van der Werf, Y. D. et al. ENIGMA-Parkinson’s Disease: An International mega- maging of anorexia nervosa: future directions in the quest for mechanisms analysis of cortical and subcortical morphometry in Parkinson’s patients underlying dynamic alterations. Biol. Psychiatry 83, 224–234 (2018). versus healthy controls. Presented at the 25th Organization of Human Brain 105. Bas-Hoogendam, J. M. et al. Voxel-based morphometry multi-center mega- Mapping Annual Meeting Jun 9–13, abstract W273 (Rome, Italy, 2019). analysis of brain structure in social anxiety disorder. Neuroimage Clin. 16, 124. Nir, T. M. et al. Smaller limbic brain volumes are associated with greater 678–688 (2017). immunosuppression in over 1000 HIV-infected adults across five continents: 106. Bas-Hoogendam, J. M. et al. Subcortical brain volumes, cortical thickness findings from the ENIGMA-HIV Working Group. Preprint at https://doi.org/ and cortical surface area in families genetically enriched for social anxiety 10.1101/724583 (2019). disorder – A multiplex multigenerational neuroimaging study. EBioMe- 125. Cole, J. H. et al. Increased brain-predicted aging in treated HIV disease. dicine 36, 410–428 (2018). Neurology 88, 1349–1357 (2017). 107. Groenewold, N. et al. F27. Subcortical volumes in social anxiety disorder: 126. Liew, S.-L. et al. A large, open source dataset of stroke anatomical brain preliminary results from enigma-anxiety. Biol. Psychiatry 83, S247–S248 images and manual lesion segmentations. Sci. Data 5, 180011 (2018). (2018). 127. Ito, K. L., Kumar, A., Zavaliangos-Petropulu, A., Cramer, S. C. & Liew, S.-L. 108. Shiroishi, M. S. et al. Brain cortical structural differences between non-central Pipeline for Analyzing Lesions After Stroke (PALS). Front. Neuroinform. 12, 63 nervous system cancer patients treated with and without chemotherapy (2018). compared to non-cancer controls: a cross-sectional pilot MRI study using 128. Ito, K. L., Kim, H. & Liew, S.-L. A comparison of automated lesion segmen- clinically-indicated scans. Proc. SPIE Int. Soc. Opt. Eng. 10572 https://doi.org/ tation approaches for chronic stroke T1‐weighted MRI data. Hum. Brain 10.1117/12.2285971 (2017). Mapp. 40, 4669–4685 (2019). 109. Shiroishi, M. S. et al. Brain morphometry in prostate cancer survivors from the 129. Liew, S.-L. et al. Abstract TMP48: Subcortical volumes associated with post- ENIGMA Cancer & Chemotherapy Working Group. Presented at the 24th stroke motor performance vary across impairment severity, time since stroke, Organization of Human Brain Mapping Annual Meeting Jun 17–21, abstract and lesion laterality: an ENIGMA Stroke Recovery Analysis. Stroke 49(Suppl_1) 1321 (Singapore, 2018). https://doi.org/10.1161/str.49.suppl_1.TMP48 (2018). 110. Altmann, A. et al. A systems-level analysis highlights microglial activation as a 130. Liew, S.-L. et al. Abstract 14: Effects of lesion laterality on post-stroke motor modifying factor in common forms of human epilepsy. bioRxiv. https://doi. performance: An ENIGMA Stroke Recovery Analysis. Stroke 48(Suppl 1) org/10.1101/470518 (2018). https://www.ahajournals.org/doi/10.1161/str.48.suppl_1.14 (2017). 111. Wilde, E. A., Dennis, E. L. & Tate, D. F. The ENIGMA Brain Injury Working Group: 131. Ito, K. L., Zavaliangos-Petropulu, A., Cramer, S. C. & Liew, S.-L. Effective con- Approach, Challenges, and Potential Benefits. Brain Imaging Behav. (Special nectivity of the ipsilesional action observation network after stroke. Presented Issue on ENIGMA Brain Injury). Preprint at https://doi.org/10.31234/osf.io/ at the 24th Organization of Human Brain Mapping Annual Meeting Jun 17–21, t96xb (2019) (submitted). abstract (Singapore, 2018). 112. Olsen, A. et al. Toward a Global and Reproducible Science for Brain Imaging 132. Zavaliangos-Petropulu, A., Jahanshad, N., Thompson, P. M. & Liew, S.-L. in Neurotrauma: The ENIGMA Adult Moderate/Severe Traumatic Brain Injury Evaluating stroke lesion overlap with subcortical structures and post-stroke Working Group. Brain Imaging Behav. (Special Issue on ENIGMA Brain Injury). motor performance. Presented at the 25th Organization of Human Brain Preprint at https://doi.org/10.31234/osf.io/jnsb2. (2019) (submitted). Mapping Annual Meeting Jun 9–13, abstract M203 (Rome, Italy, 2019). 113. Dennis, E. L. et al. ENIGMA pediatric msTBI: preliminary results from meta- 133. Zavaliangos-Petropulu, A. et al. Subcortical brain shape differences relate to analysis of diffusion MRI. in14th International Symposium on Medical Infor- post-stroke motor behavior. Presented at the American Society of Neuroreh- mation Processing and Analysis. (International Society for Optics and Photo- abilitation Conference Annual Meeting Nov 9–10, abstract F12 (Baltimore, MD, nics, 2018). 2017). 114. Dennis, E. L. et al. Brain Imaging in Young TBI Patients: A Coordinated Effort 134. Kochunov, P. et al. Heterochronicity of white matter development and aging Towards Individualized Predictors from the ENIGMA Pediatric msTBI Group. explains regional patient control differences in schizophrenia. Hum. Brain Brain Imaging Behav. (Special Issue on ENIGMA Brain Injury). Preprint at Mapp. 37, 4673–4688 (2016). https://doi.org/10.31234/osf.io/y2txh (2019) (submitted). 135. Nir, T. M. et al. Effects of Diffusion MRI model and harmonization on the 115. Dennis, E. L. et al. ENIGMA military brain injury: a coordinate meta-analysis of consistency of findings in an international multi-cohort HIV neuroimaging diffusion MRI from multiple cohorts. Proc. IEEE Int Symp. Biomed. Imaging study. In: Bonet-Carne, E., Grussu, F., Ning, L., Sepehrband, F. & Tax, C. (eds) 2018, 1386–1389 (2018). Computational Diffusion MRI. MICCAI 2019. Mathematics and Visualization. 116. Dennis, E. L. et al. ENIGMA Military Brain Injury: A Preliminary Meta-Analysis (Springer, Cham, 2019) https://doi.org/10.1007/978-3-030-05831-9_17. of Diffusion MRI Measures. Presented at the 70th Annual Meeting of the 136. Piras, F. et al. Selective white matter microstructure and its relation to clinical American Academy of Neurology Apr 21–27, abstract S46.006 (Los Angeles, features of obsessive-compulsive disorder: findings from the ENIGMA OCD CA, 2018). Working Group. bioRxiv. https://doi.org/10.1101/855916 (2019). 117. Dennis, E. L. et al. ENIGMA Military Brain Injury: Altered Subcortical Volume 137. Ho, T. C. et al. Subcortical shape alterations in major depressive disorder: Revealed by Mega-Analysis. Presented at the 71st Annual Meeting of the findings from the ENIGMA Major Depressive Disorder Working Group. Hum. American Academy of Neurology May 4–11, abstract (Philadelphia, PA, 2018). Brain Mapp. Preprint at https://doi.org/10.1101/534370 (2019) (submitted). 118. Tate, D. F. et al. Coordinating global multi-site studies of military-relevant 138. Petrov, D. et al. Machine learning for large-scale quality control of 3D shape traumatic brain injury: Opportunities, challenges, and harmonization guide- models in neuroimaging. bioRxiv. https://doi.org/10.1101/166496 (2017). lines. Brain Imaging Behav. (Special Issue on ENIGMA Brain Injury). Preprint at 139. Petrov, D. et al. Deep Learning for Quality Control of Subcortical Brain 3D https://doi.org/10.31234/osf.io/d4qs8 (2019) (submitted). Shape Modelsin Shape in Medical Imaging. 268–276 (Springer International 119. Koerte, I. K. et al. Neuroimaging of Sport-Related Brain Injury: Challenges and Publishing, 2018). Recommendations from the ENIGMA Sports-Related Brain Injury group. Brain 140. Fouche, J.-P. et al. Shape analysis of subcortical structures in obsessive Imaging and Behavior (Special Issue on ENIGMA Brain Injury) (2019) (in compulsive disorder: a multi-site analysis of the OCD Brain Imaging Con- preparation). sortium. (2020) (in preparation). 120. Esopenko, C. et al. Harmonization of Measures to Assess IPV-Related Head 141. Ching, C. R. K. et al. Subcortical Shape and Volumetric Findings from ENIGMA Trauma: Recommendations from the ENIGMA IPV Working Group. Brain 22q11.2 Working Group (N = 778). Presented at the 23rd Organization of Imaging Behav. (Special Issue on ENIGMA Brain Injury). Preprint at https://doi. Human Brain Mapping Annual Meeting Jun 25–29, abstract 3318 (Vancouver, org/10.31234/osf.io/3xzwg (2019) (submitted). Canada, 2017). 121. Bartnik-Olson, B. et al. The Clinical Utility Of Magnetic Resonance Spectro- 142. Ching, C. R. K., Gutman, B. A., Hibar, D. P., Thompson, P. M. & Andreassen, O. A. scopy In Traumatic Brain Injury: Recommendations from the ENIGMA MRS for the ENIGMA Bipolar Disorder Working Group. Subcortical Shape Analysis Thompson et al. Translational Psychiatry (2020)10:100 Page 25 of 28 from the ENIGMA Bipolar Disorder Working Group (N = 3,028). Presented at 161. Adhikari, B. M. et al. Heritability estimates on resting state fMRI data using the 23rd Organization of Human Brain Mapping Annual Meeting Jun 25–29, ENIGMA analysis pipeline. Pac. Symp. Biocomput. 23, 307–318 (2018). abstract 1205 (Vancouver, Canada, 2017). 162. Adhikari, B. M. et al. Effects of ketamine and midazolam on resting state 143. Yun, J.-Y. et al. Brain structural covariance networks in obsessive-compulsive connectivity and comparison with ENIGMA connectivity deficit patterns in disorder: a graph analysis from the ENIGMA Consortium. Brain 143, 684–700 schizophrenia. Hum. Brain Mapp. 41, 767–778 (2020). (2020). 163. Veer, I. M., Waller, L., Lett, T. A. P., Erk, S. & Walter, H. ENIGMA task-based fMRI: A 144. Pizzagalli, F. et al. The reliability and heritability of cortical folds and their workgroup studying the genetic basis of task-evoked brain activity. Presented genetic correlations across hemispheres. bioRxiv. https://doi.org/10.1101/ at the 25th Organization of Human Brain Mapping Annual Meeting Jun 9–13, 795591 (2019). abstract W293 (Rome, Italy, 2019). 145. Haukvik, U. K. et al. In vivo hippocampal subfield volumes in bipolar 164. Plis, S. M. et al. COINSTAC: a privacy enabled model and prototype for disorder – a mega-analysis from the ENIGMA consortium. (2020) (in leveraging and processing decentralized brain imaging data. Front. Neurosci. preparation). 10, 365 (2016). 146. Button, K. S. Double-dipping revisited. Nat. Neurosci. 22, 688–690 (2019). 165. Boedhoe, P. S. W. et al. An empirical comparison of meta- and mega-analysis 147. Button, K. S. et al. Power failure: why small sample size undermines the with data from the ENIGMA obsessive-compulsive disorder working group. reliability of neuroscience. Nat. Rev. Neurosci. 14, 365–376 (2013). Front. Neuroinf. 12, 102 (2019). 148. Ioannidis, J. P. A. & Khoury, M. J. Evidence-based medicine and big genomic 166. Reardon, S. The largest study involving transgender people is providing long- data. Hum. Mol. Genet. 27, R2–R7 (2018). sought insights about their health. Nature 568, 446–449 (2019). 149. Poldrack, R. A. et al. Scanning the horizon: towards transparent and repro- 167. Mueller, S., Thompson, P. & Luders E., for The Enigma Transgender Persons ducible neuroimaging research. Nat. Rev. Neurosci. 18, 115–126 (2017). Working Group. An initiative to combine MRI data in transgender persons to 150. Hallquist, M. N. & Hillary, F. G. Graph theory approaches to functional network examine structural brain differences: preliminary findings from the ENIGMA organization in brain disorders: a critique for a brave new small-world. Netw. transgender persons working group. Presented at the 3rd biennial European Neurosci. 3, 1–26 (2019). Professional Association for Transgender Health Conference Apr 11–13, oral 151. Gorgolewski, K. J. et al. NeuroVault.org: a web-based repository for collecting presentation (Rome, Italy, 2019). and sharing unthresholded statistical maps of the human brain. Front. 168. Aristotle. Metaphysics, book 2 (~350 BCE). Neuroinform. 9, 8 (2015). 169. Adams, H. H. H. et al. Novel genetic loci underlying human intracranial 152. Poldrack, R. A. & Gorgolewski, K. J. OpenfMRI: open sharing of task fMRI data. volume identified through genome-wide association. Nat. Neurosci. 19, Neuroimage 144, 259–261 (2017). 1569–1582 (2016). 153. Zuo, X.-N. et al. An open science resource for establishing reliability and 170. Nir, T. M. et al. CD4 counts predict brain white matter integrity in people reproducibility in functional connectomics. Sci. Data 1, 140049 (2014). living with HIV: A meta-analysis by the ENIGMA HIV working group. Presented 154. Kong, X.-Z. et al. Mapping cortical brain asymmetry in 17,141 healthy indi- at the Cognitive Neuroscience Society (CNS) Annual Meeting Mar 28–31, viduals worldwide via the ENIGMA Consortium. Proc. Natl Acad. Sci. USA 115, abstract D104 (San Francisco, CA, 2015). E5154–E5163 (2018). 171. Fouche, J. P. et al. A meta-analysis by the ENIGMA-HIV working group: CD4 155. Salminen, L. E. et al. Hippocampal Subfields in PTSD: Preliminary Results from counts predict subcortical volume loss in HIV-positive individuals. Presented at the ENIGMA PTSD Working Group. Presented at the 23rd Organization of the 21st Organization for Human Brain Mapping Annual Meeting Jun 14–18, Human Brain Mapping Annual Meeting Jun 25–29, abstract 1138 (Vancouver, abstract 3271 (Honolulu, HI, 2015). Canada, 2017). 172. Renteria, M. E. et al. Subcortical brain structure and suicidal behaviour in 156. Elliott, L. T. et al. Genome-wide association studies of brain imaging phe- major depressive disorder: a meta-analysis from the ENIGMA-MDD working notypes in UK Biobank. Nature 562, 210–216 (2018). group. Transl. Psychiatr 7, e1116 (2017). 157. Ferguson, A. R., Nielson, J. L., Cragin, M. H., Bandrowski, A. E. & Martone, M. E. 173. Hibar, D. et al. Novel genetic loci associated with hippocampal volume. Nat. Big data from small data: data-sharing in the ‘long tail’ of neuroscience. Nat. Commun. 8, 13624 (2017). Neurosci. 17, 1442–1447 (2014). 174. Guadalupe, T. et al. Human subcortical brain asymmetries in 15,847 people 158. Milham, M. P. et al. An open resource for non-human primate imaging. worldwide reveal effects of age and sex. Brain Imaging Behav. 11, 1497–1514 Neuron 100, 61–74.e2 (2018). (2017). 159. Palk, A., Illes, J., Thompson, P. M. & Stein, D. J. Ethical issues in global imaging 175. Dima, D. et al. Subcortical volumes trajectories across the lifespan: Data from genetics collaborations. NeuroImage (2019) (submitted). 18,605 healthy individuals aged 3–90 years. Hum Brain Mapp (2020) 160. Adhikari, B. M. et al. Comparison of heritability estimates on resting state fMRI (submitted). connectivity phenotypes using the ENIGMA analysis pipeline. Hum. Brain 176. Frangou, S. et al. Subcortical volumes across the lifespan: Normative data Mapp. 39, 4893–4902 (2018). from 10,144 individuals aged 3–90 years. Hum Brain Mapp (2020) (submitted). Author details Paul M. Thompson1, Neda Jahanshad 1, Christopher R. K. Ching1, Lauren E. Salminen 1, Sophia I. Thomopoulos 1, Joanna Bright1, Bernhard T. Baune2,3,4, Sara Bertolín5, Janita Bralten6,7, Willem B. Bruin8, Robin Bülow9, Jian Chen10, Yann Chye11, Udo Dannlowski2, Carolien G. F. de Kovel12,13, Gary Donohoe14, Lisa T. Eyler15,16, Stephen V. Faraone17, Pauline Favre18,19, Courtney A. Filippi20, Thomas Frodl21,22,23, Daniel Garijo24, Yolanda Gil24,25, Hans J. Grabe26,27, Katrina L. Grasby28, Tomas Hajek29,30, Laura K. M. Han31, Sean N. Hatton32,33, Kevin Hilbert34, Tiffany C. Ho35,36, Laurena Holleran14, Georg Homuth37, Norbert Hosten9, Josselin Houenou18,19,38, Iliyan Ivanov39, Tianye Jia40,41,42, Sinead Kelly43,44, Marieke Klein6,7,45, Jun Soo Kwon46,47, Max A. Laansma48, Jeanne Leerssen49, Ulrike Lueken34, Abraham Nunes29,50, Joseph O’ Neill51, Nils Opel2, Fabrizio Piras52, Federica Piras52, Merel C. Postema13, Elena Pozzi53,54, Natalia Shatokhina1, Carles Soriano-Mas5,55,56, Gianfranco Spalletta52,57, Daqiang Sun58,59, Alexander Teumer60, Amanda K. Tilot1, Leonardo Tozzi35, Celia van der Merwe61,62, Eus J. W. Van Someren49,63, Guido A. van Wingen8, Henry Völzke60,64, Esther Walton65, Lei Wang66,67, Anderson M. Winkler20, Katharina Wittfeld26,27, Margaret J. Wright68,69, Thompson et al. Translational Psychiatry (2020)10:100 Page 26 of 28 Je-Yeon Yun70,71, Guohao Zhang72, Yanli Zhang-James17,73, Bhim M. Adhikari74, Ingrid Agartz75,76,77, Moji Aghajani78,79, André Aleman80, Robert R. Althoff81, Andre Altmann82, Ole A. Andreassen75,83, David A. Baron84, Brenda L. Bartnik-Olson85, Janna Marie Bas-Hoogendam86,87,88, Arielle R. Baskin-Sommers89, Carrie E. Bearden58,90, Laura A. Berner39, Premika S. W. Boedhoe78, Rachel M. Brouwer45, Jan K. Buitelaar91, Karen Caeyenberghs92, Charlotte A. M. Cecil93,94, Ronald A. Cohen95,96, James H. Cole97,98, Patricia J. Conrod99, Stephane A. De Brito100, Sonja M. C. de Zwarte45, Emily L. Dennis1,101,102, Sylvane Desrivieres103, Danai Dima104,105, Stefan Ehrlich106, Carrie Esopenko107, Graeme Fairchild65, Simon E. Fisher7,13, Jean-Paul Fouche108,109, Clyde Francks7,13, Sophia Frangou110,111, Barbara Franke6,7,112, Hugh P. Garavan113, David C. Glahn114,115, Nynke A. Groenewold108, Tiril P. Gurholt75,83, Boris A. Gutman116,117, Tim Hahn2, Ian H. Harding118, Dennis Hernaus119, Derrek P. Hibar120, Frank G. Hillary121,122, Martine Hoogman6,7, Hilleke E. Hulshoff Pol45, Maria Jalbrzikowski123, George A. Karkashadze124, Eduard T. Klapwijk86,88, Rebecca C. Knickmeyer125,126,127, Peter Kochunov74, Inga K. Koerte102,128, Xiang-Zhen Kong13, Sook-Lei Liew129,130, Alexander P. Lin131,132, Mark W. Logue133,134,135, Eileen Luders136,137, Fabio Macciardi138, Scott Mackey113, Andrew R. Mayer139, Carrie R. McDonald32,140, Agnes B. McMahon1,141, Sarah E. Medland28, Gemma Modinos105,142, Rajendra A. Morey143,144, Sven C. Mueller145,146, Pratik Mukherjee147, Leyla Namazova-Baranova124,148, Talia M. Nir1, Alexander Olsen149,150, Peristera Paschou151, Daniel S. Pine152, Fabrizio Pizzagalli1, Miguel E. Rentería153, Jonathan D. Rohrer154, Philipp G. Sämann155, Lianne Schmaal54,156, Gunter Schumann42,157, Mark S. Shiroishi1,158, Sanjay M. Sisodiya159,160, Dirk J. A. Smit8, Ida E. Sønderby75,83,161, Dan J. Stein162, Jason L. Stein163, Masoud Tahmasian164, David F. Tate165,166, Jessica A. Turner167, Odile A. van den Heuvel48,78, Nic J. A. van der Wee87,88, Ysbrand D. van der Werf48, Theo G. M. van Erp168,169, Neeltje E. M. van Haren45,93, Daan van Rooij170, Laura S. van Velzen54,156, Ilya M. Veer171, Dick J. Veltman78, Julio E. Villalon-Reina1, Henrik Walter171, Christopher D. Whelan172,173, Elisabeth A. Wilde101,174,175, Mojtaba Zarei164, Vladimir Zelman176,177 & for the ENIGMA Consortium 1 Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Marina del Rey, CA, USA. 2Department of Psychiatry, University of Münster, Münster, Germany. 3Department of Psychiatry, The University of Melbourne, Melbourne, VIC, Australia. 4The Florey Institute of Neuroscience and Mental Health, The University of Melbourne, Melbourne, VIC, Australia. 5Department of Psychiatry, Bellvitge University Hospital, Bellvitge Biomedical Research Institute-IDIBELL, Barcelona, Spain. 6Department of Human Genetics, Radboud University Medical Center, Nijmegen, The Netherlands. 7Donders Institute for Brain, Cognition and Behaviour, Radboud University, Nijmegen, The Netherlands. 8Department of Psychiatry, Amsterdam UMC, University of Amsterdam, Amsterdam Neuroscience, Amsterdam, The Netherlands. 9Institute for Diagnostic Radiology and Neuroradiology, University Medicine Greifswald, Greifswald, Germany. 10Department of Computer Science and Engineering, The Ohio State University, Columbus, OH, USA. 11Turner Institute for Brain and Mental Health, School of Psychological Sciences, Monash University, Clayton, VIC, Australia. 12Biometris Wageningen University and Research, Wageningen, The Netherlands. 13Language & Genetics Department, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands. 14The Center for Neuroimaging and Cognitive Genomics, School of Psychology, National University of Ireland, Galway, Ireland. 15Department of Psychiatry, University of California, San Diego, La Jolla, CA, USA. 16Desert-Pacific Mental Illness Research, Education, and Clinical Center, VA San Diego Healthcare System, San Diego, CA, USA. 17 Departments of Psychiatry and of Neuroscience and Physiology, SUNY Upstate Medical University, Syracuse, NY, USA. 18INSERM Unit 955 Team 15 ‘Translational Psychiatry’, Créteil, France. 19NeuroSpin, UNIACT Lab, Psychiatry Team, CEA Saclay, Gif-Sur-Yvette, France. 20National Institute of Mental Health, National of Health, Bethesda, MD, USA. 21Department of Psychiatry and Psychotherapy, Otto von Guericke University Magdeburg, Magdeburg, Germany. 22Department of Psychiatry, Trinity College Dublin, Dublin, Ireland. 23German Center for Neurodegenerative Diseases (DZNE), Magdeburg, Germany. 24Information Sciences Institute, University of Southern California, Marina del Rey, CA, USA. 25Department of Computer Science, University of Southern California, Los Angeles, CA, USA. 26Department of Psychiatry and Psychotherapy, University Medicine Greifswald, Greifswald, Germany. 27German Center for Neurodegenerative Diseases (DZNE), Site Rostock/ Greifswald, Greifswald, Germany. 28Psychiatric Genetics, QIMR Berghofer Medical Research Institute, Brisbane, QLD, Australia. 29Department of Psychiatry, Dalhousie University, Halifax, NS, Canada. 30National Institute of Mental Health, Klecany, Czech Republic. 31Department of Psychiatry, Amsterdam University Medical Centers, VU University Medical Center, GGZ inGeest, Amsterdam Neuroscience, Amsterdam, The Netherlands. 32Center for Multimodal Imaging and Genetics, University of California, San Diego, La Jolla, CA, USA. 33Brain and Mind Centre, University of Sydney, Sydney, Australia. 34Department of Psychology, Humboldt-Universität zu Berlin, Berlin, Germany. 35Department of Psychiatry & Behavioral Sciences, Stanford University, Stanford, CA, USA. 36Department of Psychiatry & Weill Institute for Neurosciences, University of California, San Francisco, San Francisco, CA, USA. 37Interfaculty Institute for Genetics and Functional Genomics, University Medicine Greifswald, Greifswald, Germany. 38APHP, Mondor University Hospitals, School of Medicine, DMU Impact, Psychiatry Department, Créteil, France. 39Icahn School of Medicine at Mount Sinai, New York, NY, USA. 40Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai, China. 41MOE Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence, Fudan University, Shanghai, China. 42Centre for Population Neuroscience and Precision Medicine (PONS), MRC SGDP Centre, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK. 43Department of Psychiatry, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA. 44Department of Psychiatry, Brigham and Women’s Hospital, Boston, MA, USA. 45Department of Psychiatry, UMC Brain Center, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands. 46Department of Psychiatry, Seoul National University College of Medicine, Seoul, Republic of Korea. 47Department of Brain and Cognitive Sciences, Seoul National University College of Natural Sciences, Seoul, Republic of Korea. 48Department of Anatomy & Neurosciences, Amsterdam UMC, Location VUmc, Amsterdam Neuroscience, Amsterdam, The Netherlands. 49 Department of Sleep and Cognition, Netherlands Institute for Neuroscience, Amsterdam, The Netherlands. 50Faculty of Computer Science, Dalhousie University, Halifax, NS, Canada. 51Child & Adolescent Psychiatry, University of California, Los Angeles, Los Angeles, CA, USA. 52Laboratory of Neuropsychiatry, IRCCS Santa Lucia Foundation, Rome, Italy. 53Melbourne Neuropsychiatry Centre, Department of Psychiatry, The University of Melbourne, Melbourne, VIC, Australia. 54Orygen, The National Centre of Excellence in Youth Mental Health, Melbourne, VIC, Australia. 55CIBERSAM-G17, Madrid, Spain. 56Department of Psychobiology and Methodology Thompson et al. Translational Psychiatry (2020)10:100 Page 27 of 28 in Health Sciences, Universitat Autònoma de Barcelona, Barcelona, Spain. 57Department of Psychiatry and Behavioral Sciences, Baylor College of Medicine, Houston, TX, USA. 58Department of Psychiatry and Biobehavioral Sciences, Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, Los Angeles, CA, USA. 59Department of Mental Health, Veterans Affairs Greater Los Angeles Healthcare System, Los Angeles, CA, USA. 60Institute for Community Medicine, University Medicine Greifswald, Greifswald, Germany. 61Stanley Center for Psychiatric Research, The Broad Institute, Cambridge, MA, USA. 62Analytic and Translational Genetics Unit, Massachusetts General Hospital, Boston, MA, USA. 63Psychiatry and Integrative Neurophysiology, VU University, Amsterdam UMC, Amsterdam, The Netherlands. 64German Centre for Cardiovascular Research, Partner Site Greifswald, Greifswald, Germany. 65Department of Psychology, University of Bath, Bath, UK. 66Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medicine, Chicago, IL, USA. 67Radiology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA. 68Queensland Brain Institute, University of Queensland, Brisbane, QLD, Australia. 69Centre for Advanced Imaging, University of Queensland, Brisbane, QLD, Australia. 70Seoul National University Hospital, Seoul, Republic of Korea. 71Yeongeon Student Support Center, Seoul National University College of Medicine, Seoul, Republic of Korea. 72Department of Computer Science and Electrical Engineering, University of Maryland, Baltimore County, MD, USA. 73Department of Psychiatry and Behavioral Sciences, SUNY Upstate Medical University, Syracuse, NY, USA. 74Department of Psychiatry, University of Maryland School of Medicine, Baltimore, MD, USA. 75Norwegian Centre for Mental Disorders Research (NORMENT), Division of Mental Health & Addiction, Institute of Clinical Medicine, University of Oslo, Oslo, Norway. 76Department of Clinical Neuroscience, Centre for Psychiatric Research, Karolinska Institutet, Stockholm, Sweden. 77Department of Psychiatric Research, Diakonhjemmet Hospital, Oslo, Norway. 78Department of Psychiatry, Amsterdam UMC, Location VUmc, Amsterdam Neuroscience, Amsterdam, The Netherlands. 79Department of Research & Innovation, GGZ InGeest, Amsterdam, The Netherlands. 80 University of Groningen, University Medical Center Groningen, Groningen, The Netherlands. 81Psychiatry, Pediatrics, and Psychological Sciences, University of Vermont, Burlington, VT, USA. 82Centre for Medical Image Computing (CMIC), Department of Medical Physics and Biomedical Engineering, University College London, London, UK. 83Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway. 84Provost and Senior Vice President, Western University of Health Sciences, Pomona, CA, USA. 85Department of Radiology, Loma Linda University Medical Center, Loma Linda, CA, USA. 86Institute of Psychology, Leiden University, Leiden, The Netherlands. 87Department of Psychiatry, Leiden University Medical Center, Leiden, The Netherlands. 88Leiden Institute for Brain and Cognition, Leiden, The Netherlands. 89Department of Psychology, Yale University, New Haven, CT, USA. 90Department of Psychology, University of California, Los Angeles, CA, USA. 91Department of Cognitive Neuroscience, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, The Netherlands. 92Cognitive Neuroscience Unit, School of Psychology, Deakin University, Burwood, VIC, Australia. 93Department of Child and Adolescent Psychiatry/Psychology, Erasmus Medical Centre, Rotterdam, The Netherlands. 94Department of Epidemiology, Erasmus Medical Centre, Rotterdam, The Netherlands. 95Center for Cognitive Aging and Memory, University of Florida, Gainesville, FL, USA. 96Clinical and Health Psychology, Gainesville, FL, USA. 97Centre for Medical Image Computing (CMIC), Department of Computer Science, University College London, London, UK. 98Dementia Research Centre, Institute of Neurology, University College London, London, UK. 99Universite de Montreal, Centre de Recherche CHU Ste-Justine, Montreal, QC, Canada. 100School of Psychology and Centre for Human Brain Health, University of Birmingham, Birmingham, UK. 101Department of Neurology, University of Utah, Salt Lake City, UT, USA. 102Psychiatry Neuroimaging Laboratory, Brigham & Women’s Hospital, Harvard Medical School, Boston, MA, USA. 103Social, Genetic & Developmental Psychiatry Centre, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK. 104Department of Psychology, School of Arts and Social Sciences, City, University of London, London, UK. 105Department of Neuroimaging, Institute of Psychology, Psychiatry and Neurosciences, King’s College London, London, UK. 106Division of Psychological and Social Medicine and Developmental Neurosciences, Faculty of Medicine, TU Dresden, Dresden, Germany. 107Department of Rehabilitation and Movement Sciences, School of Health Professions, Rutgers Biomedical Health Sciences, Newark, NJ, USA. 108Department of Psychiatry and Mental Health, University of Cape Town, Cape Town, South Africa. 109SU/UCT MRC Unit on Risk & Resilience in Mental Disorders, University of Stellenbosch, Stellenbosch, South Africa. 110 Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 111University of British Columbia, Vancouver, Canada. 112Department of Psychiatry, Radboud University Medical Center, Nijmegen, The Netherlands. 113Department of Psychiatry, University of Vermont, Burlington, VT, USA. 114Department of Psychiatry, Boston Children’s Hospital and Harvard Medical School, Boston, MA, USA. 115Olin Neuropsychiatric Research Center, Institute of Living, Hartford, CT, USA. 116Biomedical Engineering, Illinois Institute of Technology, Chicago, IL, USA. 117Institute for Information Transmission Problems, Kharkevich Institute, Moscow, Russian Federation. 118Turner Institute for Brain and Mental Health & School of Psychological Sciences, Monash University, Melbourne, VIC, Australia. 119Department of Psychiatry & Neuropsychology, School for Mental Health and Neuroscience, Maastricht University, Maastricht, The Netherlands. 120Genentech, Inc., South San Francisco, CA, USA. 121Department of Psychology, Penn State University, University Park, PA, USA. 122Social Life and Engineering Sciences Imaging Center, University Park, PA, USA. 123Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA, USA. 124Research and Scientific Institute of Pediatrics and Child Health, CCH RAS, Ministry of Science and Higher Education, Moscow, Russian Federation. 125Department of Pediatrics, Michigan State University, East Lansing, MI, USA. 126Institute for Quantitative Health Science and Engineering, East Lansing, MI, USA. 127Department of Psychiatry, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA. 128CBRAIN, Department of Child and Adolescent Psychiatry, Psychosomatics, and Psychotherapy, Ludwig-Maximilians-Universität München, Munich, Germany. 129Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA. 130Chan Division of Occupational Science and Occupational Therapy, Los Angeles, CA, USA. 131Center for Clinical Spectroscopy, Brigham and Women’s Hospital, Boston, MA, USA. 132Harvard Medical School, Boston, MA, USA. 133National Center for PTSD at Boston VA Healthcare System, Boston, MA, USA. 134Department of Psychiatry, Boston University School of Medicine, Boston, MA, USA. 135Biomedical Genetics, Boston University School of Medicine, Boston, MA, USA. 136School of Psychology, University of Auckland, Auckland, New Zealand. 137Laboratory of Neuro Imaging, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA. 138Department of Psychiatry and Human Behavior, University of California, Irvine, Irvine, CA, USA. 139 Mind Research Network, Albuquerque, NM, USA. 140Psychiatry, San Diego, CA, USA. 141The Kavli Foundation, Los Angeles, CA, USA. 142Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK. 143Department of Psychiatry, Duke University School of Medicine, Durham, NC, USA. 144Mental Illness Research Education and Clinical Center, Durham VA Medical Center, Durham, NC, USA. 145Experimental Clinical & Health Psychology, Ghent University, Ghent, Belgium. 146Department of Personality, Psychological Assessment and Treatment, University of Deusto, Bilbao, Spain. 147 Radiology and Biomedical Imaging, San Francisco, CA, USA. 148Department of Pediatrics, Russian National Research Medical University MoH RF, Moscow, Russian Federation. 149Department of Psychology, Norwegian University of Science and Technology, Trondheim, Norway. 150Department of Physical Medicine and Rehabilitation, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway. 151Biological Sciences, Purdue University, West Lafayette, IN, USA. 152National Institute of Mental Health Intramural Research Program, Bethesda, MD, USA. 153Department of Genetics and Computational Biology, QIMR Berghofer Medical Research Institute, Brisbane, QLD, Australia. 154Department of Neurodegenerative Disease, UCL Queen Square Institute of Neurology, London, UK. 155Max Planck Institute of Psychiatry, Munich, Germany. 156Centre for Youth Mental Health, The University of Melbourne, Melbourne, VIC, Australia. 157Department of Psychiatry and Psychotherapy, Charite, Humboldt University, Berlin, Germany. 158Department of Radiology, Keck School of Medicine of USC, University of Southern California, Los Angeles, CA, USA. 159Department of Clinical and Experimental Epilepsy, University College London, London, UK. 160Chalfont Centre for Epilepsy, Chalfont St Peter, UK. 161Department of Medical Genetics, Oslo University Hospital, Oslo, Norway. 162Department of Psychiatry & Neuroscience Institute, SA MRC Unit on Risk & Resilience in Mental Disorders, Cape Town, South Africa. 163Department of Genetics & UNC Neuroscience Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA. 164Institute of Medical Science and Technology, Shahid Beheshti University, Tehran, I. R., Iran. 165Department of Neurology, TBI and Concussion Center, Salt Lake City, UT, USA. 166Missouri Institute of Mental Health, Berkeley, MO, USA. 167Psychology Department & Neuroscience Institute, Georgia State University, Atlanta, GA, USA. 168Clinical Translational Neuroscience Laboratory, Department of Psychiatry and Human Behavior, University of California Irvine, Irvine, CA, USA. 169 Center for the Neurobiology of Learning and Memory, University of California, Irvine, Irvine, CA, USA. 170Donders Centre for Cognitive Neuroimaging, Radboud University Medical Centre, Nijmegen, The Netherlands. 171Division of Mind and Brain Research, Department of Psychiatry and Psychotherapy CCM, Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Berlin, Germany. Thompson et al. Translational Psychiatry (2020)10:100 Page 28 of 28 172 Molecular and Cellular Therapeutics, Royal College of Surgeons in Ireland, Dublin, Ireland. 173Research and Early Development, Biogen Inc, Cambridge, MA, USA. 174 VA Salt Lake City Healthcare System, Salt Lake City, UT, USA. 175Department of Physical Medicine and Rehabilitation, Baylor College of Medicine, Houston, TX, USA. 176Keck School of Medicine, University of Southern California, Los Angeles, CA, USA. 177Skolkovo Institute of Science and Technology, Moscow, Russian Federation