Molecular Psychiatry (2021) 26:2101–2110 https://doi.org/10.1038/s41380-020-01002-z ARTICLE Structural brain imaging studies offer clues about the effects of the shared genetic etiology among neuropsychiatric disorders Nevena V. Radonjić 1 Jonathan L. Hess 2 Paula Rovira 3,4 Ole Andreassen5 Jan K. Buitelaar 6,7,8 ● ● ● ● ● Christopher R. K. Ching9 Barbara Franke 7,10,11 Martine Hoogman10 Neda Jahanshad 12 Carrie McDonald13 ● ● ● ● ● Lianne Schmaal 14,15 Sanjay M. Sisodiya16,17 Dan J. Stein 18 Odile A. van den Heuvel19 ● ● ● ● Theo G. M. van Erp 20,21 Daan van Rooij22 Dick J. Veltman19 Paul Thompson23 Stephen V. Faraone 2 ● ● ● ● Received: 29 May 2020 / Revised: 7 December 2020 / Accepted: 11 December 2020 / Published online: 17 January 2021 © The Author(s) 2021. This article is published with open access Abstract Genomewide association studies have found significant genetic correlations among many neuropsychiatric disorders. In contrast, we know much less about the degree to which structural brain alterations are similar among disorders and, if so, the degree to which such similarities have a genetic etiology. From the Enhancing Neuroimaging Genetics through Meta- 1234567890();,: 1234567890();,: Analysis (ENIGMA) consortium, we acquired standardized mean differences (SMDs) in regional brain volume and cortical thickness between cases and controls. We had data on 41 brain regions for: attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), bipolar disorder (BD), epilepsy, major depressive disorder (MDD), obsessive compulsive disorder (OCD), and schizophrenia (SCZ). These data had been derived from 24,360 patients and 37,425 controls. The SMDs were significantly correlated between SCZ and BD, OCD, MDD, and ASD. MDD was positively correlated with BD and OCD. BD was positively correlated with OCD and negatively correlated with ADHD. These pairwise correlations among disorders were correlated with the corresponding pairwise correlations among disorders derived from genomewide association studies (r = 0.494). Our results show substantial similarities in sMRI phenotypes among neuropsychiatric disorders and suggest that these similarities are accounted for, in part, by corresponding similarities in common genetic variant architectures. Introduction Subcortical brain volumes and cortical thickness/surface area dynamically change from early development through Neuropsychiatric disorders have substantial heritability, as adulthood and old age. A study of the Enhancing Neuroi- shown by many studies of twins and families [1]. Geno- maging Genetics through Meta-Analysis (ENIGMA) Plasti- mewide association studies (GWAS) have shown that city Working Group reported that changes in structural common genetic variants account for some of this herit- magnetic resonance imaging (sMRI) phenotypes have herit- ability, and that some of this heritability is shared across abilities ranging from 5% for pallidum to 42% for cerebellar neuropsychiatric disorders [2–5]. The genetic overlap across gray matter [7]. Heritability estimates of change rates were disorders may partly explain why these disorders tend to co- age-related and generally higher in adults than in children, occur with one another in both clinical and community probably due to an increasing influence of genetic factors with samples [6]. age [7]. However, it appears that later in adulthood heritability decreases most likely due to cumulative effect of environ- mental influences over the lifespan [8]. ENIGMA sMRI stu- dies of different psychiatric and neurological disorders further Supplementary information The online version of this article (https:// characterized MRI-derived phenotypes that can be used to doi.org/10.1038/s41380-020-01002-z) contains supplementary assess heritability (reviewed in [9]). material, which is available to authorized users. ENIGMA has also reported significant case vs. control * Stephen V. Faraone differences in sMRI phenotypes for: attention-deficit/ sfaraone@childpsychresearch.org hyperactivity disorder (ADHD) [10, 11], autism spectrum Extended author information available on the last page of the article disorder (ASD) [12], bipolar disorder (BD) [13, 14], 2102 N. V. Radonjić et al. common epilepsy syndromes [15], major depressive dis- Table 2 are the numbers of affected cases and unaffected order (MDD) [16, 17], obsessive compulsive disorder control participants included in each GWAS. Note, the full (OCD) [18, 19], and schizophrenia (SCZ) [20, 21]. Here we meta-analysis GWAS of MDD that included data from estimate the degree of similarity in sMRI phenotypes among 23andMe was not available for public release, thus we used these disorders and evaluate whether these similarities are the meta-analysis that combined results from the PGC influenced by corresponding similarities in common genetic cohorts and UK Biobank. variant architectures. Genetic and sMRI phenotype correlations among neuropsychiatric disorders Methods Linkage disequilibrium (LD)-score regression, a popular Collection of structural neuroimaging summary approach designed to analyze summary statistics from statistics GWAS, was used to quantify the amount of shared genetic heritability, or genetic correlation (rg), existing between Summary statistics from ENIGMA structural neuroimaging pairs of neuropsychiatric disorders, considering HapMap3 studies were collected from 12 multisite analyses published LD-scores [22]. For these analyses, the largest and latest by the ENIGMA Consortium for the following neu- GWAS available for each neuropsychiatric disorder was ropsychiatric disorders: ADHD [10, 11], ASD [12], BD selected and filtered to exclude markers with INFO < 0.90 [13, 14], epilepsy [15], MDD [16, 17], OCD [18, 19], and or within the MHC region (hg19:chr6:25–35 Mb) (Supple- SCZ [20, 21]. Prior to computing the summary statistics, the mentary Table 1). GWAS summary statistics were merged regional brain volumes had been segmented with a common with the HapMap3 reference panel (hg37 build), wherein ENIGMA protocol using FreeSurfer software. Each site variants with a MAF ≥ 5% in the HapMap3 dataset were performed these segmentations on their raw data. In addi- retained, prior to computing (co)heritability estimates. tion, quality control protocols provided by ENIGMA were To derive an estimate of the degree to which sMRI run at each site. Details are at: http://enigma.ini.usc.edu/ phenotypes were similar among disorders, we computed protocols/imaging-protocols. pairwise Spearman’s rank correlation between the Cohen’s The ADHD and ASD samples comprised both youth and d SMDs for each pair of disorders. We then used Pear- adults. The other samples comprised adults only. The eth- son’s correlation to estimate, whether the genetic corre- nicity of the patients was not available for all participants. lations for each disorder covaried with the sMRI The “epilepsy” cohort comprised temporal lobe epilepsy, phenotype correlations. We used a traditional permutation genetic generalized epilepsy, and extra temporal epilepsy. framework to generate a null distribution of sMRI phe- We analyzed 7 subcortical and 34 cortical regions (total of notype correlations by randomly shuffling Cohen’s d 41 brain regions; the mean of left and right structures) that values 10,000 times for each pair of disorders, then were included in the above specified ENIGMA studies. We recalculating sMRI correlations from the shuffled sets. extracted the covariate-adjusted Cohen’s d standardized From the null distributions, we derived an empirical per- mean differences (SMDs) denoting the case versus unaf- mutation p value for each sMRI phenotype correlation. fected comparison subject differences in subcortical volume However, a reliable p value could not be calculated due to and cortical thickness/surface area measures. The covariates nonindependence between pairwise caused by sample used in these studies adjusted SMDs for several covariates overlap between imaging studies. Adjustments for sample as indicated in Supplementary Table 1. overlap would be possible with individual-level data, but the present study only had access to summary statistics. In Collection of GWAS results among neuropsychiatric a leave-one-out analysis, we iteratively excluded one pair disorders of disorder correlations from the set and recalculated Spearman’s correlation coefficients to determine whether Publicly available summary statistics from GWAS were correlations were driven by any pair of disorders. Bino- downloaded from the Psychiatric Genomics Consortium mial sign tests were used to determine whether the number (PCG) website (https://www.med.unc.edu/pgc/results-and- of disorders showing the same direction of effect in the downloads/) with the exception of GWAS results for MDD sMRI phenotypes was greater than expected by chance coming from an online resource hosted by the University of (null probability of 50%). Per brain region, we performed Edinburgh (https://doi.org/10.7488/ds/2458) and of GWAS Cochran’s Q test implemented in the R package metafor results for epilepsy coming from the online Epilepsy Genetic (v.2.1–0) to determine whether variability among Cohen’s Association Database (epiGAD) (http://www.epigad.org/ d values was greater than expected by chance. All statis- gwas_ilae2018_16loci.html). Presented in Supplementary tical analyses were performed with R version 3.5.2 Structural brain imaging studies offer clues about the effects of the shared genetic etiology among. . . 2103 Table 1 Sample demographics for the twelve studies by the ENIGMA Consortium into structural brain alterations in neuropsychiatric disorders. Disorder MRI measure Cases (n) Controls (n) Total n Sites Weighted mean age (cases) Weighted mean age (controls) References ADHD Cortical thickness 2246 1934 4180 36 19.2 18.1 [2, 28] Surface area 2246 1934 4180 36 19.2 18.1 Subcortical volume 1713 1529 3242 23 18.6 ASD Cortical thickness 1571 1651 3222 49 15.4 [18] Surface area Subcortical volume BD Cortical thickness 1837 2582 4419 28 38.4a 35.6a [19, 30] a Surface area 1820 2582 4402 28 38.4 35.6a Subcortical volume 1710 2594 4304 20 40.1a 36.5a Epilepsy Cortical thickness 2149 1727 3876 24 34.4 33.3 [7] Surface area Subcortical volume MDD Cortical thickness 1911 7663 9574 20 44.8a 54.6a [6, 22] a Surface area 1902 7658 9560 20 44.8 54.6a Subcortical volume 1728 7199 8927 15 43.3a 56a OCD Cortical thickness 1498 1435 2933 27 32.1 30.5 [26, 41] Surface area 1497 1433 2930 27 32.1 30.5 Subcortical volume 1495 1472 2967 25 32.0 30.6 a SCZ Cortical thickness 4474 5098 9572 39 32.3 34.5a [27, 34] a Surface area 4434 5073 9507 39 32.3 34.5a Subcortical volume 2028 2540 4568 15 34.0a 31.0a a Weighted mean not provided in paper; computed from descriptive statistics. (R Core Team, 2018). We adjusted for repeated correla- affected. As indicated by the blueness of the cells, the most tion tests using the Bonferroni procedure. Correlations prominent reductions were seen for SCZ (mean Cohen’s d showing a Bonferroni-adjusted p < 0.05 were considered across all regions = −0.22, SE = 0.014), epilepsy (mean significant (threshold p = 0.00227). Cohen’s d = −0.12, SE = 0.017) and BD (mean Cohen’s d = −0.097, SE = 0.011). The smallest changes were observed for MDD (mean Cohen’s d = −0.018, SE = Results 0.006). All regions except for the caudate and putamen exhibited significant differences in the magnitude of Sample demographics for the twelve studies by the Cohen’s d across disorders (Cochran’s Q p values = ENIGMA Consortium on structural brain abnormalities in 0.012–2.8 × 10−32). Eighteen sMRI phenotypes exhibited neuropsychiatric disorders are presented in Table 1. homogeneity with respect to sign of Cohen’s d across each of the neuropsychiatric disorders evaluated (binomial sign Case–control differences in subcortical volume and test p values < 0.05): cortical thicknesses for caudal middle cortical surface area and thickness within frontal gyrus, entorhinal cortex, fusiform gyrus, inferior neuropsychiatric disorders temporal gyrus, insula, lateral orbitofrontal cortex, lingual gyrus, middle temporal gyrus, paracentral lobule, para- Figure 1 presents an anatomical graph of the standardized hippocampal gyrus, pars opercularis of inferior temporal effect sizes (Cohen’s d) measuring alterations in subcortical gyrus, precentral gyrus, precuneus, rostral anterior cingulate volume, cortical surface area and cortical thickness for 41 cortex, and supramarginal gyrus; subcortical volume for the brain regions within seven neuropsychiatric disorders— hippocampus; and surface area for middle temporal gyrus, ADHD, ASD, OCD, epilepsy, MDD, BD, and SCZ. These pars triangularis of inferior temporal gyrus, and pericalcar- have been reported on prior publications. The variation in ine cortex. For sMRI phenotypes for 39 regions of interest color from blue to red illustrates the phenomenon of SBRV, varying degrees of heterogeneity were noted in terms of with some regions showing significant reductions (blue) in discrepancy of signs of Cohen’s d. For example, individuals volume/thickness/surface areas and others not being with ASD showed a slightly thicker cortex in the rostral 2104 N. V. Radonjić et al. Fig. 1 A brain imaging plot showing standardized mean differ- brain regions in cases versus unaffected comparisons. Note: ADHD ences (Cohen’s d) measuring case–control differences in sub- attention-deficit/hyperactivity disorder, ASD autism spectrum dis- cortical volumes and cortical thickness for seven neuropsychiatric order, BD bipolar disorder, MDD major depressive disorder, OCD disorders. Results were obtained from ENIGMA working group obsessive compulsive disorder, SCZ schizophrenia. publications. Negative values for Cohen’s d indicate smaller sizes of middle frontal gyrus, individuals with ADHD showed no Correlation of shared genetic heritability with brain difference, and all other disorders showed a thinner cortex structural correlation in this region compared to controls. Figure 3 shows the pairwise correlations of sMRI pheno- sMRI phenotype correlations among types and genetic overlap across each pair of neu- neuropsychiatric disorders ropsychiatric disorders. The LD-score cross-disorder genetic correlations are positively correlated with the sMRI For each pair of disorders, we computed the Pearson corre- phenotype cross-disorder correlations (Spearman’s ρ = lation between their sMRI phenotypes listed in Fig. 1. These 0.44, p = 0.049). Leave-one-out sensitivity analyses con- are listed in Table 2 (and visualized in Fig. 2), sorted by the firmed that the direction of the correlation was positive and magnitude of the correlation. The p values reported in Table 2 remained moderate in magnitude despite removal of indi- are potentially downwardly biased due to inability to properly vidual pairs of disorders from the correlation test (range of adjust for spatial coherence of nearby brain regions. Tradi- Spearman’s ρ = 0.35–0.58), except for removing SCZ/BD tional permutation p values are provided as a column in (Spearman’s ρ = 0.35). SCZ and BD showed the highest Table 2, which attempts to correct for potential biases due to degree of concordance with respect to genetic and sMRI spatial coherence. However, we were restricted from using a phenotype correlations. spatial permutation framework to generate a null distribution of correlations, because we are jointly analyzing two cortical maps (cortical thickness and surface area) that are fully Discussion overlapped. The highest positive correlation was between SCZ and BD (r = 0.81, df = 73, p < 1.3 × 10−18, Bonferroni Our analysis of summary statistics from the ENIGMA p = 2.38 × 10−17). There were a few additional nominally ADHD, ASD, BD, MDD, OCD, SCZ, and epilepsy significant negative correlations, which did not survive mul- Working Groups and the predominantly PGC tiple testing correction: MDD and epilepsy (r = −0.37, p = case–control GWAS identified two novel findings. First, 0.02), MDD and ADHD (r = −0.33, p = 0.004), SCZ and we found substantial correlations for some disorders in the ADHD (r = −0.32, p = 0.005), ADHD and epilepsy (r = pattern of sMRI case–control differences across sub- −0.36, p = 0.02), and a positive correlation between MDD cortical and cortical regions in line with recently pub- and ASD (r = 0.26, p = 0.02). lished study of [23]. Second, these cross-disorder Structural brain imaging studies offer clues about the effects of the shared genetic etiology among. . . 2105 Table 2 Cross-disorder Disorder 1 Disorder 2 sMRI correlation df se p value Boferroni Permutation structural MRI phenotype Pearson’s r adjusted p value p value correlations (ordered from smallest to largest p value) based BD SCZ 0.81 73 0.068 1.13E−18 2.38E−17 <1E10−4 on Cohen’s d values obtained from the ENIGMA Project. BD MDD 0.69 73 0.085 1.21E−11 2.55E−10 <1E10−4 OCD SCZ 0.65 72 0.09 5.53E−10 1.16E−08 <1E10−4 MDD SCZ 0.58 73 0.095 5.55E−08 1.17E−06 <1E10−4 ADHD BD −0.53 73 0.099 1.18E−06 2.48E−05 <1E10−4 BD OCD 0.5 72 0.102 4.74E−06 9.95E−05 <1E10−4 MDD OCD 0.46 72 0.104 3.28E−05 6.89E−04 <1E10−4 ASD BD 0.38 73 0.108 8.98E−04 0.02 <1E10−4 ASD SCZ 0.36 73 0.109 1.35E−03 0.03 0.0176 ADHD MDD −0.33 73 0.111 4.27E−03 0.09 0.019 ADHD SCZ −0.32 73 0.111 4.63E−03 0.1 0.0014 Epilepsy MDD −0.37 39 0.149 0.02 0.38 0.0056 ADHD Epilepsy −0.36 39 0.149 0.02 0.41 0.004 ASD MDD 0.26 73 0.113 0.02 0.46 0.024 Epilepsy OCD −0.19 39 0.157 0.23 1 0.22 BD Epilepsy 0.17 39 0.158 0.3 1 0.3 ADHD OCD −0.1 72 0.117 0.39 1 0.39 ADHD ASD −0.06 73 0.117 0.6 1 0.6 Epilepsy SCZ −0.03 39 0.16 0.86 1 0.85 ASD Epilepsy 0.02 39 0.16 0.91 1 0.91 ASD OCD 0 72 0.118 0.97 1 0.97 Fig. 3 Scatter plot showing the correlation of correlations. Genetic correlations (rg) computed by LD-score regression are on the hor- Fig. 2 A heatmap of the cross-disorder pairwise sMRI correlations izontal axis (with standard error bars), with correlations of Cohen’s d between seven neuropsychiatric disorders examined in this study. values displayed on the vertical axis. Each dot is color-coded Colors in the plot correspond to the magnitude of the Pearson’s r according to the pairwise disorder correlations that were computed. coefficients, which are provided in each tile. Note: ADHD attention- The dotted line represents the best-fit regression line. The Spearman’s deficit/hyperactivity disorder, ASD autism spectrum disorder, BD rho (ρ) and p value are provided at the top-left corner of the panel. bipolar disorder, MDD major depressive disorder, OCD obsessive Note: ADHD attention-deficit/hyperactivity disorder, ASD autism compulsive disorder, SCZ schizophrenia. spectrum disorder, BD bipolar disorder, MDD major depressive dis- order, OCD obsessive compulsive disorder, SCZ schizophrenia. correlations in SBRV could partly be explained by the The cross-disorder correlations in SBRV are intriguing genetic correlations reported for these disorders from because, like cross-disorder genetic correlations, they sug- GWAS [3]. gest that these disorders, to varying degrees, share aspects 2106 N. V. Radonjić et al. of their etiology and pathophysiology. Any interpretation of high sMRI correlations. These are SCZ/BD, SZ/MDD, BD/ the cross-disorder sMRI correlations must keep in mind MDD, OCD/BD, and OCD/MDD. The inclusion of MDD that, for all disorders, the case–control differences in sMRI in this group is notable given that it is part of the bipolar measures are small (Fig. 1). The largest Cohen’s d values diagnosis and often occurs comorbid with other disorders. are only −0.5 for SCZ [20, 21], −0.4 for epilepsy [15], MDD also has a high genetic correlation with ADHD but a −0.3 for BD [13, 14], −0.2 for ADHD [10, 11] and ASDs negative sMRI correlation, which makes that pair an outlier [12], and −0.1 for MDD [16, 17] and OCD [18, 19]. These in Fig. 3. small case–control differences are consistent with results In the lower left region of Fig. 3, we see disorders with from GWAS and environmental risk studies, which speaks low genetic and low sMRI correlations. These involve to the fact that the effects of common risk factors are, with correlations of epilepsy, and correlations of ADHD with all some rare exceptions, individually small. Although it is disorders except ASDs and MDD, although the latter is conceivable that these small risks could accumulate to somewhat of an outlier. ASDs tend to have both modest create a more dramatic pathophysiology in the brain, the genetic correlations and modest sMRI correlations with ENIGMA data show that this is not the case for sMRI most other disorders and, hence, populates the middle range measures. Consistent with this finding, interindividual dif- of the figure. Like the sMRI correlations among disorders, ferences in neuroimaging account for only a small amount all genetic correlations with epilepsy are low, which is of the variance in symptom expression or behavioral mea- consistent with the low genetic correlation between neuro- sures of symptomatic or behavioral variance [24]. logical and psychiatric disorders as reported by [2]. The most prominent case–control differences in cortical The finding that SBRV correlations are correlated with thickness/surface area and subcortical volumes were genetic correlations suggests that future studies of SBRV observed for SCZ [20, 21] and BD [13, 14]. These disorders should consider genetic sources of etiology. Yet, because also had the highest sMRI phenotype correlations and both only about 24% of the variance in the SBRV correlations can also showed strong sMRI phenotype correlations with be accounted for by the genetic correlations, environmental MDD [16, 17] and OCD [18, 19]. As Fig. 2 shows, these sources of etiology and disease-specific genetic contributions disorders clustered together in the three-dimensional con- must also be considered. These include shared confounders, figuration required to capture cross-disorder sMRI pheno- such as chronicity and medication exposure, along with type similarity. The high sMRI correlation between SCZ shared etiologic events such as birth complications or expo- and BD is consistent with prior reports of sMRI similarities sure to toxins in utero. Our prior studies of SBRV in ADHD between the two disorders [25]. Moreover, a large body of implicated the regulation of genes in apoptosis, autophagy literature reports substantial etiologic overlap between the and neurodevelopment pathways in ADHD [42, 43]. Neu- two disorders [26–30]. Because of such data, the SCZ and rodevelopmental pathways had also been implicated in the BD have been described as sharing a continuum of etiology cross-disorder analysis of the PCG [3], which suggests that leading to psychotic [31], neurophysiological [31] and cross-disorder similarities in these pathways may account for neurocognitive [32] symptoms. The ENPACT study [33] cross-disorder similarities in SBRV. showed shared fronto–temporo–occipital gray matter Although we used data derived from very large samples volume deficits in the right hemisphere of two disorders. A (ENIGMA, iPSYCH, and the PGC), several limitations systematic review of associations between functional MRI moderate the strength of our conclusions. We inherit all the activity and polygenic risk for SCZ and BD [26] reported limitations of the constituent studies, but are further limited that genetic load for these disorders affects task-related because we analyzed summary statistics, not the original recruitment of predominantly frontal lobe brain regions. data, which would require the sharing of individual subject Many studies have reported that OCD can be a comorbid level data, an ongoing effort among the ENIGMA disorder diagnosis with SCZ or that patients with SCZ can have OCD working groups. Thus, we cannot determine whether the symptoms [34–41]. Presented findings of a significant over- possible use of controls shared among studies affected our lap in sMRI phenotypes along with the known SCZ/OCD results. It is also possible that some research participants genetic correlations suggests that more work should examine were included in the genetic and sMRI data sets for the shared pathophysiologic features between these disorders and same disorder. The p value obtained by our Spearman’s should assess the degree to which confounds, such as med- correlation test of cross-disorder sMRI and genetic corre- ication status or chronicity, might explain these results. lations may be inaccurate due to spatial autocorrelation The sMRI phenotype correlations mirror, to some extent, among sMRI Cohen’s d estimates, which can downwardly the cross-disorder correlations from GWAS. Figure 3 shows bias standard errors and lead to deflated p values. Con- a modest, yet distinct, linear correlation between the sMRI sidering we are not able to completely address with auto- phenotype and genetic correlations. In the upper right-hand correlation among brain regions using summary statistics section of the plot, we see disorders having high genetic and alone, the p value from our primary analysis (presented in Structural brain imaging studies offer clues about the effects of the shared genetic etiology among. . . 2107 Fig. 3) should be interpreted with caution. Another problem Acknowledgements SVF is supported by the European Union’s is that we could not address effects of medications or Seventh Framework Program for research, technological development and demonstration under grant agreement no. 602805, the European chronicity on brain structure. Furthermore, for some of the Union’s Horizon 2020 research and innovation program under grant disorders, we could use youth and adult data, whereas for agreements nos. 667302 and 728018 and NIMH grants others only adult effect data were used. Because findings 5R01MH101519 and U01 MH109536-01. Research Council of Nor- can differ substantially depending on the age range of the way (#223273). BF is supported by a personal Vici grant from the Netherlands Organization for Scientific Research (NWO, grant number samples included (e.g., [10, 11, 18, 19], this might have 91813669) and by a grant from the European Union’s Horizon 2020 influenced our findings. For these reasons, analyses of program for the CoCa project (grant agreement no 667302). ENIGMA participant level data will be needed to address these issues work is supported by NIH grants U54 EB020403 (PI: PT), R01 to draw stronger and more detailed conclusions. We also MH116147 (PI: PT) and R01MH117601 (MPIs: NJ and LS). MH is supported by a personal Veni grant from the Netherlands Organization did not have any longitudinal data available, which limits for Scientific Research (NWO, grant number 91619115). CMD is the ability to test hypotheses about shared and unique supported by NIH grants R01 NS065838 and R21 NS107739. PR is a developmental trajectories among disorders. recipient of a pre-doctoral fellowship from the Agència de Gestió Despite these limitations, we have documented cross- d’Ajuts Universitaris i de Recerca (AGAUR), Generalitat de Catalu- nya, Spain (2016FI_B 00899). LS is supported by a NHMRC Career disorder correlations in SBRV as assessed by sMRI. These Development Fellowship (1140764). SMS is supported by Epilepsy cross-disorder SBRV correlations are positively associated Society, UK, and the work was partly undertaken at UCLH/UCL, with the disorders’ corresponding cross-disorder genetic which received a proportion of funding from the UK Department of correlations. This finding is a novel contribution worthy of Health’s NIHR Biomedical Research Centers funding scheme.). TGMVE is supported by NIH grants U54 EB020403 (PI: PT), R01 further study that contributes to novel body of literature MH116147 (PI: PT), R01MH117601 (MPIs: NJ and LS), and focused on cross-level correspondence of genetic and neu- R01MH121246 (MPIs: Turner, TGMVE, and Calhoun). roimaging presentations of different psychiatric disorders [44–49]. Our work supports conclusions from previous Compliance with ethical standards GWAS studies suggesting a partially shared etiology and pathophysiology among many disorders [2, 50]. Disorders Conflict of interest OA has received speaker’s honorarium from like SCZ and BD or ADHD and ASD, which are distinct in Lundbeck and is a consultant to HealthLytix. In the past year, SVF received income, potential income, travel expenses continuing edu- the diagnostic nomenclature, show significant overlap in cation support, and/or research support from Tris, Otsuka, Arbor, etiology and pathophysiology. Further studies are needed to Ironshore, Shire, Akili Interactive Labs, Enzymotec, Sunovion, discern why brain regions are selectively affected by the Supernus and Genomind. With his institution, he has US patent risk factors that cause sMRI abnormalities [42, 43] and why US20130217707 A1 for the use of sodium–hydrogen exchange inhi- bitors in the treatment of ADHD. He also receives royalties from these effects are correlated across disorders. Such studies books published by Guilford Press: Straight Talk about Your Child’s may give insights into new treatment targets. Mental Health, Oxford University Press: Schizophrenia: The Facts and Elsevier: ADHD: Non-Pharmacologic Interventions. He is Program Data availability Director of www.adhdinadults.com. BF received educational speaking fees from Medice. All other authors declare no conflict of interest. URLs for GWAS SCZ from ckqny.scz2snpres.gz Publisher’s note Springer Nature remains neutral with regard to (https://www.med.unc.edu/pgc/results-and-downloads), ASD jurisdictional claims in published maps and institutional affiliations. from iPSYCH-PGC_ASD_Nov2017.gz (https://www.med. unc.edu/pgc/results-and-downloads), OCD from PGC_OC- Open Access This article is licensed under a Creative Commons D_Aug2017-20171122T182645Z-001.zip > ocd_aug2017.gz Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as (https://www.med.unc.edu/pgc/results-and-downloads), long as you give appropriate credit to the original author(s) and the ADHD from adhd_ul2017.gz (https://www.med.unc.edu/pgc/ source, provide a link to the Creative Commons license, and indicate if results-and-downloads), BD from daner_PGC_BIP32b_md- changes were made. The images or other third party material in this s7a_0416a.gz (https://www.med.unc.edu/pgc/results-and-dow article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not nloads), Epilepsy from all_epilepsy_METAL.gz (http://www. included in the article’s Creative Commons license and your intended epigad.org/gwas_ilae2018_16loci.html), and MDD from use is not permitted by statutory regulation or exceeds the permitted PGC_UKB_depression_genome-wide.txt (https://doi.org/10. use, you will need to obtain permission directly from the copyright 7488/ds/2458). holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/. Code availability References Custom R scripts used to generate results in this study can 1. Polderman TJ, Benyamin B, de Leeuw CA, Sullivan PF, van be made available upon request. Bochoven A, Visscher PM, et al. Meta-analysis of the heritability 2108 N. V. Radonjić et al. of human traits based on fifty years of twin studies. Nat Genet. 16. Schmaal L, Hibar DP, Sämann PG, Hall GB, Baune BT, Jahan- 2015;47:702–9. https://doi.org/10.1038/ng.3285. shad N, et al. Cortical abnormalities in adults and adolescents with 2. Anttila V, Bulik-Sullivan B, Finucane HK, Walters RK, Bras J, major depression based on brain scans from 20 cohorts worldwide Duncan L, et al. Analysis of shared heritability in common dis- in the ENIGMA Major Depressive Disorder Working Group. Mol orders of the brain. Science. 2018;360:eaap8757. https://doi.org/ Psychiatry. 2016;22:900. https://doi.org/10.1038/mp.2016.60. 10.1126/science.aap8757. 17. Schmaal L, Veltman DJ, van Erp TGM, Sämann PG, Frodl T, 3. Lee PH, Anttila V, Won H, Feng Y-CA, Rosenthal J, Zhu Z, et al. Jahanshad N, et al. Subcortical brain alterations in major depres- Genome wide meta-analysis identifies genomic relationships, sive disorder: findings from the ENIGMA Major Depressive novel loci, and pleiotropic mechanisms across eight psychiatric Disorder working group. Mol Psychiatry. 2015;21:806. https:// disorders. bioRxiv. 2019. https://doi.org/10.1101/528117. doi.org/10.1038/mp.2015.69. 4. Lee SH, Ripke S, Neale BM, Faraone SV, Purcell SM, Perlis RH, 18. Boedhoe PSW, Schmaal L, Abe Y, Alonso P, Ameis SH, et al. Genetic relationship between five psychiatric disorders Anticevic A, et al. Cortical abnormalities associated with pediatric estimated from genome-wide SNPs. Nat Genet. 2013;45:984–94. and adult obsessive-compulsive disorder: findings from the https://doi.org/10.1038/ng.2711. ENIGMA obsessive-compulsive disorder working group. Am J 5. Smoller JW, Andreassen OA, Edenberg HJ, Faraone SV, Glatt SJ, Psychiatry. 2018;175:453–62. https://doi.org/10.1176/appi.ajp. Kendler KS. Correction to: psychiatric genetics and the structure 2017.17050485. of psychopathology. Mol Psychiatry. 2018a. https://doi.org/10. 19. Boedhoe PSW, Schmaal L, Abe Y, Ameis SH, Arnold PD, 1038/s41380-018-0026-4. Batistuzzo MC, et al. Distinct subcortical volume alterations in 6. Buckholtz JoshuaW, Meyer-Lindenberg A. Psychopathology and pediatric and adult OCD: a worldwide meta- and mega-analysis. the human connectome: toward a transdiagnostic model of risk for Am J Psychiatry. 2017;174:60–9. https://doi.org/10.1176/appi.ajp. mental illness. Neuron. 2012;74:990–1004. https://doi.org/10. 2016.16020201. 1016/j.neuron.2012.06.002. 20. van Erp TGM, Hibar DP, Rasmussen JM, Glahn DC, Pearlson 7. Brouwer RM, Panizzon MS, Glahn DC, Hibar DP, Hua X, GD, Andreassen OA, et al. Subcortical brain volume abnormal- Jahanshad N, et al. Genetic influences on individual differences in ities in 2028 individuals with schizophrenia and 2540 healthy longitudinal changes in global and subcortical brain volumes: controls via the ENIGMA consortium. Mol Psychiatry. results of the ENIGMA plasticity working group. Hum Brain 2015;21:547. https://doi.org/10.1038/mp.2015.63. Mapp. 2017;38:4444–58. https://doi.org/10.1002/hbm.23672. 21. van Erp TGM, Walton E, Hibar DP, Schmaal L, Jiang W, Glahn 8. Ge T, Chen C-Y, Neale BM, Sabuncu MR, Smoller JW. Phe- DC, et al. Cortical brain abnormalities in 4474 individuals with nome-wide heritability analysis of the UK Biobank. PLOS schizophrenia and 5098 control subjects via the enhancing neuro Genetics 2017;13:e1006711. imaging genetics through meta analysis (ENIGMA) consortium. 9. Thompson PM, Andreassen OA, Arias-Vasquez A, Bearden CE, Biol Psychiatry. 2018;84:644–54. https://doi.org/10.1016/j. Boedhoe PS, Brouwer RM, et al. ENIGMA and the individual: biopsych.2018.04.023. Predicting factors that affect the brain in 35 countries world- 22. Bulik-Sullivan BK, Loh PR, Finucane HK, Ripke S, Yang J, wide. NeuroImage. 2017;145:389–408. https://doi.org/10.1016/ Schizophrenia Working Group of the Psychiatric Genomics, C, j.neuroimage.2015.11.057. et al. LD score regression distinguishes confounding from poly- 10. Hoogman M, Bralten J, Hibar DP, Mennes M, Zwiers MP, genicity in genome-wide association studies. Nat Genet. Schweren LSJ, et al. Subcortical brain volume differences in 2015;47:291–5. https://doi.org/10.1038/ng.3211. participants with attention deficit hyperactivity disorder in chil- 23. Opel N, Goltermann J, Hermesdorf M, Berger K, Baune BT, dren and adults: a cross-sectional mega-analysis. Lancet Psy- Dannlowski U. Cross-Disorder Analysis of Brain Structural chiatry. 2017;4:310–9. https://doi.org/10.1016/S2215-0366(17) Abnormalities in Six Major Psychiatric Disorders: A Secondary 30049-4. Analysis of Mega- and Meta-analytical Findings From the 11. Hoogman M, Muetzel R, Guimaraes JP, Shumskaya E, Mennes ENIGMA Consortium. Biol Psychiatry. 2020. M, Zwiers MP, et al. Brain imaging of the cortex in ADHD: a 24. Paulus MP, Thompson WK. The Challenges and Opportunities of coordinated analysis of large-scale clinical and population-based Small Effects: The New Normal in Academic Psychiatry. JAMA samples. Am J Psychiatry. 2019. https://doi.org/10.1176/appi.ajp. Psychiatry. 2019;76:353–4. https://doi.org/10.1001/jamapsychia 2019.18091033. try.2018.4540. 12. van Rooij D, Anagnostou E, Arango C, Auzias G, Behrmann M, 25. Rimol LM, Hartberg CB, Nesvag R, Fennema-Notestine C, Busatto GF, et al. Cortical and subcortical brain morphometry Hagler DJ Jr, Pung CJ, et al. Cortical thickness and subcortical differences between patients with autism spectrum disorder and volumes in schizophrenia and bipolar disorder. Biol Psychiatry. healthy individuals across the lifespan: results from the ENIGMA 2010;68:41–50. https://doi.org/10.1016/j.biopsych.2010.03.036. ASD Working Group. Am J Psychiatry. 2018;175:359–69. https:// 26. Dezhina Z, Ranlund S, Kyriakopoulos M, Williams SCR, Dima doi.org/10.1176/appi.ajp.2017.17010100. D. A systematic review of associations between functional MRI 13. Hibar DP, Westlye LT, Doan NT, Jahanshad N, Cheung JW, activity and polygenic risk for schizophrenia and bipolar disorder. Ching CRK, et al. Cortical abnormalities in bipolar disorder: an Brain Imaging Behav. 2019;13:862–77. https://doi.org/10.1007/ MRI analysis of 6503 individuals from the ENIGMA Bipolar s11682-018-9879-z. Disorder Working Group. Mol Psychiatry. 2017;23:932. https:// 27. Haukvik UK, Tamnes CK, Soderman E, Agartz I. Neuroimaging doi.org/10.1038/mp.2017.73. hippocampal subfields in schizophrenia and bipolar disorder: a 14. Hibar DP, Westlye LT, van Erp TG, Rasmussen J, Leonardo CD, systematic review and meta-analysis. J Psychiatr Res. 2018;104: Faskowitz J, et al. Subcortical volumetric abnormalities in bipolar 217–26. https://doi.org/10.1016/j.jpsychires.2018.08.012. disorder. Mol Psychiatry. 2016;21:1710–6. https://doi.org/10. 28. Berrettini W. Bipolar disorder and schizophrenia: convergent 1038/mp.2015.227. molecular data. Neuromolecular Med. 2004;5:109. 15. Whelan CD, Altmann A, Botia JA, Jahanshad N, Hibar DP, Absil 29. Lizano P, Bannai D, Lutz O, Kim LA, Miller J, Keshavan M. A J, et al. Structural brain abnormalities in the common epilepsies Meta-analysis of retinal cytoarchitectural abnormalities in schi- assessed in a worldwide ENIGMA study. Brain. zophrenia and bipolar disorder. Schizophr Bull. 2019. https://doi. 2018;141:391–408. https://doi.org/10.1093/brain/awx341. org/10.1093/schbul/sbz029. Structural brain imaging studies offer clues about the effects of the shared genetic etiology among. . . 2109 30. Bora E, Akgul O, Ceylan D, Ozerdem A. Neurological soft signs in and obsessive-compulsive disorder with poor insight: a neu- bipolar disorder in comparison to healthy controls and schizo- ropsychological comparison. Psychiatry Res. 2009;165:38–46. phrenia: a meta-analysis. Eur Neuropsychopharmacol. 2018;28: https://doi.org/10.1016/j.psychres.2007.07.031. 1185–93. https://doi.org/10.1016/j.euroneuro.2018.08.006. 41. Cunill R, Castells X, Simeon D. Relationships between obsessive- 31. Thaker G. Psychosis endophenotypes in schizophrenia and bipolar compulsive symptomatology and severity of psychosis in schi- disorder. Schizophr Bull. 2008;34:720–1. https://doi.org/10.1093/ zophrenia: a systematic review and meta-analysis. J Clin Psy- schbul/sbn055. chiatry. 2009;70:70–82. https://doi.org/10.4088/jcp.07r03618. 32. Kim D, Kim J, Koo T, Yun H, Won S. Shared and Distinct 42. Hess JL, Akutagava-Martins GC, Patak JD, Glatt SJ, Faraone SV. Neurocognitive Endophenotypes of Schizophrenia and Psychotic Why is there selective subcortical vulnerability in ADHD? Clues Bipolar Disorder. Clin Psychopharmacol Neurosci. 2015;13: from postmortem brain gene expression data. Mol Psychiatry. 94–102. 2017. https://doi.org/10.1038/mp.2017.242. 33. Maggioni E, Crespo-Facorro B, Nenadic I, Benedetti F, Gaser C, 43. Hess JL, Radonjić NV, Patak J, Glatt SJ, Faraone SV (2019). Sauer H, et al. Common and distinct structural features of schi- Spatial organization of cells and variable expression of autophagy, zophrenia and bipolar disorder: The European Network on Psy- apoptosis, and neurodevelopmental genes might underlie selective chosis, Affective disorders and Cognitive Trajectory (ENPACT) brain region vulnerability in Attention-Deficit/Hyperactivity Dis- study. PLoS ONE. 2017;12:e0188000. https://doi.org/10.1371/ order. bioRxiv. 2019. https://doi.org/10.1101/652792. journal.pone.0188000. 44. Gandal MJ, Haney JR, Parikshak NN, Leppa V, Ramaswami G, 34. Grover S, Sahoo S, Surendran I. Obsessive–compulsive symptoms Hartl C, et al. Shared molecular neuropathology across major in schizophrenia: a review. Acta Neuropsychiatr. 2019;31:63–73. psychiatric disorders parallels polygenic overlap. Science. https://doi.org/10.1017/neu.2018.27. 2018;359:693. 35. Hwang MLO. Schizophrenia with obsessive-compulsive features: 45. Anderson KM, Collins MA, Kong R, Fang K, Li J, He T, et al. assessment and treatment. Psychiatr Ann. 1994;24:468–72. Convergent molecular, cellular, and neural signatures of major https://doi.org/10.3928/0048-5713-19940901-08. depressive disorder. bioRxiv 2020: https://doi.org/10.1101/2020. 36. Özdemir Ö, Tükel R, Türksoy N, Üçok A. Clinical characteristics 02.10.942227. in obsessive-compulsive disorder with schizophrenia. Compr 46. Romme IA, de Reus MA, Ophoff RA, Kahn RS, van den Heuvel Psychiatry. 2003;44:311–6. https://doi.org/10.1016/S0010-440X MP. Connectome Disconnectivity and Cortical Gene Expression (03)00093-2. in Patients With Schizophrenia. Biol Psychiatry. 2017;81: 37. Poyurovsky M, Zohar J, Glick I, Koran LM, Weizman R, Tandon 495–502. R, et al. Obsessive-compulsive symptoms in schizophrenia: 47. Seidlitz J, Nadig A, Liu S, Bethlehem RAI, Vértes PE, Morgan implications for future psychiatric classifications. Compr Psy- SE, et al. Transcriptomic and cellular decoding of regional brain chiatry. 2012;53:480–3. https://doi.org/10.1016/j.comppsych. vulnerability to neurogenetic disorders. Nat Commun. 2011.08.009. 2020;11:3358. 38. Schirmbeck F, Zink M. Comorbid obsessive-compulsive symp- 48. Goodkind M, Eickhoff SB, Oathes DJ, Jiang Y, Chang A, Jones- toms in schizophrenia: contributions of pharmacological and Hagata LB, et al. Identification of a Common Neurobiological genetic factors. Front Pharmacol. 2013;4:99. https://doi.org/10. Substrate for Mental Illness. JAMA Psychiatry. 2015;72:305–15. 3389/fphar.2013.00099. 49. Disorder WCftA-DH, Disorder AS, Disorder B, Disorder MD, 39. Swets M, Dekker J, van Emmerik-van Oortmerssen K, Smid GE, Disorder O-C, Groups aSEW. Virtual Histology of Cortical Smit F, de Haan L, et al. The obsessive compulsive spectrum in Thickness and Shared Neurobiology in 6 Psychiatric Disorders. schizophrenia, a meta-analysis and meta-regression exploring JAMA Psychiatry. 2020. prevalence rates. Schizophr Res. 2014;152:458–68. https://doi. 50. Smoller JW, Andreassen OA, Edenberg HJ, Faraone SV, Glatt SJ, org/10.1016/j.schres.2013.10.033. Kendler KS. Psychiatric genetics and the structure of psycho- 40. Tumkaya S, Karadag F, Oguzhanoglu NK, Tekkanat C, Varma G, pathology. Mol Psychiatry. 2018b. https://doi.org/10.1038/ Ozdel O, et al. Schizophrenia with obsessive-compulsive disorder s41380-017-0010-4. Affiliations Nevena V. Radonjić 1 Jonathan L. Hess 2 Paula Rovira 3,4 Ole Andreassen5 Jan K. Buitelaar 6,7,8 ● ● ● ● ● Christopher R. K. Ching9 Barbara Franke 7,10,11 Martine Hoogman10 Neda Jahanshad 12 Carrie McDonald13 ● ● ● ● ● Lianne Schmaal 14,15 Sanjay M. Sisodiya16,17 Dan J. Stein 18 Odile A. van den Heuvel19 ● ● ● ● Theo G. M. van Erp 20,21 Daan van Rooij22 Dick J. Veltman19 Paul Thompson23 Stephen V. Faraone 2 ● ● ● ● 1 5 Department of Psychiatry, SUNY Upstate Medical University, NORMENT—Institute of Clinical Medicine, Division of Mental Syracuse, NY, USA Health and Addiction, Oslo University Hospital, University of 2 Oslo, Oslo, Norway Departments of Psychiatry and of Neuroscience and Physiology, 6 SUNY Upstate Medical University, Syracuse, NY, USA Radboudumc, Radboud University Medical Center, 3 Nijmegen, The Netherlands Psychiatric Genetics Unit, Group of Psychiatry, Mental Health and 7 Addiction, Vall d’Hebron Research Institute (VHIR), Universitat Donders Institute for Brain, Cognition and Behaviour, Radboud Autònoma de Barcelona, Barcelona, Spain University Medical Center, Nijmegen, The Netherlands 4 8 Department of Psychiatry, Hospital Universitari Vall d’Hebron, Department of Cognitive Neuroscience, Radboud University Barcelona, Spain Medical Center, Nijmegen, The Netherlands 2110 N. V. Radonjić et al. 9 16 Imaging Genetics Center, USC Mark and Mary Stevens UCL Queen Square Institute of Neurology, Department of Clinical Neuroimaging and Informatics Institute, Keck School of Medicine and Experimental Epilepsy, University College London, of the University of Southern California, Marina Del Rey, CA, London, UK USA 17 Chalfont Centre for Epilepsy, Epilepsy Society, Bucks, UK 10 Department of Human Genetics, Radboud University Medical 18 Center, Nijmegen, The Netherlands SA MRC Unit on Risk & Resilience in Mental Disorders, Department of Psychiatry & Neuroscience Institute, University of 11 Department of Psychiatry, Radboud University Medical Center, Cape Town, Cape Town, South Africa Nijmegen, The Netherlands 19 Department of Psychiatry and Department of Anatomy & 12 Imaging Genetics Center, Department of Neurology and Neurosciences, Amsterdam UMC/VUmc, Amsterdam, The Biomedical Engineering, USC Mark and Mary Stevens Netherlands Neuroimaging and Informatics Institute, Keck School of 20 Medicine of USC, University of Southern California, Marina Del Clinical Translational Neuroscience Laboratory, Department of Rey, CA, USA Psychiatry and Human Behavior, University of California Irvine, Irvine, CA, USA 13 Department of Psychiatry, Center for Multimodal Imaging and 21 Genetics (CMIG), University of California, San Diego, CA, USA Center for the Neurobiology of Learning and Memory, University of California Irvine, Irvine, CA, USA 14 Centre for Youth Mental Health, The University of Melbourne, 22 Parkville, VIC, Australia Donders Centre for Cognitive Neuroimaging, Radboud University Medical Center, Nijmegen, The Netherlands 15 Orygen, The National Centre of Excellence for Youth Mental 23 Health, Parkville, VIC, Australia Neuro Imaging Institute for Neuroimaging and Informatics, Keck School of Medicine of the University of Southern California, Marina Del Rey, CA, USA