NEUROIMAGING OF SUBCORTICAL BRAIN VOLUME ALTERATIONS IN PEADIATRIC AND ADULT OBSESSIVE-COMPULSIVE DISORDER: Preliminary findings from the ENIGMA Obsessive-Compulsive Disorder working group

Objective Structural MRI studies investigating the neural correlates of OCD have been numerous. Nevertheless, results of these studies have not always been consistent. Variability in study results may partially be explained by small sample sizes resulting in limited statistical power, clinical heterogeneity between patient samples, and methodological differences between studies. In this context, we initiated the ENIGMA-OCD working group. Our aim is to identify robust imaging markers of OCD using coordinated standardized image processing and statistical analysis protocols. Here, we perform the largest study to date of subcortical brain volumes in OCD patients and healthy controls using an individual participant data (IPD) based meta-analysis approach. Methods Structural T1-weightred MRI scans from 3722 subjects (361 paediatric OCD patients, 1574 adult OCD patients, 314 paediatric controls and 1473 adult controls) from 25 research sites worldwide were analysed using FreeSurfer 5.3. Segmentations of subcortical regions, lateral ventricles and total intracranial volumes (ICV) were visually inspected for accuracy and compared between patients and controls using regression models controlling for age, sex, and ICV. Each site followed standardized protocols designed to facilitate harmonized image analysis across multiple sites. Separate stratified analyses assessing effects of age of onset, disease duration, symptom dimensions, and symptom severity were performed. Results were combined in random-effects meta-analysis models. Meta-regression analyses were used to test whether mean age of each sample and field strength of MR images explained a significant proportion of the variance in effect seizes across sites in the meta-analysis. Results were considered significant if they exceeded a Bonferroni corrected P-value threshold (p=0.05/9 regions = 5.6x10 -3 ). Results Adult patients, compared to adult controls, had significantly smaller hippocampal volumes (Cohen’s d= -0.15, p= 7.4x10 -4 ) and bigger pallidum volumes (d= 0.16, p=1.8x10 -3 ). This effect was notably stronger in medicated patients versus controls (d= -0.29, p=2.8x10 -6 and d= 0.24, p=2.5x10 -4 , respectively). The hippocampus effect seemed to be driven by patients with comorbid depression d= 0.29, p=4.8x10 -5 ). Also, symptom severity was associated with a smaller hippocampus (R= -0.08, p=4.9x10 -3 ). Sample characteristics such as mean age and field strength did not moderate brain volume alterations. None of the subcortical structures, neither ICV nor lateral ventricles were significantly different between paediatric OCD cases and controls after Bonferroni correction. However, unmedicated paediatric patients show bigger thalamus volumes (d= 0.40, p=9.0x10 -4 ) relative to control children. Additionally, longer disease duration and ad trend-significant level, younger age of onset was associated with bigger thalamus volume in paediatric OCD (d= 0.18, p=5.1x10 -3 and d= -0.17, p=5.7x10 -3 , respectively). Conclusion Results of this subcortical meta-analysis indicate a key role of the pallidum in the pathophysiology of OCD. Different ENIGMA disease working groups, such as MDD, schizophrenia, and bipolar disorder also observed hippocampal abnormalities, which suggests that our findings regarding hippocampal abnormalities are aspecific for OCD. In addition, we showed that unmedicated paediatric OCD patients have significantly enlarged thalamus volumes. Moreover, our findings suggest a different pattern of subcortical abnormalities in paediatric OCD patients and adult OCD patients in line with the developmental nature of the disease.