Background: Regional gray matter volume (GMV) differences between individuals with mental disorders and comparison subjects may be confounded by co-occurring disorders. To disentangle the disorder-specific GMV correlates, we conducted a large-scale multi-disorder meta-analysis using a novel approach that explicitly models co-occurring disorders.
Methods: We systematically reviewed voxel-based morphometry studies indexed in PubMed and Scopus up to January 2023 comparing adults with major mental disorders (anorexia nervosa, schizophrenia-spectrum, anxiety, bipolar, major depressive, obsessive-compulsive, and post-traumatic stress disorders, plus attention-deficit/hyperactivity, autism spectrum, and borderline personality disorders) to comparison subjects. Two authors independently extracted data and assessed quality using the Newcastle-Ottawa Scale. We derived GMV correlates for each disorder using: a) a multi-disorder meta-analysis accounting for all co-occurring mental disorders simultaneously; b) separate standard meta-analyses for each disorder ignoring co-occurring disorders. We assessed the alterations' extent, intensity (effect size), and specificity (inter-disorder correlations and transdiagnostic alterations) for both approaches.
Results: We included 433 studies (499 datasets) involving 19,718 patients and 16,441 comparison subjects (51% females, aged 20-67 years). We provide GMV correlate maps for each disorder using both approaches. The novel approach, which accounted for co-occurring disorders, produced GMV correlates that were more focal and disorder-specific (less correlated across disorders and fewer transdiagnostic abnormalities).
Conclusions: This work offers the most comprehensive atlas of GMV correlates across major mental disorders. Modeling co-occurring disorders yielded more specific correlates, supporting this approach's validity. The atlas NIfTI maps are available online.
Keywords: anxiety disorders; comorbidity; gray matter volume; major depressive disorder; meta-analysis; psychiatric conditions.
Copyright © 2024. Published by Elsevier Inc.