Two subtypes of major depressive disorder are identified from individualized gray matter morphological abnormalities in a large multi-site dataset.

Abstract

Background: Neuroimaging studies provide compelling evidence that major depressive disorder (MDD) is associated with widespread gray matter morphological abnormalities. However, significant interindividual variability complicates the interpretation of group-level findings, highlighting the need for investigating potential MDD subtypes.

Methods: In this study, we aimed to identify subtypes of MDD based on individualized deviations from normative gray matter volumes (GMVs), as estimated using a normative model derived from healthy controls (HCs). We leveraged a large, multi-site dataset of high-resolution structural MRI scans, comprising 1,276 MDD patients and 1,104 matched HCs. To explore the transcriptional and molecular mechanisms underlying the observed structural abnormalities, we examined the relationships between GMV deviations, transcriptomic similarities (as measured by the correlated gene expression [CGE] connectome), and the distribution of neurotransmitter receptors/transporters.

Results: Our results revealed two reproducible MDD subtypes, each exhibiting distinct patterns of GMV abnormalities across study sites. Subtype 1 displayed increased GMVs in cerebral regions and decreased GMVs in cerebellar regions, whereas subtype 2 showed the opposite pattern, with decreased GMVs in cerebral regions and increased GMVs in cerebellar areas. The identified GMV abnormalities were differentially associated with neurotransmitter receptor/transporter distributions. Furthermore, these abnormalities were linked to transcriptionally connected gene networks, suggesting genetic underpinnings for both subtypes. Notably, the two subtypes exhibited distinct CGE-informed disease epicenters.

Conclusions: This study identifies two robust MDD subtypes, providing new insights into the neurobiological and genetic bases of MDD and offering a potential advancement in the nosology of the disorder.