780. Classification of Major Depressive Disorder Subtypes via Surface-based Brain Imaging and Clinical Features

Abstract

Abstract Background Major Depressive Disorder (MDD) is characterized by substantial heterogeneity in both symptomatic presentation and underlying neurobiology, posing significant challenges for accurate diagnosis and effective intervention. While prior research has attempted to delineate MDD subtypes using only neuroimaging features, the clinical interpretability of these subtypes often remains limited. Aims & Objectives Aiming to identify clinically meaningful MDD subtypes through an integrative analysis of clinical symptoms and multimodal neuroimaging data, this study leveraged samples from the REST-meta-MDD consortium. The final dataset included 474 MDD patients with Hamilton Depression Rating Scale (HAMD) scores and 418 healthy controls (HCs) across six sites. Method Structural and functional MRI data underwent surface-based preprocessing. Partial correlation analyses were performed between imaging metrics and HAMD scores to identify brain regions exhibiting clinically meaningful associations. Subsequently, Regularized Canonical Correlation Analysis (rCCA) was applied to these selected regions and clinical scores for feature dimensionality reduction. K-means clustering was then implemented on the derived components to delineate MDD subtypes. Finally, the identified subtypes were compared with HC to characterize their distinct profiles in brain structure, function, and clinical symptomatology. Results Two distinct MDD subtypes were identified (Subtype 1: n=299; Subtype 2: n=175). Relative to HC, Subtype 1 exhibited reduced cortical thickness and volume in multiple brain regions, including the frontal, temporal, limbic, and parietal lobes, whereas no significant functional alterations were detected. Conversely, Subtype 2 showed decreased amplitude of low-frequency fluctuation (ALFF) and regional homogeneity (ReHo) values across several resting-state networks—such as the default mode, salience/ventral attention, dorsal attention, visual, and control networks—without notable structural deficits. Clinically, Subtype 1 demonstrated higher scores than Subtype 2 in items related to depressive mood and psychic anxiety. Discussion & Conclusions This study delineated two MDD subtypes with distinct neuropathological profiles: one characterized by structural abnormalities across multiple brain regions and the other by widespread functional disruptions. Notably, the subtype defined by structural alterations presented greater severity in depressive symptoms and psychic anxiety. These findings elucidate the clinical and biological heterogeneity of MDD, offering a theoretical foundation for refined diagnostic classification and the development of targeted therapeutic strategies.