ABSTRACT Depressive symptoms are commonly observed in stress‐related psychiatric disorders, such as major depressive disorder (MDD) and posttraumatic stress disorder (PTSD). To date, emerging evidence from behavior and psychology suggests the possibility of underlying neurobiological mechanisms in transdiagnostic depression. This study aims to identify predictive signatures of depression severity across MDD and PTSD using a whole‐brain connectivity machine learning analysis based on resting‐state functional magnetic resonance imaging (rs‐fMRI). Patients with MDD ( n = 84) and PTSD ( n = 65), all medication‐free at the time of enrollment, underwent rs‐fMRI scans along with a battery of clinical assessments. Using a multivariate machine learning approach, we applied sparse connectome predictive modeling to identify a functional connectivity signature that predicts individual depression severity, as assessed by Hamilton Depression Rating Scale‐17 items. The cross‐validated model explained 42% of the variance in depression severity across MDD and PTSD. The identified connectome signature predominantly involved regions in the fronto‐limbic circuit (e.g., middle frontal gyrus and temporal pole), subcortical areas (e.g., hippocampal, caudate, and brainstem), and the cerebellum. Our findings highlight diffuse whole‐brain dysfunction patterns associated with depressive symptom severity, emphasizing the importance of transdiagnostic research in understanding the neurobiological mechanisms underlying key clinical features across disorders.
Yang et al. (Wed,) studied this question.