Disentangling individual heterogeneity reveals robust network and molecular signatures of major depressive disorder with suicidal ideation

The neurobiological mechanisms of major depressive disorder with suicidal ideation (MDDSI) remain unclear, partly due to individual heterogeneity among patients with MDDSI. We developed a multi-level framework to extract individual-shared (IShN) and individual-specific brain networks (ISpN) using personalized principal component analysis (perPCA), construct structure-function coupling (SFC) network via graph embedding, and map network alterations to transcriptomic and neurotransmitter distributions. Structural, functional, and SFC networks were examined in 528 participants and replicated in 123 participants of an independent cohort. After removing individual heterogeneity, patients with MDDSI showed convergent disruptions within the default-mode network and action mode network across structural, functional, and SFC networks. These alterations corresponded to 5-HT2a and to the expression of genes involved in neurotransmitter transport, synaptic signalling, and neurodevelopmental pathways. By disentangling subject-specific components, the ISpN captured symptom-relevant variations that were obscured in the original brain networks, enabling more accurate diagnostic classification. Our findings identify reproducible, cross-modal network abnormalities and their molecular correlates underlying MDDSI, demonstrating the importance of disentangling individual heterogeneity for advancing the neurobiological understanding of MDDSI.