Shift to the core: Abnormal core-periphery global topography in unipolar and bipolar depression

This study explores the global signal topography of core and periphery brain networks in Major Depressive Disorder (MDD), Bipolar disorder (BD-Dep) and healthy controls (HC) using resting-state fMRI. In a sample of 140 depressed MDD and BD patients, and 70 HC, we observed a significant shift toward increased activity in the transmodal-core regions (e.g., default mode network, frontoparietal network) at the expense of unimodal-periphery regions (e.g., visual, sensory-motor cortices) in both depressed MDD and BD patients compared to HC. Whole brain machine learning analyses further demonstrated that altered global signal dynamics can effectively distinguish MDD and BD from HC (ACC = 79% and 77% respectively). Notably, we identified a significant negative correlation between global signal correlation in unimodal-periphery networks and depressive symptom severity. Additionally, in a smaller sample of BD during mania (N = 22) a distinct topographic pattern was observed, with increased global representation in the unimodal-periphery compared to depressive states, suggesting mood state-dependent shifts in network organization. To assess multivariate discriminability across diagnostic groups, a Partial Least Squares (PLS) analysis revealed that higher Core and related network activity (DMN, FPN) predicted diagnostic assignment to MDD and BD-Dep, whereas higher Periphery and related network (e.g., visual and sensory-motor networks) predicted assignment to BD-Man and HC. The Core–Periphery (C–P) ratio emerged as the strongest predictor (VIP = 1.65). These results underscore the critical role of global signal topography in mood disorders, particularly the imbalance between core and peripheral brain networks, as a potential neurobiological marker for depressive states.