Edge-centric network reveals altered functional integration and dispersion in major depressive disorder

Abstract

Brain networks are composed of nodes representing neural elements, such as brain regions, and edges indicating functional or anatomical connections between these nodes. By shifting our focus from traditional node-centric perspectives to examining second-order similarity patterns between pairs of network edges, we captured and illuminated the co-fluctuation profiles between brain regions, revealing overlapping communities and the intensity of interactions within brain networks. Specifically, we mapped edge-centric networks and then computed edge-community normalized entropy and edge functional connectivity (eFC) to assess perturbations in normal brain network organization associated with major depressive disorder (MDD). Sample data were sourced from a cohort of 400 MDD patients and 441 healthy controls. Edge-community entropy was measured by clustering edge time series derived from resting-state functional magnetic resonance imaging data, while eFC was quantified using the Pearson correlation coefficient between edge time series. Our results showed that MDD patients exhibited increased entropy in the subcortical and frontoparietal networks and decreased eFC within the visual and sensory-motor networks compared to controls. These differences were less evident in first-episode drug-naive patients. However, in recurrent patients, the same abnormalities were observed and the entropy of subcortical network was positively correlated with depression severity, while the eFC of visual network was negatively correlated with depression and anxiety scores. This study provides new insights into the abnormal changes in MDD from a spatiotemporal flexibility and diversity perspective based on high-order edge-centric networks and offering potential novel biomarkers for MDD.