Edge communities in functional brain networks reveal heterogeneous, overlapping organization across the human lifespan

Abstract

Understanding changes in functional brain organization and their implications in development and aging is one of the central questions in neuroscience. In this study, we used an edge-centric approach to examine cross-sectional differences in functional brain network organization across the human lifespan using resting state functional MRI data from the Nathan Kline Institute - Rockland Sample dataset. By creating edge time series - a framewise multiplication of nodal time series - and clustering them based on their temporal similarities, we were able to identify clusters of edges instead of nodes. This method naturally allows multiple community affiliations per node (brain region), providing a nuanced perspective on network participation compared to conventional hard-partition approaches. To do so, we created age-neutral templates of edge communities - or "eFC lures" - that, when applied, yielded consistent edge communities across non-overlapping subsamples of data. The communities of edges revealed a trajectory of desegregation with aging, suggested to be linked to neural dedifferentiation of activity and cognitive decline in older adults. Additionally, age group-specific lures significantly enhanced the detection of edge community organization compared to the age-neutral version. Combined, these results offer new insights into the heterogeneous, event cluster-level shifts in brain functional organization as well as underscore the importance of age-targeted analytical frameworks throughout the human lifespan.