Investigating cortico-subcortical interactions in functional connectivity with edge community structure

COEX 

The structural interconnectivity between cortical and subcortical regions is well established, and both can be organized within functional networks. However, the predominant emphasis has been on region (node)-centric analysis of connectivity. Edge time-series, a framework that temporally unravels functional connectivity, allows us to study functional dynamics among edges of a network. Applying this framework to 92 participants from the Human Connectome Project (HCP), edge time-series were clustered into communities, where each edge was assigned into a community, allowing nodes to have multiple/overlapping community affiliations. Focusing on the subnetwork comprised only of the interaction edges between cortex and subcortex, we find that cortical and subcortical regions are functionally coupled via multiple edge communities, with hippocampus and amygdala showing a distinct patten from striatum and thalamus. Additionally, cortical functional organization could be estimated solely from the cortico-subcortical interaction edges. Finally, we applied a motif analysis, where edge community triads between any two cortical nodes and a subcortical node can be classified based on number of unique communities in the triad. In the HCP sample, 2-community triads where a single community coupled a subcortical to two cortical nodes, were found to be overrepresented relative to a null model. These findings are aligned with our current understanding of the role the subcortex plays in integrating and distributing information from/to cortical regions and highlight the potential for use of edge community motifs to study network interactions.