FIG. 3.

The most central nodes and connections of the social brain network. After the four centrality measures (closeness centrality, betweenness centrality, vulnerability, and dynamic importance) were calculated and ranked for each node, a hub was defined as a node with at least three of four centrality measures ranked at the highest percentile; again, to avoid arbitrary selection of a single percentile threshold, “highest” was systematically varied from 3% (with only one highest-ranking node passing the threshold) to 100% (with all nodes passing the threshold). After identifying hubs in this manner across all percentile thresholds, the frequency of hub occurrence was counted and ranked. Nodes that passed the threshold for inclusion as a high-ranking hub more frequently are plotted with larger spheres and brighter yellow color. The most frequently identified putative hubs are bilateral DMPFC and AMG, suggesting the critical roles of these nodes in the structural social brain network. The most central connections in the social brain network were identified using edge betweenness centrality; their relative centrality is plotted proportionally to the width of each connection line in the figure.