Correct retrieval was not only indexed by the density of network connections but also by the precise organization of its edges to particular nodes. Specifically, we investigated regional importance using a combination of two metrics: node degree and betweenness centrality. Node degree captures the local connectivity of a node, while betweenness centrality describes more global aspects of the network by tracking the tendency for the connections from other nodes to travel through a given node. Previous studies have found these two graph theory metrics to be reliable predictors of functional hubs181920. We found that several areas within the medial temporal lobes (MTL), frontal and parietal lobes, and the visual cortex, exhibited significantly greater levels of betweenness centrality than other nodes in the network for correct compared to incorrect retrieval trials. Figure 2D shows graphs for brain hubs displaying high betweenness centrality, with chi-squared tests indicating significant differences between the two conditions. Tests were Bonferroni corrected at p < 0.05 to account for multiple comparisons between regions (all pcorrected < 0.05; Calcarine sulcus (Calc): χ2 (1) = 93.86; anterior hippocampus (aHPC):
