The graph theoretic analysis of these networks was straightforward, and simply involved visualizing the graphs in a pseudo-anatomic layout and examining their structure, shown in Figure 4. In children ages 7–9 years, the graph was a single component, the dorsal anterior cingulate/medial superior frontal cortex (dACC/msFC) node was embedded within the FP nodes, and the bilateral anterior prefrontal cortex (aPFC) nodes bridged between FP and CO groupings (Figure 4A). In adolescents ages 10–15 years, the graph broke into two components, but the dACC/msFC node remained in the FP component (Figure 4B). In adults ages 21–31 years, the graph existed as 2 components corresponding to the FP and CO task control networks, as expected (Figure 4C). Thus, the network reorganized over development, such that anterior cingulate and aPFC nodes dissociated from other frontal nodes to more strongly associate with insular and thalamic nodes. The authors also measured the Euclidean distance between all pairs of ROIs, and noted that short-distance edges tended to be strong in children and weaken over development, whereas long-distance edges tended to be weak in children and strengthen
