Network Parameter Analysis

Graph theoretical analysis was performed on the interregional connectivity matrix by using GRETNA¹, a graph theoretical network analysis toolbox for imaging connectomics. The weighted network properties were calculated, with a sparsity range of 0.05–0.4 with a step size of 0.01. Sparsity was defined as the total number of edges divided by the maximum possible number of edges. Because there is no gold standard to select a single threshold, we calculated the parameters with different thresholds. Finally, the networks were constructed at the sparsity of 0.14, which ensured all nodes included in the networks to present the nodal characteristics of the networks and ensured the most characteristic small-world topology. GRETNA was used to calculate the structural network topological properties, including the network efficiency properties (local efficiency and global efficiency), local Cp, global clustering coefficient [M(Cp)], local shortest path length (Lp), global shortest path length [M(Lp)], and the node degree and betweenness centrality for the node properties. For each subject, 1,000 times randomization was applied, and each time a corresponding random network was generated. Then, the random distribution of Cp and Lp was used to transform real Cp and Lp into a Z score by their position in the random distribution as previous studies (Wang et al., 2015). The brain networks were visualized with BrainNet Viewer² (Xia et al., 2013).