Table 2.

Graph theory metrics.

Term	Description	Computation
Clustering coefficient	A fraction of a node’s neighbors that are also neighbours of each other; a measure of clustered connectivity around individual nodes	$C=\frac{1}{n}{\sum }_{i\in n}{C}_i=\frac{1}{n}{\sum }_{i\in n}\frac{2t_i}{k_i(k_i-1)}$
	In the context of CT networks, it reflects uniformity of CT with respect to individual nodes	n = the total number of nodes
		Ci = the clustering coefficient of node i
		ki = the degree of node i
		Ci = 0 for ki < 2
Normalized clustering coefficient	Ratio of the mean clustering coefficient C and normalization factor Crnad computed as the mean clustering coefficient of 10 random networks (see below) with the same number of nodes and edges as the tested input network	$C_{\mathit{norm}}=\frac{C}{C_{\mathit{rand}}}$
Characteristic pathlength	A measure of network integration representing the number of edges typically required to connect pairs of nodes in the network	$L=\frac{1}{n}{\sum }_{i\in N}{L}_i=\frac{1}{n}{\sum }_{i\in N}\frac{{\sum }_{j\in N·j\ne i}{d}_{\mathit{ij}}^{-1}}{n-1}$
	In the context of CT networks, path length represents the number of required indirect correlations surpassing the sparsity threshold	Li = the average distance between node i and all other nodes
		dij = the distance from node i to node j
Normalized pathlength	The ratio of characteristic path length L and a normalization factor Lrnad based on 10 random networks, as described above	$L_{\mathit{norm}}=\frac{L}{L_{\mathit{rand}}}$
Small-world index	Describes a topology featuring numerous short-range connections with an admixture of few long-range connections; balances specialized and distributed processing while minimizing wiring costs. Small-world networks lie on a continuum between regular networks, in which each node has the same number of edges, and random networks, in which nodes are connected to other nodes with a random probability	$S=\frac{C_{\mathit{norm}}}{L_{\mathit{norm}}}$