Recurrently connected and localized neuronal communities initiate coordinated spontaneous activity in neuronal networks
Fig 5
Analysis of the structural connectivity motifs in the functional communities of the network.
Clustering coefficient of structural motifs tested (up to six nodes and isomorphic subgraphs) against the relative abundance of structural motifs in fCOMs and the null-models: rndCOMs (A), rCOMs (B) and sCOMs (C) (see Materials and Methods section). Red dots mark the structural motifs occurring with a significant frequency difference (p-value<0.05, t-test) for all null-models. Regression lines visualize the trend (where a and p denote the value of the slope and the p-value respectively) among all tested motifs (gray) and the statistically different ones (red). (D) Illustration of overabundant (positive index) and under-represented (negative indexes) structural motifs found in fCOMs with respect to the null models depicted in A-C. As highlighted by the red lines, a high clustering coefficient characterizes the structural motifs that are significantly over-expressed in fCOMs; conversely, significantly under-represented structural motifs resulted in low clustering coefficients (0.56±0.30 vs. 0.19±0.27, respectively).