Network Community Structure Detection for Directional Neural Networks Inferred From

Abstract—In many neuroscience applications, one is interested in identifying the functional brain modules from multichannel, multiple subject neuroimaging data. However, most of the existing network community structure detection algorithms are limited to single undirected networks and cannot reveal the common commu-nity structure for a collection of directed networks. In this paper, we propose a community detection algorithm for weighted asym-metric (directed) networks representing the effective connectivity in the brain. Moreover, the issue of finding a common commu-nity structure across subjects is addressed by maximizing the total modularity of the group. Finally, the proposed community detec-tion algorithm is applied to multichannel multisubject electroen-cephalogram data. Index Terms—Community detection, directed information, ef-fective connectivity, electroencephalogram(EEG), group analysis. I