Using Statistical Distribution to Identify the Influence Connections in Brain Networks.

Functional connectivity is a widely used method to explore the connections among different regions in the brain. However, it is still a challenge to extract structures from multiple similar brain networks that can characterize the different groups of patients in functional connectivity. In this paper, we commence from the binomial distribution and then extend to double gamma distributions to binarize brain network. This produces an optimal threshold in the process of network construction. We introduce a community-based nodal finding algorithm to handle the influence of the rich effect on finding important nodes. The experiments present an effective classification result on different brain network structures.