This is an implementation of ADSD algorithm to identify connectomic subgraphs that are systematically associated with brain disorders from group-level brain connectome data.
"greedy_ad.m" is the function for greedy algorithm (Algorithm (1) in Web Appendix) to maximize the adaptive density function with a given lambda.
"greedy_ad_blik.m" is the complete ADSD algorithm with lambda selected by integrated likelihood (Algorithm (2) in Web Appendix).
"perm_label.m" is for the network-level inference based on non-parametric permutation tests.
We include a demo to reproduce the ADSD results on the brain network data from schizophrenia study. "edge_wise_res.mat" is the matrix of edge-wise inference results. "demo.m" illustrates the ADSD procedure on the weighted adjacency matrix. To see the demonstrative document with codes and output, please refer to "demo.html" observed from the publish option of Matlab.