A cluster p-value (PFDR) representing significance after correction for multiple comparisons was then calculated for each candidate by applying Benjamini-Hochberg on raw p-values within the cluster-specific PA domain of the S-matrix. Candidate clusters with PFDR > 0.05 were excluded from further consideration. Cluster-specific comparisons of PAC by AUD diagnosis were conducted on the remaining candidates by calculating median PAC for each participant, then performing group comparisons of the cluster median PAC estimates with Mann-Whitney U test (two-tailed, α = 0.05). Cluster PAC strength for each participant was calculated by taking the median PAC from within the boundaries of the cluster-specific PA domain from that participant's average comodulogram (described in previous section). Only candidate clusters satisfying all validation and statistical testing conditions were reported as significant PAC clusters.
