The goal of the study was to identify features that may accurately classify AUD individuals from unaffected controls using multi-domain measures such as EEG-based FC of the DMN, neuropsychological performance, and impulsivity in an RF classifier algorithm. Results showed that the RF model achieved 80% classification accuracy and correctly identified 24 of 30 subjects in each group. The top features that contributed to group classification (ALC vs. CTL) were: 29 FC connections (across key DMN nodes representing all frequency bands), three neuropsychological variables (total number of correctly performed trials in forward and backward sequences, and average time of correct trials in forward sequence) and all four impulsivity scores (motor, non-planning, attentional, and total). The AUD group showed a predominant pattern of hyperconnectivity among all significant connections (25 of 29 connections, p < 0.05) (Figure 5, panels a–e) as well as among 12 highly significant connections (11 of 12 connections, p ≤ 0.001) (Figure 6), where parahippocampal connections with other DMN regions were predominant (13 connections overall; six highly significant connections). On the other hand, four prefrontal connections of PFC and
