The top significant variables were also shown in a multi-way importance plot based on the RF importance measures Gini decrease, number of trees, and p-value (Figure 2). A variable is deemed significant if that variable is used for splitting more often than at random. As listed in Table 3, the variables that were found to be important for group classification are: 29 FC connections (across multiple 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).
