The goal of the present study was to identify specific features from a set of multi-domain measures, including functional connectivity in the reward network, neuropsychological performance, and impulsivity, to classify individuals with AUD from healthy controls. The results showed that the random forest algorithm was highly successful in identifying the key features that contributed significantly to differentiating AUD from CTL individuals. Relative to controls, AUD individuals manifested (i) alterations in functional connectivity across reward network regions (including ventral tegmental area, nucleus accumbens, anterior insula, anterior cingulate cortex, and other cortical and subcortical structures), showing hypoconnectivity in nine connections and hyperconnectivity in three connections, (ii) increased impulsivity in motor and non-planning categories, and (iii) poorer neuropsychological performance, in terms of total number of correct trials in the forward sequence of the visual-spatial memory span test. In summary, relative to healthy controls, AUD individuals manifested aberrant functional connectivity in the reward network, increased impulsivity, and poor neuropsychological performance in visual–spatial memory.
