Fig. 2: Multivariable analysis of psychometric scales across groups.

This regression model depicts the association between alcohol craving, drinking levels, depression, and anxiety across race and AUD diagnosis. A Alcohol craving (PACS) and alcohol drinking (TLFB7) showed a significant positive correlation (R = 0.62, p = 0.004) in Caucasian controls. B In African-American control group, alcohol craving (PACS) and alcohol drinking (TLFB7) showed a significant positive correlation (R = 0.73, p = 0.001). Significant positive correlations were also found between anxiety and depression in this sample group (R = 0.78, p = 0.001). C Significant negative correlation between alcohol intake (TLFB) and depression (R = −0.79, p = 0.001) as well as anxiety (R = −0.69, p = 0.001) for the Caucasian AUD group, which is not present in African-American AUD group. D African-American AUD groups showed a significant positive correlation (R = 0.81, p = 0.001) between depression and anxiety. Red circle represents positive correlation, while blue circle represents negative correlation between the x-axis and y-axis. Circle size indicates p-values. Green boxes indicate the statistically significant correlation (p < 0.05; correlation coefficient of >0.60). The correlations were estimated by Row-wise method. E Principle component analysis (PCA) of psychometric scales can clearly discriminate healthy controls from AUD subjects across the x-axis, but not by race. The algorithm was not able to distinguish race accurately, as seen by the overlapping of race clusters and misclassified study subjects indicated by the dots.