The cross-sectional cohort was used to (1) examine correlates of mild, moderate, mild-to-moderate, and severe AUD; (2) identify AUD criteria indicative of heightened risk using IRT analysis; and (3) evaluate whether individuals with mild-to-moderate AUD who endorsed high-risk criteria differed from those who did not, and from those with severe AUD, across relevant factors, including alcohol-related, comorbid psychiatric, and electroencephalography (EEG)-derived traits and polygenic indices. Given extant literature demonstrating substantive differences between severe AUD and mild or moderate AUD,16,17,18 and statistical considerations for examining effects of criteria heterogeneity separately within mild and moderate AUD (ie, reduced power, large number of statistical tests17), we elected to combine mild and moderate AUD into a single referent group for analyses in both cohorts.
