This study, presented by Hani Zainal and Van Doren, uses machine learning applications to evaluate how sleep disturbances predict alcohol use disorder (AUD) risk in midlife adults; introducing the Sleep-Anxiety Dysregulation Model of AUD Risk. Through the administration of clinical interviews, self-reports, and a seven-day actigraphy protocol assessing demographics, psychiatric symptoms, anxiety severity, subjective sleep, and objective actigraphy sleep indices, 1,054 community-dwelling midlife adults were examined. In their procedures, Hani Zainal & Van Doren employed a five-folded nested cross-validated random forest to identify potentially non-linear and interactive predictors. The results demonstrated that baseline objective sleep and wake activity patterns and psychiatric comorbidities strongly predict the severity of an AUD even a decade later. Additionally, poor sleep quality, low daytime activity, and high levels of anxiety and depression symptoms tend to led to worsened AUDs. While self-reported sleep issues were not predictive, objective sleep measurement was a significant predictor in understanding long-term AUD risk. Sleep and anxiety are heavily emphasized factors in developing long-term risk for alcohol use disorders. The Sleep-Anxiety Dysregulation Model emphasizes that chronic arousal dysregulation, shown by physiological hyperreactivity and disrupted sleep, are strong indicators for developing an AUD.