1113 Prospective Associations Between School Night Sleep, Pre-Weekend Reward Function, and Weekend Alcohol use in High School Students

Abstract Introduction Sleep and circadian factors longitudinally predict alcohol use during adolescence and adulthood, but the underlying mechanisms remain unclear. Furthermore, extant studies have largely focused on timeframes of months-to-years; studying more proximal sleep/circadian-alcohol associations may better inform prevention/intervention efforts. To address these gaps, we assessed objective sleep/circadian characteristics in high school drinkers, examining the proximal associations with one putative mechanism (reward function) and alcohol use across the school week and weekend. Methods One hundred and fourteen high school students reporting prior alcohol use completed an 8-day protocol during the school year that included both naturalistic and laboratory methods. Sleep was assessed throughout via actigraphy; circadian phase was assessed during an overnight laboratory visit on Thursday via the salivary dim light melatonin onset (DLMO). Self-report (positive affect; alcohol craving) and behavioral [Balloon Analogue Risk Task (BART)] measures of reward function were administered Friday in the lab. Self-reported alcohol use was assessed from Friday through Sunday. We hypothesized that later circadian phase (DLMO), greater circadian misalignment (shorter DLMO-midsleep intervals), and shorter total sleep time (TST) on weekdays would predict altered pre-weekend (Friday) reward function, and that the sleep/circadian variables and altered pre-weekend reward function would predict greater weekend alcohol use. Results Only 24% of participants endorsed alcohol use on the study weekend. Weekday DLMO and DLMO-midsleep were unrelated to pre-weekend reward function or weekend alcohol use. Moreover, weekday TST did not predict positive affect, alcohol craving, or weekend alcohol use. Shorter weekday TST did predict greater reward-related risk taking based on the BART (greater explosions; B=-0.23; p=0.017). Furthermore, greater explosions (BART) predicted a higher likelihood of weekend alcohol use (OR=1.19; p=0.004). Conclusion Findings did not support hypothesized associations between circadian markers and proximal reward function or alcohol use, although the latter association may have been limited by relatively low alcohol use on the study weekend. Findings were consistent with insufficient schoolnight sleep amplifying reward-related risk taking, which in turn was associated with a higher likelihood of weekend alcohol use. This suggests a path by which insufficient sleep impacts alcohol use in adolescents; experimental studies will be required to determine causality. Support (if any) R01AA025626 (Hasler)