Abstract
BACKGROUND: The preference for immediate rewards and high sensation seeking are both potent risk factors for alcohol use disorder (AUD), but how they interact during intoxication is poorly understood. To model decision making linked to AUD risk, we tested heavy drinkers for impulsive choice (delay discounting with alcohol:money or money:money) and behavioral sensation seeking using a novel odor choice task. Laboratory tasks measured actual behavior with real contingencies. Our goals were to determine, in heavy drinkers, (i) alcohol's effects on delay discounting, and (ii) how AUD risk factors relate to delay discounting, and (iii) how delay discounting with alcohol choices compares with strictly monetary choices. METHODS: Thirty-five heavy drinkers (≥2 binges per month; age = 22.8 ± 2.2; 20 male; 5.8 ± 2.3 drinks/drinking day) performed cross-commodity discounting (CCD) of immediate alcohol vs. delayed money, a monetary delay discounting (DD), and behavioral sensation-seeking tasks. CCD and DD were performed while sober and during controlled alcohol infusion targeting 0.08 g/dl. The behavioral sensation-seeking task presented binary choices of odorants varying in intensity and novelty, and the risk of exposure to a malodorant. RESULTS: CCD and DD behaviors were highly correlated across conditions, mean r = 0.64. Alcohol increased delayed reward preference in DD, p = 0.001, but did not alter mean CCD, p > 0.16. However, alcohol-induced changes in CCD correlated with behavioral sensation seeking, such that higher sensation seekers' immediate alcohol preference increased when intoxicated, p = 0.042; self-reported sensation seeking was uncorrelated, ps > 0.08. Behavioral sensation seeking also correlated with "want" alcohol following a priming dose targeting 0.035 g/dl, p = 0.021. CCD and DD did not correlate with self-reported drinking problems or other personality risk traits. CONCLUSIONS: Alcohol increased impulsive alcohol choice in high sensation seekers, suggesting an interaction that may underlie impaired control of drinking, at least in a subset of heavy drinkers-consistent with models highlighting high novelty/sensation-seeking AUD subtypes. Discounting behavior overall appears to be a generalized process, and relatively stable across methods, repeated testing, and intoxication. These findings further support the utility of behavioral tasks in uncovering key behavioral phenotypes in AUD.