Individuals with problematic alcohol use have implicit cognitive biases towards alcohol cues, which may influence drinking behavior. A previous study found that implicit approach/avoidance tendencies towards alcohol stimuli in such individuals was correlated with future level of alcohol use. The current study sought to explore the concurrent relationship between approach/avoidance, reward/punishment sensitivity and risk-taking.

Participants were 17 adults (M=51.7 years; SD=10.8) reporting heavy alcohol consumption (M=31.3 weekly drinks; SD=15.0; 92% with an Alcohol Use Disorder) and seeking treatment for problematic alcohol use. The majority of participants were women (55%) who were college-educated (84.2%), employed (73.7%), and racially White ( 78.9%). At treatment baseline, they completed the computerized Alcohol Approach/Avoidance (AAT), Reward/Punishment Learning (RPLT) and Balloon Analogue Risk (BART) tasks in randomized order. The AAT assesses alcohol approach/avoid tendencies by measuring the reaction time (RT) of participants to execute forward (approach) and backward (avoid) joystick movements in response to alcohol and neutral visual cues. The RPLT assesses reward/punishment sensitivity by measuring % optimal responding during categorization of abstract stimuli based on feedback that is probabilistically positive, negative or neutral. The BART assesses risk-taking by requiring participants to “pump” digital balloons via mouse-click to earn money (25¢/pump), with successive clicks increasing the risk of a pop and loss of accrued money. Primary indices on these tasks were subjected to partial correlation analyses (controlling for approach/avoidance RTs to neutral stimuli).

AAT alcohol approach RTs were inversely and exclusively correlated with RPLT % optimal responding on positive trials (r = -0.82; p<0.001) and AAT alcohol avoidance RTs were inversely and exclusively correlated with RPLT % optimal responding on negative trials (r = -0.98; p<0.001). The BART was uncorrelated (p>0.05) with both other measures.

Increased alcohol approach tendency was strongly associated with decreased reward learning, and increased alcohol avoid tendency was strongly associated with decreased punishment learning. Risk-taking was unrelated to these functions. Although sensitivity to reward is hypothesized to be associated with appetitive responses to alcohol, this relationship may reverse itself after alcohol use reaches the chronic/heavy stage. Future research should examine these associations in moderate adult drinkers as well as adolescents who are just initiating alcohol use.