Intraindividual fluctuations in decision-making predict real-life drinking in alcohol use disorder

Abstract

Alcohol use disorder (AUD) is a major contributor to global disability and mortality. Predicting how people transition from occasional to excessive substance use remains a challenge. Previously, AUD has been linked to reduced cognitive control and increased risky decision-making in cross-sectional studies. However, these relationships can reflect changes that could either be a consequence or precedent of substance use. Thus, an untested hypothesis remains whether fluctuations in cognitive control and decision-making may temporally precede fluctuations in substance use. Here, we test this hypothesis based on a unique, preregistered, one-year longitudinal ecological momentary assessment (EMA) study. We employed EMA of real-life alcohol use in combination with a battery of smartphone-based gamified cognitive control and decision-making tests in n=288 participants with AUD. As hypothesized, we found that more risky decision-making (mixed gambles, information sampling) in a given month predicted increased alcohol consumption in the subsequent month. These results are also supported by a mechanistic computational model of information sampling in risky decision-making. Follow-up analyses further supported a specific temporal direction of the effect such that changes in decision-making preceded subsequent drinking but not vice versa. In contrast to measures of decision-making, cognitive control was not linked to subsequent alcohol consumption. However, we found, as hypothesized, that the detrimental intraindividual impact of risky decision-making on subsequent drinking was buffered in individuals with high working memory. In sum, we report first-time real-life longitudinal results from smartphone-based experiments that reveal intraindividual fluctuations in decision-making as a driving mechanism underlying subsequent fluctuations in alcohol consumption. Our smartphone-based experimental readouts can open new avenues for innovative mechanism-based and just-in-time interventions in AUD.