Overview
- The study, published Oct. 19–20, 2026, followed 603 people with mostly mild-to-moderate alcohol use disorder for one year using gamified decision tasks each month and alcohol reports every two days.
- Within-person increases in specific risk-taking on mixed-gamble tasks and drops in information sampling reliably forecasted higher alcohol consumption in the following month.
- Measures of general cognitive control such as working memory and response inhibition did not predict drinking changes, indicating the signal came from narrow decision-making markers rather than broad executive deficits.
- Analyses showed the relationship was unidirectional because prior alcohol intake did not predict later deterioration in the tested decision-making measures.
- Authors and the journal say smartphone-based gamified monitoring could power digital phenotyping and just-in-time support, but they caution clinical use will need replication, randomized trials, and attention to privacy and implementation before being offered to patients.