Decision making occurs in dynamic contexts in which the individual and reward attributes change, requiring that predicted reward values of options be updated continually and applied to future choices to maximize reward. Value processing relies on corticostriatal dopamine, which is dysregulated in people with cocaine use disorder (pwCocUD), but whether these individuals differ in value-based decision making has only recently been considered. A probabilistic concurrent monetary choice task with reversals, reinforcement learning modeling, and functional magnetic resonance imaging were used to assess value-based decision making in nontreatment-seeking pwCocUD and matched controls (n = 17 8F and 9M per group; n = 34 total). Both groups were sensitive to reinforcement probabilities, but pwCocUD made significantly fewer choices for the high probability reward option. Reinforcement learning modeling revealed that pwCocUD had lower β parameter estimates, indicating reduced consistency in using relative option values to make choices. Significant correlations between the application of value (i.e., β parameter) and value-modulated activity during choice deliberation were identified in brain regions previously linked to reinforcement learning and exploiting versus exploring choice options in both groups. A significant group difference in the strength of this relationship was found in medial frontal control regions, including the dorsal anterior cingulate cortex, which has been associated with the explore-exploit trade-off, suggesting that the differential engagement of these areas contributed to group differences in choice behavior. This research contributes to our understanding of suboptimal decisions made by pwCocUD in uncertain, low-reward contexts and support targeting networks involved in value-based decision making for neuromodulation intervention development. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Lile et al. (Thu,) studied this question.