Bayesian Modeling of Context-Specific Goal-Directed and Habitual Control in Sequential Decision Making

In a three-step sequential decision task, action sequence repetition improves habit-context response speed and optimal-choice performance, but it initially reverses after contingency changes and does not fully recover.

• Why it matters: It remains unclear how context-specific habits form and persist while goal-directed planning adapts when task contingencies shift.
• What they did: Participants performed interleaved planning and habit contexts while action sequence repetition, contingency change, and context switching were manipulated; a Bayesian contextual control model fit choice patterns using learned sequence priors, goal-directed evaluation, and context inference.
• The result: After contingencies changed, the previously habitual sequence became suboptimal and advantages reversed; later, habit-context optimal-choice performance recovered without significantly exceeding planning, while responses stayed slower, and context switches reduced optimal-choice performance in both directions.

AI summary of the abstract · may contain errors · not medical advice · read the paper.