Modeling decision dynamics disentangles working memory, cognitive control and reinforcement learning and reveals clinical differences.

Joint modeling of decision dynamics improves disentangling of working memory, reinforcement learning, and cognitive control, revealing clinical differences in schizophrenia.

• Why it matters: Understanding the distinct contributions of neurocognitive processes to decision making is crucial for diagnosing and treating mental health disorders, but current models often conflate these processes, limiting insights into underlying mechanisms.
• What they did: The study employed hierarchical Bayesian modeling of choices and response times during a reinforcement learning task with manipulated working memory demands, comparing models with and without decision dynamics.
• The result: Incorporating decision dynamics enhanced parameter recovery and out-of-sample prediction, uncovered proactive control strategies, and revealed specific deficits in schizophrenia, advancing computational phenotyping of psychopathologies.