Adaptive Decision-Making in Mouse Models of Neurodevelopmental Disorders

We have posted a new preprint on bioRxiv investigating adaptive decision-making in mouse models of neurodevelopmental disorders using the Operant House. [Distinct Patterns of Adaptive Decision-Making in Mouse Models of Neurodevelopmental Disorders | bioRxiv]

Decision-making involves multiple cognitive processes, which may be affected differently across neurodevelopmental disorders. In this study, we used an operant-based probabilistic reversal learning task and reinforcement learning models to investigate decision-making behavior in three mouse models of neurodevelopmental disorders: Mecp2 deficiency, Mecp2 overexpression, and Kmt2c deficiency.

Our analysis revealed distinct alterations in decision-making processes across the three models. Notably, although conventional behavioral measures did not detect significant abnormalities in Kmt2c-deficient mice, reinforcement learning analysis uncovered specific changes in the learning processes underlying their decision-making behavior.

These findings demonstrate that reinforcement learning models can reveal subtle alterations in decision-making that are not captured by conventional behavioral measures, providing a valuable approach for understanding the diversity of behavioral phenotypes associated with neurodevelopmental disorders.

James and Kerda (internship students in our lab) and Takumi (Takata Lab, RIKEN CBS / Kato Lab, Juntendo University) conducted the behavioral experiments and data analysis. We also thank Takeru Matsuda (Unit of Statistical and Mathematical Sciences, RIKEN CBS) for his invaluable contributions to the reinforcement learning analysis.