Predictive Modeling of Decision-Making in Mouse Virtual Reality Behavior Task

Decision-making and reinforcement learning (RL) are processes of learning that play significant roles in our cognitive functioning. In cognitive and learning disorders, such as Autism Spectrum Disorder, learning rates and decision-making abilities might be impaired. Thus, through research focused on understanding the driving factors behind healthy learning and comparing this to diseased states, we can further the development of treatment for such disorders.

In order to identify what behavioral factors drive value-based decision-making, we had mice perform in a VR binary choice task. We then fitted mouse performance data with predictive models for RL to identify behavioral factors that drive this learning process and their individual weights. Future applications of fiber photometry and 2-photon-imaging will allow for a deeper understanding of the neural mechanisms behind RL.

Responsibilities for this project included data collection and management, as well as conducting daily animal training sessions with over 20 mice. Furthermore, I aided in the construction of the set-up used in the project and troubleshooting of the VR behavior task.