A mechanistic account of how our environments shape how we learn is critical for understanding individual differences in behavior. In this talk, I will propose that specific statistics of experienced environments shape reward learning across multiple nested timescales via a meta-learning process, focusing on the example of environmental controllability. The ...
