Some experimental findings are inconsistent with the dominant model of reward prediction error, highlighting the need for alternative testable and falsifiable models for dopamine function. Vijay Namboodiri explains: