PredOpt · Neural circuits for optimal prediction of timing

Predicting the timing of events is not only crucial to our survival but seamlessly contributes to everyday activities like riding bicycles, dancing or playing sports. Human and animal prediction of timing is statistically optimal, however, how such behavior arises from neural circuitry is poorly understood. Here I outline a theoretical framework, testable hypotheses and a set of experiments that will help uncover the neural circuits that integrate previous experience to make optimal estimates about the timing of events. In this proposal, the systematic investigation of optimality in complex forms of timing behavior is based on exploiting known neural circuits belonging to simple timing behavior. The anatomical origins of simple timing tasks like eyeblink conditioning are extremely well-characterized in the cerebellum of different species. Using this, I propose a framework for extending this circuit to understand optimal timing in a series of targeted physiology, behavior and optogenetics experiments. This research, if successful, will reveal the neural basis for how the nervous system combines previously encountered stimuli and recent observations to generate statistically optimal estimates and will generate valuable insights into the underpinnings of biological intelligence.