In Pursuit of Activity Dependent Synaptic Plasticity Rules for Cerebellar Motor Learning: A Computational Study

摘要

The cerebellar circuit is one of the stereotypical structures of the brain in which learning at the behavioral level can be directly associated with plastic changes at the synaptic level. Therefore, understanding the regulatory foundations of plasticity in the cerebellum and translating them to behavioral output is of great importance. In this paper, we proposed a computational framework to identify and investigate different synaptic plasticity rules as a function of population activity in deep cerebellar nuclei and Purkinje cell. Comparing the robustness of the rules against parameter variations, the most plausible scenario was selected and applied to model gain adaptation of optokinetic reflex (OKR), a typical cerebellar-dependent task, which promotes fixation of images on the center of vision. The model could successfully regenerate the process of OKR gain adaptation confirms to the experimental data of wild-type mice. The model could also predict the effect of flocculus shutdown, immediately after the 4th training session, on long-term learning. The results of this work illustrate the great potential of population models and activity-dependent plasticity rules in reproducing a simple whole-cerebellum network, and open a possibility to predict behavioral dynamics from neural activity.