Complex dynamics in the granual layer of the cerebellum: Large-scale computational reconstructions

Realistic modeling of the cerebellum granular layer has remained an unresolved issue. Now the availability of detailed single neuron codes and of a wealth of experimental results on single cell and network functions in vitro and in vivo has allowed reconsidering the issue. We have constructed a detailed computational model of the cerebellum granular layer and mapped it against several cellular and network constraints including: 1) center-surround organization 2) time-windowing 3) non-linear filtering 4) theta-frequency oscillations the model, purely based on microscopic properties (connectivity and single-neuron functions) has been able to reproduce these four higher-level effects. Therefore, the granular layer model is proving a powerful tool for predicting cerebellar computations.