The cerebellum is a large brain structure that is important for
proper execution of movements.  Evidence suggests that the role of
the cerebellum is making feed-forward predictions, that it uses
learning to improve these predictions based on experience, and the
temporal coding is a key aspect of this learning.  I will describe
our approach to cerebellar function and how it enables the use of
biologically constrained computer simulations.  I will discuss how we
have used these simulations in attempts to understand two key aspects
of the temporal dynamics displayed by the cerebellum: 1) timing and
temporal coding and 2) learning over a limited time span.