2018 Feb 19 – Bayes’ law might fit the capacity of brain circuits to encode certain prior distributions of events and use them to derive posterior probabilities. In a work published in Nature Communications, authors hypothesized that the cerebellum would be best suited to learn sub-second to second prior temporal distributions of time intervals and support Bayesian estimates. To validate their hypothesis, they elaborated and tested a mathematical model called TRACE (Temporally Reinforced Acquisition of Cerebellar Engram) that synthesizes known anatomical and physiological mechanisms of the cerebellum. TRACE performed in a manner highly consistent with Bayesian estimation theory. When human subjects were tested in two established cerebellar timing tasks, the results were consistent with the predictions of the model. This work poses the prospect of a new implication for the cerebellum.