Learning a Bayesian prior in interval timing

Task and Experimental Design Two potential sources of Perceptual Learning (PL; improvement in task performance through repetition) within the framework of Bayesian probabilistic inference [1-5] are (1) Likelihood: enhancement in sensory processing of stimuli (2) Prior: learning-induced changes in prior expectation To assess relative contributions of the prior and likelihood to PL of Interval timing (IT)[5-7], we fitted Bayesian observer models to the distributions of timing latency data while human subjects were learning a novel timing task for an extended period of time (10 days).