Fig. 6.

Bayesian model predictions of CCE as function of the causal prior and temporal disparity τ_O − τ_A. a The different effects of the causal prior on CCE across the three conditions is evident even with equal causal priors, which means that CCE depends on outcome uncertainty. b When plotted as functions of the temporal disparity τ_O − τ_A, given the condition-dependent causal priors (marked by the colored dots in a), CCE falls off with the disparity of sensory inputs from the prior mean, |τ_O − τ_A − μ_AO|, faster when the outcome uncertainty is higher