A Bayesian perspective on the interaction between numerical and temporal perception

Abstract

This literature research analyzed to what extent Bayesian inference can explain the interaction between numerical and temporal perception. A Theory of Magnitudes (Walsh, 2003) suggests that space, time, and number all interact with each other in a generalized magnitude system. The Bayesian brain hypothesis suggests that integration of sensory input is accomplished by Bayesian inference. Utilizing three potential characteristics of Bayesian perceptual processing, cue integration, adaptation, and the central tendency effect, this research concludes that the interaction between numerical and temporal perception can be understood within a Bayesian framework. The found directionalities of interaction (uni-, bi-, and non-directional) can be explained by optimal cue integration according to the most reliable (and least noisy) cue. These findings suggest that a more liberal interpretation of ATOM could provide a more integrative perspective on the discrepancy in the literature, and in light of Bayesian theories of perception, humans behave and perceive in a statistically optimal manner.