Acquisition of Multiple Prior Distributions in Tactile Temporal Order Judgment

Yasuhito Nagai, Mayu Suzuki, Makoto Miyazaki, Shigeru Kitazawa

The Bayesian estimation theory proposes that the brain acquires the prior distribution of a task and integrates it with sensory signals to minimize the effect of sensory noise. Psychophysical studies have demonstrated that our brain actually implements Bayesian estimation in a variety of sensory-motor tasks. However, these studies only imposed one prior distribution on participants within a task period. In this study, we investigated the conditions that enable the acquisition of multiple prior distributions in temporal order judgment of two tactile stimuli across the hands. In Experiment 1, stimulation intervals were randomly selected from one of two prior distributions (biased to right hand earlier and biased to left hand earlier) in association with color cues (green and red, respectively). Although the acquisition of the two priors was not enabled by the color cues alone, it was significant when participants shifted their gaze (above or below) in response to the color cues. However, the acquisition of multiple priors was not significant when participants moved their mouths (opened or closed). In Experiment 2, the spatial cues (above and below) were used to identify which eye position or retinal cue position was crucial for the eye-movement-dependent acquisition of multiple priors in Experiment 1. The acquisition of the two priors was significant when participants moved their gaze to the cues (i.e., the cue positions on the retina were constant across the priors), as well as when participants did not shift their gazes (i.e., the cue positions on the retina changed according to the priors). Thus, both eye and retinal cue positions were effective in acquiring multiple priors. Based on previous neurophysiological reports, we discuss possible neural correlates that contribute to the acquisition of multiple priors.