[P2] Poster: Day 2

[P2-23] Distributional Variability Increases Uncertainty in Mean Duration Judgments

*Taku Otsuka^1,2, Hakan Karsilar¹, Hedderik van Rijn¹ (1. University of Groningen (Netherlands), 2. The University of Tokyo (Japan))

Keywords:

distributional variability,contextual effect,Bayesian,EEG,contingent negative variation

Prior studies on contextual effects in duration perception have focused on how current perception is influenced by traces of past stimuli. However, real-world performance often requires extracting and retaining summary statistics, such as the mean and variance, of temporal distributions. For example, in baseball, it is advantageous for a batter to estimate the average speed of pitched balls and the variability around this mean to prepare for the next game. In order to investigate such summary representations in time perception, we explicitly instructed participants to estimate the mean duration of stimulus distributions. Critically, these distributions had identical means but differed in their variability. We found that the variability of participants’ mean estimates increased with the variability of the distributions, even though the actual mean remained constant. We further examined how this variability-related effect was reflected in EEG signals during the task. The contingent negative variation (CNV) correlated not only with single-trial reaction times but also with the extent to which participants’ mean estimates were influenced by the distributions’ variability. Conversely, the post-interval P2 component was associated with the perceived duration of the current stimulus. These findings suggest that while humans can accurately estimate the mean of a temporal distribution, the uncertainty of this representation increases as distributional variability increases, as reflected in the preparation-related CNV during temporal judgments.