Neural-latency noise places limits on human sensitivity to the timing of events

Abstract

The brain-time account posits that the physical timing of sensory-evoked neural activity determines the perceived timing of corresponding sensory events. A canonical model formalises this account for tasks such as simultaneity and order judgements: Signals arrive at a decision centre in an order, and at a temporal offset, shaped by neural propagation times. This model assumes that the noise affecting people's temporal judgements is primarily neural-latency noise, i.e. variation in propagation times across trials, but this assumption has received little scrutiny. Here, we recorded EEG alongside simultaneity judgements from 50 participants in response to combinations of visual, auditory and tactile stimuli. Bootstrapping of ERP components was used to estimate neural-latency noise, and simultaneity judgements were modelled to estimate the precision of timing judgements. We obtained the predicted correlation between neural and behavioural measures of latency noise, supporting a fundamental feature of the canonical model of perceived timing.