Effects of short- and long-term experience on two classical measures of the multisensory temporal integration window

The relative timing between sensory signals strongly determines whether they are integrated in the brain. Two classical measures of temporal integration are provided by simultaneity judgments, where one judges whether cross-modal stimuli are synchronous, and violations of the race model inequality (RMI) due to faster responses to cross-modal than unimodal stimuli. While simultaneity judgments are subject to trial history effects (rapid temporal recalibration) and long-term experience (musical training), it is unknown whether RMI violations are similarly affected. Musicians and nonmusicians made simultaneity judgments and speeded responses to brief auditory-visual stimuli with varying onset asynchronies. We derived a so-called temporal integration window for both measures, via an observer model for simultaneity judgments and a nonparametric test for detecting observer-level RMI violations. Simultaneity judgments were subject to rapid recalibration and musicians were less likely than nonmusicians to perceive stimuli as synchronous. Proportionally, twice as many musicians as nonmusicians exhibited RMI violations within a temporal window spanning −33 to 100 ms. Response times (and RMI violations) were unaffected by rapid recalibration and modality shift costs, suggesting that rapid recalibration is not caused by changes in early sensory latency. Our findings show that perception- and action-based measures of multisensory temporal processing are affected differently by experience.