Multisensory temporal numerosity judgment

Abstract

In temporal numerosity judgment, observers systematically underestimate the number of pulses. The strongest underestimations occur when stimuli are presented with a short interstimulus interval (ISI) and are stronger for vision than for audition and touch. We investigated if multisensory presentation leads to a reduction of underestimation. Participants were presented with 2 to 10 (combinations of) auditory beeps, tactile taps to the index finger and visual flashes at different ISIs (20 to 320 ms). For all presentation modes, we found underestimation, except for small number of pulses. A control experiment showed that the latter is due to a (cognitive) range effect. Averaged over conditions, the order of performance of sensory modalities is touch, audition and last vision. Generally, multisensory presentation improves performance over the unisensory presentations. For larger ISIs (160 and 320 ms), we found a tendency toward a reduction in variance for the multisensory presentation modes. For smaller ISIs (20 and 40 ms), we found a reduction in underestimation, but an increase in variance for the multisensory presentation modes. In the discussion, we relate these two findings to Maximum Likelihood Estimation (MLE) models predicting that multisensory integration reduces variance.