The effect of an unexpected modality on time reproduction: Clock speed or memory mixing?

The pacemaker-counter model (PCM) has been the core architecture of scalar expectancy theory for decades. PCM assumes the same timing mechanism applies to every stimulus input, regardless of modality, and has been used to explain differences between perceived durations. In violation, previous studies demonstrate a robust effect of memory-mixing, occurring when the memory trace of a previous time interval influences perception of upcoming ones. We examined the influence of unexpected modality for a trained duration on temporal reproduction, using auditory/visual stimuli with short/long (500/1000ms) standard durations—testing the PCM pure ‘clock speed’ prediction against the memory-mixing account. Here, we report different outcomes of unexpected modality on time reproduction based on modality assignments with specific intervals. When the short interval was trained with an auditory stimulus, and long interval with a visual stimulus, unexpected presence of the opposite modality led to a memory-mixing pattern, in which reproductions shifted toward the other interval. However, no such effect occurred when the short interval was trained with a visual stimulus and long interval with an auditory stimulus. We propose that durations stored in reference memory retain its paired modality, but that during averaging durations are combined based on both their modality and relative duration.