The spatial-tapping task to reveal the coexistence of event-based and emergent timing for the control of rhythmic sequences

International audienceThe control of rhythmic motor sequences may involve two distinct timing processes, i.e. event-based and emergent timing. Event-based timing refers to the mode of action control in which the task goal is to maintain timing accuracy (Wing & Kristofferson, 1973a, 1973b), while in emergent processes, the timing emerges from the dynamics of control of the spatial trajectory (Robertson et al., 1999; Turvey, 1977). These timing modes have been revealed through finger tapping and circle drawing tasks, respectively (Zelaznik, Spencer, & Doffin, 2000; Zelaznik, Spencer, & Ivry, 2002). In the present study, we used a hybrid-pointing task in order to assess whether the two modes could co-exist within a unique movement, as suggested by Repp & Steinman (2010). Sixty-eight participants performed a spatial-tapping task in which they were instructed to produce discrete tapping actions around a circular trajectory, across nine distinct tempi (1100 to 300 ms of inter-onset-interval). Autocorrelation functions (AC) of the inter-response-intervals were calculated up to ten lags to reveal series dependencies. Significant negative AC-1 were revealed at tempi ≥ 700 ms, suggesting that the timing was event-based at these tempi, and significant positive AC-6 were revealed at tempi ≤ 500 ms, suggesting that the timing was emergent at these slow tempi. Furthermore, an analysis of the spatial errors indicated that the timing errors were the smallest between 1100 to 900 ms of IOI, intermediate between 800 to 600 ms of IOI, and the largest between 500 to 300 ms of IOI, pattern that follows the index of difficulty of the task. Finally, between 600 to 300 ms of IOI the endpoint distributions were significantly more oriented in function of the tangent to the circle, with the emergence of an anchor point in the spatial trajectory, suggesting that the task goal at faster tempi was to smooth the tapping actions within a global circular pattern rather than maintaining timing accuracy per se (see Roerdink, Ophoff, Peper, & Beek, 2008 for descriptions of the anchoring phenomenon). Overall, our results suggest that for sequential motor control, two different timing modes can be used in function of task constraints. Autocorrelation analyses suggest that event coding is used at slower tempi (≥ 900 ms), and that an emergent timing mode is used at faster tempi (≤ 500 ms). For intermediate tempi, the temporal pressure was higher and the control was maintained event-based between 800 to 700 ms of IOI, with however a significant decrease in subjects’ performances. A combination of modes was revealed around 600 ms of IOI. Hence, we conclude that our results argue in favour of the coexistence of the two timing within the same motor sequence as a balance, with one mode taking over when timing is the priority (event-coding at slow tempi) and the other being dominant when the spatial aspect of the task is set as the priority (emergent-coding at fast tempi)