Spontaneous bimanual independence during parallel tapping and sawing

Abstract

The performance of complex polyrhythms-rhythms where the left and right hand move at different rates-is usually the province of highly trained individuals. However, studies in which hand movement is guided haptically show that even novices can perform polyrhythms with no or only brief training. In this study, we investigated whether novices are able to tap with one hand by matching different rates of a metronome while sawing with the other hand. This experiment was based on the assumption that saw movement is controlled consistently at a predictable rate without the need for paying primary attention to it. It would follow that consciously matching different stipulated metronome rates with the other hand would result in the spontaneous performance of polyrhythms. Six experimental conditions were randomised: single handed tapping and sawing as well as four bimanual conditions with expected ratios of 1:1 (performed with and without matching a metronome) as well as 3:4 and 4:3 (performed matching a metronome). Results showed that participants executed the saw movement at a consistent cycle duration of 0.44 [0.20] s to 0.51 [0.19] s across single and bimanual conditions, with no significant effect of the condition on the cycle duration (p = 0.315). Similarly, free tapping was executed at a cycle duration of 0.48 [0.22] s. In the bimanual conditions, we found that for a ratio of 4:3 (4 taps against 3 sawing cycles per measure), the observed and predicted ratio of 0.75 were not significantly different (p = 0.369), supporting our hypothesis of the spontaneous adoption of polyrhythms. However, for a ratio of 3:4 (3 taps against 4 sawing cycles per measure), the observed and predicted ratio differed (p = 0.016), with a trend towards synchronisation. Our findings show that bimanual independence when performing complex polyrhythms can in principle be achieved if the movement of one hand can be performed without paying much-if any-attention to it. In this paradigm, small rhythmic arm movements are possibly driven by an intrinsic timing which leads to spontaneous convergence on a cycle duration of around 0.5 s, while the movement of the other hand can be controlled consciously to occur at desired rates.

Fig 1. Study setup, illustrating the dual movement task where participants saw into a piece of wood while simultaneously tapping on a box to the beat of a metronome.

Data from accelerometers attached to saw and box (inset) are recorded wirelessly.

Fig 2

Left: boxplot for the ratios of tapping to sawing cycle duration calculated for the four dual movement tasks. Right: Absolute difference (‘ABS’) between expected and observed ratio for each participant. Experimental conditions: ‘synch’–synchronised tapping and sawing (1:1), expected since the metronome beat was set according to the preferred sawing cycle duration; ‘4:3’—4 beats tapping against 3 beats preferred sawing cycle duration per measure; ‘3:4’—3 beats tapping against 4 beats preferred sawing cycle duration per measure. Dashed line: ratio 1:1 (identical cycle duration for tapping and sawing); dotted lines: ratio 0.75 for condition ‘4:3’ (tapping duration is shorter than sawing duration) and 1.33 for condition ‘3:4’ (tapping duration is longer than sawing duration). Please note that condition labels are given as the equivalent of a frequency for ease of understanding the experimental condition.

Fig 3. Boxplots for the ratios of the set metronome beat to tapping cycle duration (left) and sawing cycle duration (right) calculated for the three dual movement tasks where a metronome beat was matched through tapping.

For experimental conditions and legend, please refer to Fig 2. The three large outliers for tapping represent participants tapping at around twice the set metronome beat, which was confirmed by video footage.

Fig 4. Boxplots showing within-participant variation as an index of stability.

Top: within-participant IQR for the hand performing tapping (left) and the hand performing sawing (right), for both the single handed task (‘single’) and the dual movement tasks (remainder). Bottom, left: participant-specific difference between within-participant variation for the hand performing tapping and the hand performing sawing. Dashed line: no difference in variation. Bottom, right: participant-specific ratio of the variation observed during the single hand task compared to the equivalent action in the four dual movement tasks. Dashed line: no difference in variation. A value of 0.5 corresponds to half the amount of variation in the single hand task, a value of 2 corresponds to twice the variation. For further details on experimental conditions and legend, please refer to Fig 2.

Fig 5

Top: relationship of the median absolute difference between observed and expected cycle duration for the hand engaging in sawing compared to the hand engaging in tapping. Most of the deviation from the expected ratio is due to the difference between expected and observed sawing cycle duration. Bottom: median absolute difference between observed and expected cycle duration (left) and within-participant variation thereof (right) for sawing and tapping. Complementary to the data presented in Fig 3, the deviation from the expected cycle duration is larger for sawing, however the within-participant variation in the deviation is similar for sawing and tapping.