AbstractMusicians coordinate the upper-limbs in both temporal and spatial domains to execute sound-producing movements. From a motor control perspective, temporospatial control mechanisms underlying sound production remain poorly understood. Four-mallet marimba performance was assessed from spatial and temporal perspectives at three timepoints spaced over twelve weeks, here representing early (S1), intermediate (S2), and late phases (S3) of motor learning. Motion tracking measured movements of nine percussionists and computed acceleration of the elbows, wrists, hands, and mallets during the individual sessions. Mallets were held proximal to the pinky (mallet 1 in the left hand, mallet 4 in the right hand) and thumb (mallet 2 in the left hand, mallet 3 in the right hand). Mallet acceleration assessed temporal aspects of learning by converting data from each mallet and session into a frequency-domain representation using the pwelch method. The relative signal amplitude was extracted (i.e., relative power) to indicate relative contributions of specific frequencies to the time-series. Spatial contributions to learning were assessed via movement variability. It was hypothesized that the magnitude of learning would be greatest between S1 and S2, as demonstrated by reductions in spatial variability in upper-limb movements and temporal variability of mallet acceleration. As hypothesized, spatial variability in right and left elbow movements reduced between S1 and S2. However, temporal acceleration of mallets 2 and 3 increased from S1 to S2. Overall, percussionists may constrain movement degrees of freedom at the elbows to focus on temporal characteristics of the mallets during intermediate phases of motor learning in marimba performance.
Acknowledgments: The Canada Foundation For Innovation and the Percussion Department in the Faculty of Music at U of T