Presenter Name:Alyssa Murdoch

School/Affiliation:McMaster University

Co-Authors:Emily A. Wood, Andrew Chang, Dan Bosynak, Lucas Klein, Elger Baraku, Dobromir Dotov & Laurel Trainor

Abstract:

Classical ensemble musicians have the challenge of synchronizing their playing in addition to creating a shared understanding of how to interpret the musical piece, such as interpreting expressive timing variations and dynamics as a group. To predict how other musicians will play, ensembles can attend to sensorimotor signals in each other’s body sway movements. In our lab, we use motion capture to measure the body sway of musicians during performance, and Granger causality (GC) to calculate predictive information flow between the musician time-series (Chang et al., 2017, 2019). Recently, we used this technique to show that information flow in a professional string quartet decreased as they learned to play two unfamiliar pieces over eight trials on average (Wood et al., 2022). Here, we expand on this study to examine information flow during particularly expressive sections of these pieces. We picked three phrase ending sections from each of the two pieces that were characterized by expressive timing and dynamic changes. Information flow was calculated for these expressive sections for all eight repetitions of each piece. Expert string players (N = 5, 4 females) then rated the expressive sections in terms of quality, synchrony, and expressivity for each of the repetitions. Preliminary results show no correlation between any of the quality ratings and information flow. However, data collection from raters is ongoing. We are currently completing an additional cross-correlation analysis during these phrase endings sections, as well as examining information flow during sections where mistakes were made.