Exploring the effect of training on muscle synergies and kinematics of a task

In this preliminary study we aim to extract muscle synergies in spiral tracking task when performed with the non-dominant hand, and investigate the effect of training on muscle synergies. Three young subjects who were university students (all Females, 21±1 years old, right-handed) volunteered to participate in this study. The subjects were healthy with no history of neuromuscular disease. Each subject performed one of the three tasks chosen randomly in five sessions. The tasks were to track either a circle in a limited time, or a circular spiral or a square spiral. During the task, the trajectory of the pen and EMG signals from eight muscles of the non-dominant hand were simultaneously recorded. Afterwards, the EMG signals were factorized using the Non-Negative Matrix Factorization (NMF) method and synergy modules were extracted. While extracting muscle synergies using NMF, we used the Variance Accounted For (VAF) criteria in order to determine the optimized number of modules. Three synergy modules were determined as the optimum number for the primary EMG signals reconstruction. The P index was introduced to compare the changes in muscle synergies during the training. This index stands for the correlation between the muscle synergies of the trials of each session with the average of muscle synergies in the same session. The results show that mean square error between the desired and the actual trajectories, regardless of the type of training, has decreased for all three subjects. This indicates that the subjects have learned the task during the training sessions. This was incorporated as an index of participants' skill. Moreover, changes in the P index correspond noticeably with subjects tracking skill. In addition, this index can be adjusted during training and there have been noticeable changes in its value between the sessions. The results convey that central nervous system (CNS) adjusts muscle synergies defining given motor behaviors to the desired synergy space, which is the average of synergies, during the training.