A novel approach to the segmentation of sEMG data based on the activation and deactivation of muscle synergies during movement

The muscle-synergy hypothesis is a widely used method for describing simple motion patterns and how the central nervous system deals with the complexity of controlling a large set of muscles in parallel. However, the physiological interpretation of synergies and the mathematical techniques behind their computation are still far from being properly standardized. This letter proposes a novel approach to obtaining detailed and accurate information about how muscular synergies are triggered during different stages of a given motion. In that regard, a way to find the number of muscle synergies working in parallel is introduced based on the appearance of unexpected high-frequencies in the time series of electromyographic (EMG) signals during synergy extraction. This phenomenon allows the definition of a robust threshold for the number of synergies. The proposed methods are tested on muscle synergies computed from superficial EMG signals recorded during wheel steering. The results show the advantage of the proposed approach in relation to the currently used methods that require segmenting the signal according to the stages of the movement, and suggest future increases in the understanding of motor control from a neurological point of view. © 2016 IEEE.