Global features for the estimation of discharge rate from intramuscular EMG

The use of intramuscular EMG for proportional control of prostheses requires an effective means of estimating the magnitude of neural drive to the muscles of interests. This implies the quantification of the motor unit (MU) discharge rate by which the central nervous system encodes information. Algorithms for full decomposition of signals exist, but they are time-consuming and work only at low to moderate force levels. This study investigates whether global features of the intramuscular EMG can represent the discharge rate of MU action potentials (AP). Motor unit action potential (MUAP) trains were simulated at different signal-to-noise ratios (SNR). The relationship between the number of MUAPs and global features were quantified using the coefficient of determination (R²). The same analysis was also performed on recorded intramuscular data. Results showed that MUAP discharge rate can be estimated using global features, however the estimation performance depends on the threshold for three features, and that the thresholds are SNR dependent.