CYCLOSTATIONARY ANALYSIS OF ELECTROMYOGRAPHIC SIGNALS

Mean firing rate estimation is an important step in electromyographic (EMG) signals analysis. Its application is of great interest for the conception and implementation of algorithms in various research domains, ranging from neuromuscular diseases diagnosis to biomechanics. The proposed work is focused on the study of the intrinsic cyclostationary properties of single motor unit action potential train. It must be of direct interest to provide information about the neuromuscular command. It is shown that individual motor unit firing rates can be better estimated using second order cyclostationary analysis than traditional statistical tools, such as Fourier transform. After a brief state-of-the-art on cyclostationary analysis of EMG signals, the basic concepts and measures of cyclostationarity are presented. Follows a presentation of the EMG model. Results after application of the cyclostationary analysis tools for simulated and real data are provided next. A discussion on the obtained results concludes this work.