Analysis of muscle fatigue using entropybased measures from the recurrence pattern of surface electromyography signals

Abstract Neuromuscular fatigue can be monitored using complexity domain analysis of surface electromyography (sEMG) signals. This study provides a technique to detect fatigue induced change in signal complexity using the entropic measurements of its recurrence pattern. For this purpose, fifteen participants have been recruited to perform isometric fatiguing exercise and signals are acquired from the biceps brachii muscle of the dominant hands. Thresholded recurrence plot is constructed from the non-fatigue and fatigue segments of the signal. Bidimensional dispersion entropy (DispEn2D) is calculated to quantify the textural change in the resulting recurrence plot. The results show that the present approach can distinguish between non-fatigue and fatigue states. Decrease in the DispEn2D values in the fatigue state shows there is an increase in periodic component of sEMG signal during fatiguing contraction.