Assessing Increasing Load Cycling Induced Muscle Fatigue Based on Multiscale Entropy Analysis of Surface Electromyography Signals

Abstract

This paper presents a multiscale entropy(MSE) based complexity analysis of surface electromyography(sEMG) signals recorded from quadriceps for assessing muscle fatigue induced by increasing load cycling. Ten male undergraduate students were recruited to perform cycle ergometer experiment with 5 step increasing load from 50 Watts to 200 Watts. The last 5 seconds sEMG signals for each load were acquired for MSE analysis. The MSE values over scale 1 to 40 and the mean value for all scales were computed to obtain MSE curves and multiscale complexity Cτ , respectively. Results show that MSE curves and Cτ decline with load increment that is consistent with the progress of muscle fatigue. The ideal scale range to separate MSE curves between neighboring loads is from 10 to 30. It was demonstrated that Cτ index outperform traditional sample entropy in its ability to capture the complex temporal fluctuations in fatigue-free state so that it is well suited for assessing muscle fatigue induced by varied-loading dynamic tasks.