A Feasibility Study of Fatigue and Muscle Contraction Indices Based on EMG Time-dependent Spectral Analysis

Abstract

Parameters derived from EMG power spectrum are beneficial methods for assessing muscle fatigue. Two important indices are median frequency and mean frequency (MDF and MNF). They have been established as the global muscle fatigue indices, particularly in static muscle contraction of biceps brachii muscle. However, a major problem of these parameters is a nonlinear relationship between muscle load/force and feature value, particularly in cyclic dynamic muscle contraction of forearm muscles. In this study, EMG feature indices that can simultaneously identify both fatigue and load are proposed. Instead of using an entire EMG signal fast Fourier transformation (FFT), a concept of using consecutive FFT has been proposed as known as the “Time-Dependent MDF and MNF” (TD-MDF and TDMNF). The success of TD-MDF and TD-MNF features used to determine muscle fatigue have been presented in a lot of research works, whereas the performance of TD-MDF and TD-MNF used to determine muscle load have been proposed in a few research works. Moreover, it has not previously been evaluated with the forearm muscles. Thus it has been investigated in this study. Four forearm muscles are evaluated including flexor pollicis longus, supinator longus, extensor carpi radialis longus and pronator radii teres. The results showed that the proposed methods have a linear relationship with muscle load in some selected range of TD-MNF for flexor pollicis longus muscle that were not found for traditional MDF and MNF. Moreover, a statistically significant difference between TD-MNF values for different loading conditions (p < 0.2) was observed. The optimal method of TD-MNF was successful when overlapping consecutive window was performed with 512-sample window size and 128-sample window increment.