Bandwidth optimization of the fatigue index to estimate muscle fatigue during dynamic contractions

In this study, a filter-based fatigue index (FIhlrOPT) was optimized to maximize the correlation with peak power. Cut-off frequencies of high-pass-filter were optimized to determine bandwidth for fatigue estimation. Thirty two subjects performed 5 sets of 10 knee extensions with 2 minutes of rest between each set. Surface electromyography (sEMG) signals were obtained from rectus femoris (RF), vastus medialis (VM), and vastus lateralis (VL) muscles for each subject. The correlation of the proposed index was compared with sEMG-based fatigue indices, such as the mean root-mean-square (RMS) values, median frequency, median frequency after continuous wavelet transform, Dimitrov spectral index and Gonzalez-Izal wavelet index. The correlation of the index proposed in this study was higher (r = 0.81 ± 0.08 (RF), 0.58 ± 0.25 (VM & VL)) than those of the cited fatigue indices. The optimized cut-off frequencies in this study (353.30 ± 49.45 Hz (RF), 348.71 ± 34.28 Hz (VM & VL)) reflects the decreased firing rate during fatigue, and thus cut-off frequencies were appropriate for fatigue estimation.