Muscle fatigue detection during dynamic contraction under blood flow restriction: Improvement of detection sensitivity using multivariable fatigue indices

At present, in order to detect peripheral muscle fatigue, it is customary to perform a frequency domain analysis using surface electromyography data measured with bipolar leads, and extract characteristic quantities like mean power frequency and median frequency. However, there are times when muscle fatigue variables, typically represented by decreasing wave frequencies, do not occur under certain conditions such as dynamic contraction causing a blood flow restriction. This study has tried to improve muscle fatigue detection sensitivity by utilizing two approaches: alteration of electrode arrangements to monopolar leads and use of nonlinear analysis in addition to frequency domain analysis and time domain analysis. The multivariable fatigue indices calculated by using each analysis method are combined into one synthetic variable using the method named “end to end projection” proposed by Rogers and Macisaac. After the effective indices and combination patterns have been investigated, it was showed that the optimum number of combinations is around four. The results also indicate that the combination method has a potential to improve the detection accuracy and monopolar leads can detect muscle fatigue change more clearly than bipolar ones.