Logarithmic Power-Frequency: An Alternative Method for Emg-Based Fatigue Assessment

Abstract

The purpose of this study was to explore different methods of processing EMG signals with a goal of obtaining more sensitive measures of fatigue. Two groups of 28 and 16 participants performed isometric efforts in arm abduction and torso extension until exhaustion respectively at three exertion levels (30, 50 and 70% of maximum). Electromyographic signals were recorded from the middle deltoid muscle and the erector Spinae muscles at the L4/L5 level for the first and the second group, respectively. Three new parameters, peak amplitude frequency, slope of lower frequency, and slope of higher frequency, were derived from logarithmic power versus logarithmic frequency over a 20–200 Hz bandwidth. Changes of these parameters over time were found to be sensitive to fatigue and to different levels of exertion. The first two parameters were more sensitive than median power frequency during arm abductions, but less sensitive than mean power frequency. In torso extensions, the slope of higher frequency gave better performance than the slope of lower frequency. This alternative method appears to have promise for future EMG-based assessments of localized muscle fatigue.