Correction of EMG Frequency Spectrum Using Physical Characteristics for Muscle Fatigue Estimation

Abstract

A method using electromyograms to estimate muscle fatigue associated with exercise has been proposed. Muscles can be classified into three types: fast, intermediate, and slow, each of which being activated by physiological metabolic mechanisms that correlate with fatigue. The percentage of muscle fibers in use, which correlates with muscle fatigue, can thus be calculated using frequency analysis. In some participants, this muscle fatigue analysis method failed because actual muscle fatigue did not correlate with percentage of muscle fibers used. Assuming that this phenomenon was due to the characteristics of the subcutaneous fatty tissues between the muscle and skin surface, we attempted to cluster the participants based on their physical characteristics, construct a correction model using physical characteristics as explanatory variables within the clusters, and correct the percentage of muscle fibers used. Consequently, the muscle fatigue in all seven participants, which initially deviated from the ratio of muscle fibers used, was adjusted to align with the expected pattern, thereby enhancing the validity of the muscle fatigue analysis.