Estimating muscle fatigue of the biceps brachii using high to low band ratio in EMG during isotonic exercise

The principal objective of this study was to estimate biceps brachii muscle fatigue using the fatiguing time (FT) and the initial slope (S_init) of the high to low frequency band ratio (H/LFB) in EMG during isotonic exercise until the endurance time (ET) at which the subject could no longer follow the fixed speed. Surface EMG was measured on the biceps brachii muscle of ten subjects (5 males and 5 females) as the subject performed repetitive elbow contraction tasks. The first task was with no load (“no load” task). The other tasks were performed until exhaustion (“load” task) with a load weighted at 10%, 20% and 30% of the subject’s maximum voluntary contraction (MVC). We extracted decreasing parameters of the EMG, such as median frequency (MDF) and three types of H/LFB (H1; 95∼500Hz, H2; 150∼350Hz, H3; 470∼500Hz and LFB; 15∼45Hz). Statistical analysis was conducted to select an appropriate parameter. The FT was defined as the time at which the selected H/LFB became smaller than the threshold (50% at 10% MVC, 60% at 20% MVC, and 65% at 30% MVC) with respect to the initial value, and S_init was defined as the slope of the linear regression line using H/LFB from the onset of exercise to FT. In order to predict ET, we proposed using an analytical method that employed a multiple linear regression model with two independent variables, FT and S_init. The results from this novel model were compared with those of previously established methods. In the “no load” task, there was no FT in the MDF and three H/LFBs. In the “load” task, the MDF and three H/LFBs gradually decreased, but the H3/LFB decreased most rapidly. A significant difference between the “no load” task and the “load” task at a 95% significant level was only observed in H3/LFB. ETs of all subjects were predicted at an error of approximately 30.4%. This error was better than that obtained using other methods (34.8% by Mathiassen’s method, 39.7% by Ma?setti’s method). FT detected from H3/LFB was a useful indicator to distinguish the EMG in the “load” task from that observed in the “no load” task, and the ET of a subject could be predicted using FT and S_init. Therefore, the H3/LFB provides more information on muscle fatigue than the other features under isotonic conditions, and may be suited to specific experimental protocols in workplace fatigue studies.