Classification of Muscular Nonfatigue and Fatigue Conditions Using Surface EMG Signals and Fractal Algorithms

Abstract

The application of surface electromyography (sEMG) technique for muscle fatigue studies is gaining importance in the field of clinical rehabilitation and sports medicine. These sEMG signals are highly nonstationary and exhibit scale-invariant self-similarity structure. The fractal analysis can estimate the scale invariance in the form of fractal dimension (FD) using monofractal (global single FD) or multifractal (local varying FD) algorithms. A comprehensive study of sEMG signal for muscle fatigue using both multifractal and monofractal FD features have not been established in the literature. In this work, an attempt has been made to differentiate sEMG signals recorded nonfatigue and fatigue conditions using monofractal and multifractal algorithms, and machine learning methods. For this purpose, sEMG signals have been recorded from biceps brachii muscles of fifty eight healthy subjects using a standard protocol. The signals of nonfatigue and fatigue region were subjected to eight monofractal (Higuchi, Katz, Petrosian, Sevcik, box counting, multi-resolution length, Hurst and power spectrum density) and two multifractal (detrended fluctuating and detrended moving average) algorithms and 28 FD features were extracted. The features were ranked using conventional and genetic algorithms, and a subset of FD features were further subjected to Naïve Bayes (NB), Logistic Regression (LR) and Multilayer Perceptron (MLP) classifiers. The results show that all fractal features are statistically significant. The classification accuracy using feature subset of conventional method is observed to be from 83% to 88%. The highest accuracy of 93.96% was achieved using genetic algorithm and LR classifier combination. The result demonstrated that the performance of multifractal FD features to be more suitable for sEMG signals as compared to monofractal FD features. The fractal analysis of sEMG signals appears to be a very promising biomarker for muscle fatigue classification and can be extended to detection of fatigue onset in varied neuromuscular conditions.

What are the best indicators of myoelectric manifestation of fatigue?

“…Concerning chaos, correlation, entropy, and fractals MMF indicators, Higuchi fractal dimension and fuzzy entropy were the only indicators that had high VIP values (1.26-2.03) and changed significantly for a large proportion of the participants (65%-90%) for the anterior and medial deltoids. Higuchi fractal dimension therefore confirms that the fractal dimension is sensitive to intrinsic changes of EMG signals occurring with muscle fatigue [18,49]. As to fuzzy entropy, our results are in line with Xie et al, [54] which stated that it provides an improved evaluation of time series complexity.…”

“…In a recent study, Ni et al, [68] showed that time, frequency, and time- Consequently, the aim of the present study was to determine the best MMF indicators, f rom 15 MMF indicators identified in the literature as potentially relevant to assess muscle fatigue during a repetitive pointing task (RPT) performed with the upper limb. We hypothesized that for the anterior and medial deltoids, which are the muscles showing the largest signs of fatigue during the upper limb repetitive task used in the present studies [72][73][74], the mobility, median frequency, spectral entropy, fuzzy entropy, multiscale entropy [24,28,41,75,76], and degree of multifractality [40,47,49,53] will have greater importance to predict the evolution of the RPE. In addition, these MMF indicators will significantly change for a large proportion of the participants in the anterior and medial deltoid in comparison to approximate and sample entropy, recurrence quantification analysis, correlation dimension, and the largest Lyapunov exponent [28,40,46,55].…”

“…During sustained isometric contractions, fractal dimension was found to decrease with fatigue [18,[43][44][45] while the degree of multifractality increased during fatiguing static and dynamic contractions and was a more sensitive MMF than median frequency [40,[46][47][48]. Moreover Marri and Swaminathan., [49] demonstrated that the performance of multifractal indicators were more suitable for sEMG signals as compared to monofractal fractal dimension indicators such as Higuchi fractal dimension, in their study of the classification of muscular non-fatigue and fatigue conditions using EMG and fractal algorithms. The correlation group includes correlation dimension and recurrence quantification analysis.…”

Muscle Performance Investigated With a Novel Smart Compression Garment Based on Pressure Sensor Force Myography and Its Validation Against EMG

“…The FD methods for assessing muscle fatigue have increased in importance over the last 10 years, with researchers using different methods, such as the box-counting method ( Troiano et al, 2008 , Beretta-Piccoli et al, 2015 ; Boccia et al, 2016 ) to understand the fractal behavior. Marri and Swaminathan (2016) used several methods [e.g., Higuchi (1988) , Katz, Sevcik, box counting; multifractal analysis]. In most cases, Marri and Swaminathan’s (2016) monofractal algorithms delivered smaller FDs for fatigued muscles compared to non-fatigue; while the opposite was true for multifractal algorithms where the FD was mostly smaller than 1.…”

“… Marri and Swaminathan (2016) used several methods [e.g., Higuchi (1988) , Katz, Sevcik, box counting; multifractal analysis]. In most cases, Marri and Swaminathan’s (2016) monofractal algorithms delivered smaller FDs for fatigued muscles compared to non-fatigue; while the opposite was true for multifractal algorithms where the FD was mostly smaller than 1. In general, a signal’s FD ranges between a value of 1 and 2, i.e., between a straight line or smooth curve, and a maximally noisy signal filling up an area ( Fuss, 2013 ).…”