Fractals, chaos and entropy analysis to obtain parametric features of surface electromyography signals during dynamic contraction of biceps muscles under varying load

Abstract

Objective: The purpose of this paper is to illustrate the findings of non-linear dynamic analysis techniques in characterization of the Electromyography Signal during biceps muscle contraction under varying loading conditions. The variation of Chaos, Fractals and Entropy of the EMG signal is represented for biceps muscle contraction under varying load to reveal the underlying system dynamics. Methods: The EMG signal is acquired through a developed EMG signal acquisition circuitry to obtain the time series for offline data analysis. We have used the tools of Non-Linear dynamics to calculate Fractal dimension, Chaos and Entropy. We have implemented some specific algorithms to obtain the optimum parameters required for successful estimation of Chaos, Fractal Dimension and Entropy of surface EMG signal during dynamic muscle contraction. Results: The presence of deterministic chaos was clearly evident in case of arm flexion with different loads. The complexity of the signal, as evident from the fractal dimension calculation, revealed enough information about the complexity associated during arm flexion condition. Further application of the Entropy estimate helped us to estimate the change in uncertainty or rate of information transfer with increased loading of muscles. Significance: Since the non-linear dynamics technique proves to be an efficient tool to address the changing dynamics associated with muscle contraction, it can help in quantitative assessment of muscular activity. Therefore we propose here a technique of biomedical signal processing and analysis to be effectively applied in EMG signal analysis and interpretation.