Research On SEMG Software System And Non-linear Analysis Technique

Nervous system is a neural network, consisted of billions of nervous cells connected with each other by electric or chemical synapses. Non-linear characters are presented in it. Surface Electromyography(sEMG) can reflect neuromuscular function activity level in some degree. Different from linear analysis, non-linear analysis can excavate the core of neuromuscular, explore new road for the feature extraction of sEMG.In the thesis, a sEMG analysis and bio-feedback software system was designed and a study based on non-linear analysis using the sEMG recorded from the system was elaborated. The contents can be concluded into three part.First of all, based on existing hardware, an interactive software system was designed, which include collection of sEMG, analysis and evaluation, and bio-feedback treatment three main functions. The custom collection and analysis module can do time domain, frequency domain, time-frequency domain and cooperativity analysis. Secondly, an algorithm was presented. The study using LZ complexity, improved LZ complexity, Sample entropy, Fuzzy entropy parameters to detect the differences between healthy persons and stoke patients on their tibialis anterior muscle sEMG during the maximum isometric voluntary contraction. At last, the law of muscle fatigue was researched. Based on multi-fractal theory, the fatigue sEMG characters were presented. A magnitude in spectrum of multifractal was found and a new parameter was brought out to measure fatigue.The thesis used software engineering to design and complete a software system, 60 stroke patients clinical test was conducted to prove the effectiveness of the software. Analyzing the non-linear characteristic of sEMG, and exploring the possible physiological mechanism. The results show that IZ complexity, improved LZ complexity and Fuzzy entropy can distinguish the healthy sEMG and the patients sEMG on the both good and impaired side (P<0.001), Sample entropy is related to the brunnstrom stage of stroke patients(P<0.05), A magnitude in spectrum of multifractal was found and a new parameter can measure fatigue. The findings manifest that ccmplexity and entropy may reflect the number of activated Motor Unit and their discharge frequency, they have potential clinical value in the diagnose and evaluation of stoke patients and muscle fatigue.The results above have a great potential in clinical application. It lays the foundation of further exploration.