Study On The Detection Of Motor Unit Action Potentials In Surface Electromyography

The surface electromyographic(sEMG)signal caused by voluntary muscle contraction or passive muscle stimulation is recorded non-invasively using surface electrodes directly located on the skin,it can be considered as the compound of the superposition of motor unit action potential(MUAP)trains generated by the activated motor units and various recorded noise,the surface potential distribution is influenced by the volume conductor and the signal acquisition equipments.Because of the different muscle contraction processes and force levels,the recruitment and firing behavior of motor units will also be different.For studying the functional information of motor units stably,the sEMG signal during the voluntary muscular contraction at the constant force is applied.According to the influence of the different contraction force,the sEMG signal decomposition processing was considered at the lower contraction force,and the MUAP number estimation was considered at the higher contraction force.The research results can be used to explore the controlling functional characteristic of neuromuscular system,help the diagnosis of neuromuscular diseases,and so on.The focus of this dissertation is to detect and analyze the MUAP information in the sEMG signal,then,the recruitment and firing information of motor units can be discussed.The main work and achievement of the dissertation could be presented as follows:1.The sEMG signal measurement and preprocessing.A multi-channel sEMG signal detection system was developed,and the bipolar electrode configuration was utilized.During the signal detection,the noise was inevitable,it could be affected by many factors,such as electronic devices,motion artifact,physiological structure,etc, and the noise interfere should be restricted.The signal could be separated into a series of intrinsic mode functions(IMF)and the residue by the method of empirical mode decomposition(EMD),then,the signal of noise reduction could be reconstructed by threshold processing of the IMF components and removing the residual component. The 50Hz information was contained in the sEMG signal,the spectrum interpolation was utilized to suppress the power line component.The experimental results showed that the multi-channel signal detection system worked well;the sEMG signal quality was improved and the action potential shapes were clearer.2.The study on the sEMG signal simulation.A physiological model of sEMG signal was established to help the subsequent analysis about MUAP detection.Based on the simulation of single fiber action potential(SFAP)and MUAP,the controlling function about the recruitment and firing process of motor units was modeled,some other factors were also considered in this model,such as contraction force,muscle fatigue,etc.The sEMG signals and contraction forces of different excitation levels could be simulated,and the simulation interface was accomplished.Comparing with the previous laboratory work,much progress had been made in this study.The experimental results showed that,in the time domain,the simulated superimposed MUAP shapes could be matched with the recorded MUAP shapes satisfactorily by adjusting some model parameters;in the frequency domain,the spectral variation of simulated sEMG signal could be similar with the real signal,and the muscular fatigue phenomenon during the actual prolonged contraction process could be represented by the simulated sEMG signal.In conclusion,the physiological generation process of sEMG signal could be effectively described in this model.3.The study on the sEMG signal decomposition.The major and precise MUAP information could be obtained by the decomposition method,and several blind source separation(BSS)algorithms were used.After the comparison with the separable performance of the JADE,SOBI,FastlCA,SEONS and TFBSS algorithms,the last three algorithms were selected to be applied in the sEMG signal decomposition at lower contraction force.The decomposition results of the simulated sEMG signals based on the BSS algorithm model showed that the simulated signals could be effectively separated by the algorithms,and the SEONS algorithm presented better. The decomposition results of the simulated signals based on physiological model and the real multi-channel sEMG signals showed that the MUAP information of the separated signals came from the major motor units,however,some different results could be obtained due to the complexity of sEMG signals and the algorithm theories. 4.The study on the MUAP number estimation in the sEMG signal.A satisfying result could not be acquired by a decomposition method at the higher contraction force.To be a substitute,the global firing information of the motor units could also be used to explore their recruitment and firing behavior,it could be obtained by the MUAP number estimation,because the MUAP number could be described as the product between the numbers of activated motor units and their mean firing rate.In this study,the MUAP location detection was the key and the shape did not need be cared.After the comparison with the methods of nonlinear energy operator(NEO) and scalogram analysis of continuous wavelet transform(CWT),the improved detection method based on CWT and hypothesis testing was applied in the MUAP location detection.The experimental results showed that the improved method presented better than the methods of NEO and scalogram analysis;when the contraction force enhanced,the MUAP estimation number was also increased,and a little variation of the estimation number would occur around the maximal contraction force.The research is supported the National Natural Science Foundation of China "Study on simulation and decomposition of surface electromyography"(60371015) and the Postgraduate Innovation Foundation of USTC "The influence of electrode configurations and physiological parameters on surface electromyography based on simulation study"(KD2006047).