Electromyographic Signal Analysis Method Based On HHT And Its Application In Spasticity Assessment

The surface electromyographic signal(sEMG)is the result of comprehensive effects of the active motor units potentials,which can characterize the function and state of the neuromuscular system.The use of sEMG for spasticity detection in clinical medicine,rehabilitation medicine and other fields has significant practical values.The paper focused on the key techniques of sEMG de-noising and feature extraction,which were applied to the detection of the stretch reflex starting point of spasticity.The quantitative evaluation of spasticity finally could be achieved combing with the angle.Firstly,as to spasticity patients with different levels,signal acquisition system was designed and the relevant data collection experiments were carried out in the process of patients' elbow joint extension.As sEMG acquisition was accompanied by a lot of interference signals,this paper proposed a threshold de-noising method based on EMD entropy.Compared with the wavelet threshold de-noising method,the proposed method is proved to be effective in removing noise and retaining more details.In order to obtain the characteristics of sEMG accurately,a feature extraction method based on HHT marginal spectral entropy was proposed in consideration of the nonlinear non-stationarity of sEMG.Based on the characteristics of the changes of EMG in different functional states,combined with the changes of the signals before and after spasticity,HHT marginal spectral entropy was used to detect stretch reflex starting point.Combined with the angle information,stretch reflex threshold(SRT)could be measured.After de-noising and feature extraction of all collected signals,the SRTs of all patients were obtained.The correlation between SRTs and MAS scores was analyzed by Spearman.Final experimental results show that SRTs are significantly correlated with MAS scores(R =-0.921,*P<0.01).All above means that our method can provide an objective quantitative analysis method for clinical spasticity assessment.