Analysis and Discrimination of Surface Electromyographic Features for Parkinson’s Disease during Elbow Flexion Movements

Abstract

Parkinson’s disease (PD) is the second most common neurodegenerative disorder. Early intervention/treatment relies on early diagnosis of PD. There is increasing interest in methods based on electromyography measurements of PD patients because of its noninvasiveness. Thus, this study was to investigate electromyographic (EMG) characteristics of the upper limb between PD patients and healthy control subjects using EMG, and to distinguish PD patients from healthy control subjects according to the EMG information using a support vector machine (SVM) classifier. Sixteen right-handed PD patients and 25 right-handed healthy subjects participated in experiments involving elbow flexion movement. The frequency power, duration, skewness, recurrence rate, and correlation dimension of EMG signals and success rate for the right hand and the skewness of EMG signals for the left hand were found to be significantly different between the two groups. This information was subsequently used to distinguish PD patients from healthy control subjects using the SVM classifier to obtain a mean accuracy of 87.02%. Although the results may not be immediately available to use in clinical applications, the safety, simplicity and speed of the system still merits further consideration. Enhancing performance accuracy and examining PD patients in different stages of disease are anticipated in future investigations.