Analysis of EMG signals for automated diagnosis of myopathy

Abstract

Myopathy is a very common muscular disease in which muscle fibers do not work properly resulting in muscle weakness, stiffness, cramps etc. Clinically, analysis of Electromyography (EMG) signals plays an important role in diagnosis of myopathy. This work presents signal processing based method for automated diagnosis of myopathy from EMG signals. EMG signals collected from biceps brachii (long head) muscles were analyzed for identification of myopathy using artificial intelligence method. Basic statistical features from EMG signals were extracted and studied to find out discrimination between normal and myopathy. Artificial neural network classifier was used for identification of myopathy. Experiments were carried on a comprehensive database of EMG signal and results are encouraging. The proposed method achieved 87% accuracy with 90% sensitivity for diagnosis of Myopathy disease.