Title: EMG signal classification using ANN and ANFIS for neuro-muscular disorders

Authors: Pallikonda Rajasekaran Murugan; Stephy Mariam Varghese

Addresses: Department of Electronics and Communication Engineering, Kalasalingam University, Tamil Nadu 626126, India ' Department of Electronics and Instrumentation Engineering, Kalasalingam University, Tamil Nadu 626126, India

Abstract: Neuro-muscular disorders can be caused by immunological and autoimmune disorders. Electromyography (EMG) may aid with the diagnosis of nerve compression, nerve rest injury, with other problems of muscles and nerves. Electromyography signal studies the electrical activity of the muscles and forms a valuable neurophysiologic test for the assessment of neuromuscular disorders. When the numbers of Motor Unit action Potentials (MUPs) are increased, it is very difficult for the neurophysiologist to distinguish the individual waveforms. Thus the classification of the EMG signal becomes necessary. The existing quantitative analysis methods have several limitations such as lower recognition rate of MUP waveforms, sensitive to continuous training, less accuracy and provide inconsistent output. In this paper, in order to classify the tasks, the EMG signals are trained using soft computing techniques like Artificial Neural Network (ANN) and Adaptive Neuro Fuzzy Inference System (ANFIS). ANFIS combines both neural networks and fuzzy logic principles; it can capture the benefits of both in a single framework. The principal component analysis and wavelet transform are used as dimensionality reduction methods and the classification is done by using the soft computing techniques. These proposed techniques have advantages like higher recognition rate, insensitivity to overtraining, higher reliability and accuracy.

Keywords: signal classification; electromyography; neuromuscular disorders; artificial neural networks; ANNs; adaptive neuro fuzzy inference system; ANFIS; fuzzy logic; EMG signals; principal component analysis; PCA; wavelet transform; dimensionality reduction.

DOI: 10.1504/IJBET.2014.065657

International Journal of Biomedical Engineering and Technology, 2014 Vol.16 No.2, pp.156 - 168

Received: 06 Feb 2014
Accepted: 18 Aug 2014

Published online: 25 Apr 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article