Feature Extraction of EMG Signals in Time and Frequency Domain for Myopathy, Neuropathy and Healthy Muscle

Neuromuscular actions of the brain generate electrical signals for muscles are Electromyograpic signals (EMG). It is dependent on the structure, composition of muscle, a peripheral nervous system and method of detection. Due to the nonstationary and the random nature of the EMG signal, it is difficult to differentiate different muscle conditions (neuropathy, myopathy and normal) only with EMG. In this paper Feature Extraction of EMG signals in Time and frequency domain is done for different muscle conditions. Time domain and frequency domain features such as peak amplitude, Root Mean Square (RMS), mean, median, variance and total peaks are extracted. Features of EMG are set up so that differentiation of muscle condition can be done only with EMG. The results show that electrical activity is more for neuropathy than healthy and myopathy muscle. A comparison of the results is done on the basis of RMS value, integration, mean amplitude and the total peaks both domains. Neuropathy can be identified using features such as variance, mean amplitude and RMS values. While myopathy can be differentiated using features total peaks.