ALS Detection Based on T-Location Scale Statistical Modeling of the DWT Coefficients of EMG Signals

Amyotrophic lateral sclerosis (ALS) is a motor neuron disease (MND) responsible for affected people to be paralyzed and unable to talk, walk, swallow or breathe over time. Electromyography (EMG) is an investigation procedure to record and evaluate the electrical activity made by skeletal muscles. ALS disease can be easily identified from the recorded EMG signals. In this paper, discrete wavelet transforms (DWT) and statistical modeling of those attained DWT coefficients have been performed to detect ALS disease. T-location scale distribution has been chosen to statistically model the DWT coefficients and justified by Chi-square goodness of fit test. The modeling parameters are extracted from the probability density function (PDF) to prepare the feature set to feed several classifiers. Goodness of features has been justified by figure-of-merits for inter-class distance and intra-class compactness. The proposed method is found capable of making greater performance outcome while evaluating compared to some state-of-the-art processes.