Computer aided detection of abnormal EMG signals based on tunable-Q wavelet transform

The information present in the electromyogram (EMG) signals can be used for the diagnosis of the neuro-muscular abnormalities such as: amyotrophic lateral sclerosis (ALS) and myopathy. In this paper, a technique for detection of ALS and myopathy is presented, which is based on tunable-Q wavelet transform (TQWT). For the purpose of detection of these abnormalities, motor unit action potentials (MUAPs) are extracted from the EMG signals. Different entropy features computed from sub-bands obtained using TQWT along with time-domain based features are used for classification of MUAPs. The classification is performed using random forest classifier. The results obtained from proposed methodology show the effectiveness of the technique to distinguish ALS and myopathy signals.