Automatic diagnosis of neuromuscular diseases from electromyographic (EMG) records

• Published in: 2017 5th International Conference on Electrical Engineering - Boumerdes (ICEE-B) pp. 1 - 6
• Main Authors: Sadi-Ahmed, Nafissa, Messekher, Anis, Namane, Salim, Kedir-Talha, Malika, Mekaoui, Slimane
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2017
• Subjects: ElectroMyoGraphic (EMG)) signals; Electromyography; Feature extraction; Huang-Hilbert transform (HHT); Intrinsic mode functions (IMFs); Kernel; Myopathy; Neuromuscular; Neuropathy; Support vector machine (SVM); Support vector machines; Transforms
• DOI: 10.1109/ICEE-B.2017.8192195

In this study, we proposed an analysis method of ElectroMyoGraphic (EMG) signals in order to diagnose and to identify neuromuscular pathologies (i.e.; myopathy and neuropathy). Analysis is performed fully automatically without expert assistance and without prior segmentation of muscle contractions. The method is based on Huang-Hilbert transform (HHT) which is a data-driven algorithm that decomposes the signal in a natural way without prior knowledge. From real EMG records of both healthy and pathological subjects, the proposed method selected relevant intrinsic mode functions (IMF) and then extracted the most relevant features from their HHT. Thus, EMG signals were classified with support vector machine (SVM) basing on these selected features. Mean accuracy rate of 96.08% was obtained for classification of healthy and pathological EMG signals. However, myopathic and neuropathic pathologies were discriminated with a mean classification accuracy rate of 100%. These results proved efficiency of our features selection procedure and its adequacy to recognize the pathology origin automatically.