Classification of Myopathy and Amyotrophic Lateral Sclerosis Electromyograms Using Bat Algorithm and Deep Neural Networks.

Electromyograms (EMG) are a recorded galvanic action of nerves and muscles which assists in diagnosing the disorders associated with muscles and nerves. The efficient discrimination of abnormal EMG signals, myopathy and amyotrophic lateral sclerosis, engage crucial role in automatic diagnostic assistance tools, since EMG signals are nonstationary signals. Hence, for computer-aided identification of abnormalities, extraction of features, selection of superlative feature subset, and developing an efficient classifier are indispensable. Initially, time domain and Wigner-Ville transformed time-frequency features were extracted from abnormal EMG signals for experiments. The selection of substantial characteristics from time and time-frequency features was performed using bat algorithm. Extensively, deep neural network classifier is modelled for selected feature subset using bat algorithm from extracted time and time-frequency features. The performance of deep neural network exerting selected features from bat algorithm was compared with conventional artificial neural network. Results demonstrate that the deep neural network modelled with layers 2 and 3 (neurons = 2 and 4) using time domain features is efficient in classifying the abnormalities of EMG signals with an accuracy, sensitivity, and specificity of 100% and also exhibited finer performance. Correspondingly, the developed conventional single layer artificial neural network (neurons = 7) with time domain features has shown an accuracy of 83.3%, sensitivity of 100%, and specificity of 71.42%. The work materializes the significance of conventional and deep neural network using time and time-frequency features in diagnosing the abnormal signals exists in neuromuscular system using efficient classification.

Automated diagnosis of amyotrophic lateral sclerosis using electromyograms and firefly algorithm based neural networks with fractional position update.

Amyotrophic Lateral Sclerosis (ALS) is a disorder of the neuromuscular system that causes the impairment of nerve cells from brain to spinal cord and to the voluntary muscles in every part of the human physiological system, which totally leads to paralysis. The examination of ALS using Electromyograms (EMG) is a challenging task which requires experts to investigate and diagnose. Hence, the development of an efficient and automated procedure is significant for the analysis of ALS signals. In this work, eighty time-frequency features were extricated from EMG signals transformed into time-frequency images. Further, fifteen highly substantial features were chosen using the firefly algorithm with fractional position update. Further, fractional firefly neural network is introduced and developed to examine the EMG signals. The performance metrics of the fractional firefly based neural network diagnostic system were analyzed with different fractional orders (α) and hidden neurons. Results demonstrated that the proposed technique is highly efficient and yields good statistical significance. Further, the accuracy of the fractional firefly neural network classifier with α = 0.5 and 15 hidden neurons is higher (93.3%) when compared to the accuracy of the classifier with different α values and hidden neurons. The proposed fractional order-based feature selection algorithm and classifier model are highly suitable for development of systems for evaluation of ALS and normal EMG signals, since the proficient discrimination of normal and ALS EMG signals is essential for the identification of neuromuscular disorders.