Remote Monitoring and Detection of Amyotrophic Lateral Sclerosis Disease Through Wireless EMG Measurement System

Abstract

Amyotrophic Lateral Sclerosis (ALS) is one of the most common neuro-muscular diseases and it affects both lower and upper motor neurons in human beings. ALS requires periodic monitoring of the patient, and an early diagnosis helps prevent the disorder from progressing further, thereby improving the quality of life for patients. In this study, we demonstrated a proof-of-concept, point-of-care diagnostics system to measure the surface-EMG signal of a patient. It automatically detects whether the patient is subjected to ALS condition or not by appropriately analyzing the EMG signal. The proposed system consists of required electronics hardware in terms of sensors, front-end signal-processing circuits, and wireless modules to transmit the measured EMG signals onto a remote server through the Internet. The proposed system consists of necessary software modules, machine-learning models, and signal-processing algorithms to detect and classify whether the measured EMG signal from a patient corresponds to the given pathological condition (ALS) or not. The developed framework contains a modified VGG-5 Network, called SpinaINet-VGG, and classifies the Spectrogram of EMG signals to identify the ALS condition. The variations in the spectrogram of the EMG signals form the basis of the proposed framework. The proposed methodology for ALS disease detection involves two major segments: the generation of 2D spectrograms from Raw EMG Signals and the classification of spectrograms using SpinalN et. The performance was evaluated using popular metrics such as overall accuracy, sensitivity, and specificity and was compared with existing methods. The proposed method showed an accuracy of 96.08 %. The proposed system can be used as a point-of-care diagnostic device to help patients monitor by themselves and transmit EMG signals over the Internet to a remote server so that a physician can use the measured data for remote diagnosis. © 2024 IEEE.