Research on Intelligent Diagnosis Method of Swallowing Signal Based on Complex Electrical Impedance Myography

• Published in: IEEE sensors journal Vol. 25; no. 4; pp. 5969 - 5977
• Main Authors: Chu, Xu, Yu, Shaoshuai, Zhang, Fu, Yang, Yuxiang, Fu, Letian, Liu, Qi
• Format: Journal Article
• Language: English
• Published: New York IEEE 15.02.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Amplifiers; Amplitudes; Assessment of swallowing disorders (ASDs); Biological activity; complex electrical impedance myography (C-EIM); Demodulation; Diagnosis; Diseases; Disorders; Electrical impedance; Esophagus; GA-generalized regression neural network (GRNN); Impedance; integer-period digital lock-in amplifier (IPD-LIA); Interference; Lesions; Lock in amplifiers; Motion perception; Neural networks; Neuromuscular; Neuromuscular diseases; Pharynx; Recognition; Sensors; Swallowing
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2024.3519559

Assessment of swallowing disorders (ASD) is the first step to diagnosing whether the patient's swallowing function is normal or not. Accurate ASD methods can reduce the incidence rate of swallowing disorders. Electrical impedance myography (EIM) is an early diagnosis method of neuromuscular diseases for muscle electrical impedance detection. It can diagnose the early lesions of neuromuscular tissue and bring new hope for the early diagnosis of swallowing disorders. However, the traditional EIM technology only detects the amplitude information of biological electrical impedance and cannot collect the phase information, which cannot comprehensively and dynamically record the swallowing process. Based on the traditional EIM technology, this article proposes a dynamic detection and recognition method of pharyngeal complex EIM (C-EIM) based on the integer-period digital lock-in amplifier (IPD-LIA). First, this method proposes and designs a C-EIM hardware detection system based on IPD-LIA. Then, the C-EIM system is used to dynamically record the relevant amplitude and phase information of the subjects' swallowing events. Finally, the amplitude and phase information are combined with the GA-generalized regression neural network (GRNN) intelligent algorithm to identify swallowing events. The experimental results show that the method based on the combination of the C-EIM system and GA-GRNN can effectively reduce the interference of motion artifacts on swallowing events and the recognition rate of 95.2% for swallowing events, which is much higher than traditional EIM detection methods. This lays a theoretical foundation and technical guidance for the subsequent evaluation of swallowing disorders.