Accurate and Noninvasive Dysphagia Assessment via a Soft High-Density sEMG Electrode Array Conformal to the Submental and Infrahyoid Muscles.

Abstract

Accurate, noninvasive dysphagia assessment is important for rehabilitation therapy but current clinical diagnostic methods are either invasive or subjective. Surface electromyography (sEMG) that monitors muscle activity during swallowing, offers a promising alternative. However, existing sEMG electrode arrays for dysphagia assessment remain challenging in combining the advantages of a large coverage area and strong compliance to the entire swallowing muscles. Here, we report a stretchable, breathable, large-area high-density sEMG (HD-sEMG) electrode array, which enables intimate contact to complex surface of the submental and infrahyoid muscles to detect high-fidelity HD-sEMG signals during swallowing. The electrode array features a 64-channel soft on-skin sensing array for comprehensive data capture, and a stiff connector for simple and reliable connection to an external acquisition setup. Systemically experimental studies revealed the easy operability of the soft HD-sEMG electrode array for effortless integration with the skin, as well as the excellent mechanical and electrical characteristics even subject to substantial skin deformations. By comparing HD-sEMG signals collected from 38 participants, three objective indicators for quantitative dysphagia evaluation were discussed. Finally, a machine learning model was developed to accurately and automatically classify the severity of dysphagia, and the factors affecting the recognition accuracy of the model were discussed in depth.