The Swallowing Intelligent Assessment System Based on Tongue Strength and Surface EMG

Dysphagia is a very important issue in modern society, and it is common in stroke patients and the elderly. Many studies have shown that tongue strength can be used as an evaluation criterion for swallowing function. This work has implemented a methodology to assess and improvise the swallowing function of dysphagia patients. To execute the same, a tongue pressure instrument was used as a tool to assess tongue strength, and a surface electromyography (sEMG) instrument was used to collect electrical data on larynx muscles. In addition to this, an assessment task has been carried out that is combined with interesting games to increase users' willingness. After completing the task, the system collects tongue pressure and muscle electrical data. We use the scoring function calculation to quantify the user's swallowing performance quality. In addition to calculating the quality score of the motion, we also extract features from the collected data, build a variety of machine learning (ML) models to compare each model's classification effectiveness, and select the best model to correctly predict the level of the user's swallowing function. Through this evaluation system, we hope to provide fast and accurate evaluation results so that medical personnel can have more convenient and effective tools for dysphagia diagnosis and rehabilitation training.