Real‐time assessment of swallowing sound using an electronic stethoscope and an artificial intelligence system

Abstract

Objectives: Daily assessments of swallowing function and interventions such as rehabilitation and dietary adjustments are necessary to improve dysphagia. Cervical auscultation is convenient for health care providers for assessing swallowing ability.Although this method allows for swallowing sound evaluations, sensory evaluations with this method are difficult. Thus, we aimed to assess swallowing sound by the combined use of an electronic stethoscope and an artificial intelligence (AI) system that incorporates sound recognition.Material and Methods: Herein, 20 fifth-year dentistry student volunteers were included; each participant was drank 10 ml and then 20 ml of water in different positions (sitting and supine). We developed an algorithm for indexing bolus inflow sounds using AI, which compared the swallowing sounds and created a new index. Results:The new index value used for swallowing sound was significantly higher in men than in women and in the sitting position than in the supine position. A software for acoustic analysis confirmed that the swallowing index was significantly higher in men than in women as well as in the sitting position than in the supine position.These results were similar to those obtained using the new index. However, the new index substantially differed between sexes in terms of posture compared with effective sound pressure. Conclusions:We developed a new algorithm for indexing swallowing sounds using a stethoscope and an AI system, which could identify swallowing sounds. For future research and development, evaluations of patients with dysphagia are necessary to determine the efficacy of the new index for bedside screening of swallowing conditions.

Novel Approach Combining Shallow Learning and Ensemble Learning for the Automated Detection of Swallowing Sounds in a Clinical Database

“…Nevertheless, it is noteworthy that with the technique reported here, we were able to achieve a performance comparable to theirs even when using recording data from patients with dysphagia that contained more complex swallowing-sound patterns than those in healthy individuals, and when using shallow learning [39,40]. Swallowing detection was previously possible with various sensors, such as microphones, respiratory flow, and electromyograms [25,36,[41][42][43]. In recent years, studies of swallowing-sound detection have combined high-resolution CA (HRCA) signals and deep learning.…”

Impact of neck percutaneous interferential current sensory stimulation on swallowing function in patients with Parkinson's disease: A single-arm, open-label study protocol

“…Multichannel surface electromyography can evaluate the recruitment of motor units and the effect of electrical stimulation [ 25 , 26 ]. Recently, electronic stethoscopes have been developed for the evaluation of swallowing sounds with the help of artificial intelligence [ 27 ]. These novel methods of evaluation can provide new insights on dysphagia in patients with PD and the usefulness of percutaneous neck interferential current stimulation.…”

Use of the Stethoscope to Diagnose Gastrointestinal and Hepatic Disorders

“…Cervical auscultation has been used to assess the sounds associated with swallowing and aspiration (1)(2)(3)(4)(5). The stethoscope is best positioned over the median line of the cricoid cartilage, which acts as a potential resonator.…”