Renascience has been developing the software as a medical device to evaluate the decline in swallowing function utilizing artificial intelligence (AI) in collaboration with Tohoku University. We are pleased to announce that an interview with our collaborator, Professor Hiroyasu Kanetaka of Tohoku University Graduate School of Dentistry, regarding this project, was published in the January 1, 2023 issue of Kagaku Shimbun (The Science News).
(Outline of this project)
When a healthy person eats food, it passes from the mouth through the esophagus into the stomach. However, when the swallowing function deteriorates, food and other substances enter the trachea through the mouth, which is called ‘aspiration’. Pneumonia is one of the leading causes of death among Japanese people. As the aging of the population progresses in recent years, the incidence of aspiration pneumonia, which occurs when bacteria enter the bronchi and lungs after being aspirated along with saliva and food, has been increasing. To prevent aspiration pneumonia, it is effective to diagnose a decline in swallowing function at an early stage and provide rehabilitation to improve the function. However, video endoscope and video fluorography are required for definitive diagnosis, and the specialized equipment required for the tests limits the number of medical institutions that can perform the tests and also increases the burden on the patient, including the cost.
We have noted that the organs used in “swallowing” and “speaking” share many common parts, such as the tongue, oral cavity, and pharynx, and we have been working with Tohoku University to develop the software to assist in the determination of the decline in swallowing function from speaking voice as an alternative to swallowing function evaluation. In fact, swallowing and articulation are highly related since speech therapists use vocalization and speech training to treat dysphagia. If swallowing function could be easily evaluated from voice by the AI without the use of endoscopic or fluorographic equipment, such convenient voice testing will be widely used to evaluate the swallowing function. The current method of assessing articulation however requires listening by five speech therapists, which makes it difficult to evaluate a large number of subjects, and the method may not be objective
We have completed baseline analysis of the articulation (pronunciation) by healthy subjects using our proprietary AI engine, and have classified the pronunciation by gender, age difference, individual difference, word difference, etc. We will further develop the AI to discriminate subjects with swallowing dysfunction from normal subjects by their voice.
The AI that can diagnose swallowing dysfunction from voice data during conversation will not only help reduce the workload of medical professionals such as speech therapists, doctors, and dentists by reducing their working hours, but will also provide optimal medical care by identifying hospitalized patients who need intensive care. As people age and their muscles begin to weaken, their oral functions also deteriorate, which is known as oral frailty. If patients with impaired swallowing function can be screened using the AI, it will be possible to diagnose and treat them at an early stage, resulting in the extension of healthy life expectancy in an aging society.
Please see the following link for the article in the Science News.