A foreign study has found that it is possible to diagnose early Parkinson’s disease patients through artificial intelligence (AI) that can analyze changes in language habits.
Although Parkinson’s disease cannot yet be cured, it can be managed to some extent if detected early, so early diagnosis is crucial. The research team plans to see if it can be used at home via a smartphone app in the future.
On the 30th, a research team from Kounus University of Technology (KTU) and Lithuanian University of Health Sciences (LSMU) in Lithuania said that changes in language habits could be an early sign of Parkinson’s disease, and developed a method to identify early symptoms of Parkinson’s disease using voice data. The researcher explained that AI can be used to analyze and evaluate voice signals to make a diagnosis in just a few seconds.
The results of the research were recently published in ‘Applied Science’ published by MDPI, an international academic journal.
Parkinson’s disease is a degenerative brain disease caused by the loss of various nerve cells, such as dopaminergic neurons in the brain. Symptoms can vary from patient to patient, but in general, symptoms such as hand tremors, muscle stiffness, and loss of balance appear.
The research team explained that as exercise capacity decreases in Parkinson’s patients, vocal cords, diaphragm, and lung function may also decrease. This is evidence that early symptoms can be identified by verbal command.
Indeed, according to the research team, early Parkinson’s patients speak more quietly. It can also be monotonous, slow and less expressive. However, as these changes appear to be very small, it is very difficult to notice in the ear. As Parkinson’s disease progresses, symptoms such as hoarseness, stuttering, slurred speech, or loss of distinction between words can become severe.
Considering these symptoms, the research team developed a system that can detect diseases early. In fact, the link between Parkinson’s disease and language skills is not new. This is an area that has been known and studied since the 1960s. However, as technology advances, it is possible to extract and analyze more information through voice signals in just a few seconds through AI.
The research team used AI to analyze and evaluate voice signals and calculate them in just a few seconds. Afterwards, as a result of actually applying it to patients, he explained that it accurately distinguished healthy people from Parkinson’s patients.
First, we recorded the speech of a healthy person and a Parkinson’s patient using microphones in a soundproof booth. AI was then trained to evaluate and process it as a signal. The research team emphasized that this algorithm can be made into a mobile app in the future as there is no need to use a device with high specifications.
“It is not intended to replace the routine tests that people with Parkinson’s have,” said the research team, “but the method is designed to facilitate early diagnosis of the disease and track the effectiveness of treatment .”
“As a result of testing with healthy people and negative samples of Parkinson’s patients, this algorithm was more accurate than other diagnostic methods that have been released before,” he said. It’s very big,” he added.
The research team plans to increase the number of participants in the future and compare the algorithm with the method used to detect Parkinson’s early. In addition, we intend to ensure that it can work well not only in the laboratory, but also in the doctor’s office or at home.