An artificial intelligence has been developed to identify people who suffer from Parkinson’s disease by simply analyzing their voice.
- According to INSERM, more than 167,000 people are affected by Parkinson’s disease. There are 25,000 new diagnoses per year.
- Parkinson’s disease is caused by a progressive degeneration of dopamine neurons in the brain. Dopamine is a neurotransmitter that plays an essential role in the control of many functions including movement.
- Symptoms of the disease begin to appear when 50-70% of dopamine neurons are destroyed.
Hand tremors, muscle stiffness, balance problems… The motor symptoms of Parkinson’s disease are the best known. However, they are not necessarily the first to appear. Changes in the way of speaking come before.
This is why researchers from the Kaunas University of Technology (KTU) and the Lithuanian University of Health Sciences (LSMU) wanted to develop an artificial intelligence capable of detecting neurodegenerative pathology based on an analysis of the voice.
Parkinson’s: an artificial intelligence that listens to patients
Patients who are in the early stages of Parkinson’s disease may begin to speak in a calmer, more monotonous, less expressive or more fragmented way. Then as the disease progresses, hoarseness, stuttering or even difficult pronunciation of words are likely to appear. However, these changes are sometimes difficult to notice by ear.
Professor Virgilijus Ulozas came up with the idea of using artificial intelligence (AI) to analyze and evaluate patients’ voice signals. The system was trained to discover abnormalities in the voice of individuals affected by the pathology. The algorithm succeeded in distinguishing between oral recordings of patients and those of healthy volunteers. The results are “satisfactory from a medical point of view, confirming the possibility of generalizing the proposed model”indicate the authors in their article published in Applied Sciences.
A possible application to analyze the voice
For the researchers, their artificial intelligence offers a simple and inexpensive approach to the assessment and screening of Parkinson’s disease. They also point out that “the algorithm does not require powerful hardware and could be transferred to a mobile application in the future”.
“We are not creating a substitute for routine patient examination – our method is designed to aid in the early diagnosis of disease and to monitor the effectiveness of treatment”says Professor Rytis Maskeliūnas of Kaunas University of Technology (KTU).
“Our results, which have already been published, have a very high scientific potential. Of course, there is still a long and difficult way to go before they can be applied in daily clinical practice”, he adds. The project should, for example, be extended to other languages. Indeed, the tests were only carried out with Lithuanian and Italian.