We recently covered MIT’s Parkinson’s detecting AI model that offers diagnosis based on the patient’s breathing pattern. A recent research published in the IEEE Journal of Translational Engineering in Health and Medicine, IEEE Access, and Computers in Biology and Medicine demonstrates promising results about a novel screening test application that may well-assist in the timely detection of Parkinson’s disease and severe COVID-19, thereby improving disease monitoring. 

A group of engineers and neurologists at RMIT University have devised a voice screening app that helps to detect Parkinson’s and severe COVID-19 illness just by listening to a person’s voice recordings. The smartphone app, driven by artificial intelligence, analyzes a person’s speech and takes only about 10 seconds to determine whether they have Parkinson’s disease and should be examined by a neurologist. Presently, Parkinson’s disease is diagnosed through a neurological examination by a neurologist, which can take up to 90 minutes.

The researchers employed voice recordings of Parkinson’s patients and a healthy control group speaking three sounds – A, O, and M. It was observed that these sounds led to a more accurate diagnosis. The same test can be used for people with COVID-19 pulmonary disease symptoms and an alteration in voice caused by a pulmonary infection. However, due to the significant diversity in people’s voices, respiratory disorders are more challenging to spot in their early stages. 

The new app functions by comparing speech samples of persons with Parkinson’s disease and those without it, thereby yielding exact findings. The researchers believe that the app is quicker, more efficient, and more effective than any previous AI-based technology and could be ideal to use in a national screening program.