Researchers at RMIT University in Melbourne, Australia, have developed an app that is intended to provide early diagnosis for Parkinson’s disease and severe COVID-19. The artificial intelligence-powered technology works by analyzing voice recordings, having previously been trained to recognize the vocal hallmarks of these diseases by listening to recordings of patients. The app takes just ten seconds to assess a voice sample and provide a recommendation that someone should seek further treatment. The technology could be useful in large community-wide screening programs given its convenience and speed.
What can our voice reveal about us? Well, perhaps more than you may have thought. The researchers behind this latest technology aim to diagnose Parkinson’s disease and identify patients at risk of progressing to severe COVID-19 simply by listening to their voices. The app relies on AI to identify the hallmarks of these disease while listening to the spoken word.
At present, Parkinson’s is typically diagnosed following a lengthy appointment with a neurologist. However, someone may not be referred to such a specialist until their symptoms have advanced to the point that they are obvious to anyone. Discovering and treating the disease early may help to improve patient outcomes, but the signs of early Parkinson’s are subtle.
There are some changes that impact a patient’s voice, including tremor, rigidity, and slowness, but these are difficult to spot, especially given the natural variation in people’s voices. Similarly, the early stages of progression to a more serious form of respiratory disease can also provide vocal hallmarks, but inter-individual differences in our natural speaking voices can make them difficult to identify.
To address this, these researchers identified a method to help standardize the voice samples provided to the software. “As part of our research, we used voice recordings of people with Parkinson’s and a controlled group of so-called healthy people saying three sounds — A, O and M — which is similar to the Hindu meditation chant,” said Dinesh Kumar, a researcher involved in the study. “These sounds result in a more accurate detection of the disease.”
The Australian team hopes that their advancement could form a useful tool in community-wide screening programs for Parkinson’s. “We are also keen to test the efficacy of this technology for other diseases, such as other neurological conditions and sleep disorders,” said Kumar. “We are looking for a commercial partner and clinical partner ahead of a clinical trial planned for next year.”
Study in journal Computers in Biology and Medicine: Convolutional neural network ensemble for Parkinson’s disease detection from voice recordings
Via: RMIT University
