Parkinson's is a progressive disease with no known cure that affects as many as six million people worldwide. Diagnosis and tracking currently depends on extensive testing and a ranking of symptoms. All that could soon change, as a new voice based system using a speech-processing algorithmcould diagnose the disease with a three-minute phone call.
Parkinson's Disease often manifests itself early on with subtle changes in person's voice. While the telltale quavers, breathiness or hoarseness may not be recognized by the human ear, Max Little of the Massachusetts Institute of Technology Media Lab has tackled this by developing the non-invasive speech-processing algorithm that analyzes vowel sounds.
Little and his team processed 263 recordings of 43 people as they sustained six or seven vowel "ah" sounds and the algorithm was trained to recognize 10 impairments or variations in the recordings. After the training, the program was then able to recognize the speech markers for Parkinson's and diagnosed them 99 percent of the time under lab conditions.
The next step comes in testing the algorithm in the real world.
After explaining the breakthrough at the TEDglobal conference in Edinburgh, Scotland at the end of June, Little appealed to the global community to help test the application in the real world. Visitors to the Parkinson's Voice Initiative website can find out more information and obtain local phone numbers to help test its efficacy under real conditions and in multiple languages.
An easy, non-invasive diagnostic testing system for Parkinson's could be a game changer. It's estimated that a fifth of Parkinson's cases are never diagnosed due to the extensive testing which can be time consuming and costly. Further, delays in diagnosis can cut down on a patient's quality of life when they could potentially manage symptoms via drugs and surgical intervention.
Parkinson's is just one of the many diseases and conditions that could be diagnosed via phone — an area of research called telemedicine. While Little's project is focused only Parkinson's, his comments at TEDglobal could very well apply to the entire field of study:
"This kind of non-invasive technology, which can be seamlessly integrated into people's lives, could give you data on their social life, daily patterns, and track them over time," Little says. "We end up with a giant database with far fewer risk factors, which will give researchers a way to streamline the hunt."
The video below provides additional information on Little's project as does the Parkinson's Voice Initiative website.
Looking for more of the latest on how technology impacts health and well-being? Check out our Healthy Tech series.
Via New Scientist