A young graduate in IT engineering has created an online site to detect early signs of Parkinson’s disease from voice recordings.
Says Diogo Braga, working out of Porto’s Superior Institute of Engineering, his hope is the software will help patients discover the illness without delays that can see it worsen through lack of medical treatment, and that correspondingly the physical and emotional tolls of Parkinson’s can be reduced.
The technology is still being perfected, but it involves users downloading onto the website a ‘wav’ or ‘mp3’ audio file of “around five seconds”.
The results are given “immediately”, explains TVI – thus people will know straight away whether they need to seek medical help, “or repeat the test another day”.
Indications of the slowly crippling motor neuron disease are detected through a system called ‘machine learning’ and based (so far) on the data of 24 patients with Parkinson’s (at different stages), and 30 ‘healthy people’ (or people without Parkinson’s).
Say reports, the success rate so far has been 92.38%.
The website is not yet ‘online’ as Braga and his backers want first to get the new tool publicised through scientific journals to give it “a greater degree of credibility”.
Long-term plans include adapting the software to be able to detect other neuro degenerative diseases, like Alzheimer’s and Huntington’s.
