[youtube https://www.youtube.com/watch?v=ZzDNRDypz1I?feature=oembed&w=100&h=100]
Progressive neurodegenerative brain disorder Parkinson’s disease (PD) can be tough to spot early on. A failure to do so can mean that patients miss out on valuable time during which they could get treatment or have the opportunity to make lifestyle changes. That’s something an EU-funded project called i-Prognosis wants to change. A four-year project led by Greece’s Aristotle University of Thessaloniki, it has launched a smartphone and smartwatch app which could help diagnose onset Parkinson’s much earlier than is currently the case.
“PD is a progressive and chronic neurological disease that often begins with mild symptoms that advance gradually over time,” Professor Leontios Hadjileontiadis, coordinator of the project, told Digital Trends. “Symptoms can be so subtle in the early stages that they go unnoticed, as there are no PD-related biomarkers, such as blood tests, and findings on routine magnetic resonance imaging and computed tomography scans are unremarkable — leaving the disease undiagnosed for years.”
It is here that i-Prognosis aims to help, thanks to the various sensors found in modern mobile devices and some amazing insights from machine learning.
“i-Prognosis proposes a radically novel approach to capture the risk of transition from healthy status towards PD by unobtrusive behavioral sensing and large scale collection of elderlies’ data, acquired from their natural use of smart devices,” Hadjileontiadis continued.
Once a user’s consent has been gathered, the app proceeds to analyze an extraordinary amount of data on their phone. This includes (but isn’t limited to) their voice characteristics while having conversations, the steadiness of their hands while holding their device, the way they enter keystrokes, and even their facial expressions on stored photos or the emotional content of text messages. In other words, it combs through every inch of data in a way that no doctor would be able to, and uses this to predict whether you might want to see a physician. This level of detail probably doesn’t make it an app for the more privacy conscious of people, but if it is able to successfully make you aware of onset Parkinson’s the trade-off may be worth it.
For now, the app is firmly in the testing phase. Hadjileontiadis said that the models will continued to be refined through February 2020. At present, it is available in the Google Play app store for Android users, although only in Germany, Greece, Portugal, and the United Kingdom. Hopefully, a U.S. version — and possible iOS support — could follow at a later date.
