A team from MIT’s world-famous Media Lab is now using a computer system to diagnose Parkinson’s disease over the phone.
A recent New Scientist article discusses how they used vocal markers of neurological damage that Parkinson’s causes in order to determine the likelihood that someone has the disease. This approach was facilitated by the fact that Parkinson’s impacts an individual’s control of their voice, introducing a “quaver in the voice, softer speech and breathiness or hoarseness, though they can be subtle at first.”
The MIT team assessed hundreds of recordings from people with Parkinson’s and employed machine learning techniques in order to “train” the computer system to recognize Parkinson’s disease. Max Little, a member of that team, called for volunteers to aid the team by calling in to a hotline and lending their voices in order to give the team more test data to work with. The team’s goal is to reach 10,000 recordings, and, as of this point, they are still far from reaching that goal.
While this approach might not work in all cases, it does showcase the fact that computer systems are now able, and have been for some time, to use reasoning similar to what a human in order to troubleshoot a situation or process.
Ubiquitous access to the Internet may further aid the process, since efforts such as Google’s cloud supercomputing initiative mean that no longer does the computer system recording the data need the horsepower necessary to analyze all of that data. It is easy to imagine a world where data on physical symptoms, mechanical problems, or a host of other phenomena are collected en masse in order to make for much more robust statistical sampling and, thus, a much more confident understanding of the world around us.
Perhaps, in the future, our mobile devices will be able not only to track our vital signs, but also to employ machine learning and other artificial intelligence techniques to diagnose any problems that might be indicated by those signs. Imagine a day where the doctor can’t enlist the aid of your mobile device in order to track heart rate, blood pressure, or even EKG for high-risk patients and send that information to your physician or, if necessary, a hospital wirelessly.