Two researchers are using smartphone data to try to diagnose Parkinson's disease.
In a paper published Thursday, Patrick Schwab and Walter Karlen of the Institute for Robotics and Intelligent Systems in Switzerland proposed using smartphone data and machine learning to help diagnose Parkinson's disease. The idea is to use data collected by a smartphone and analyze it for signs of the disease, as well as its progression.
Parkinson's disease is a disorder of the nervous system that affects a person's ability to control his or her movements, according to the Mayo Clinic. The disease progresses over time and can affect a sufferer's ability to speak or walk. Currently, no cure exists for Parkinson's, though treatments are available to make the symptoms more manageable.
Parkinson's affects more than 6 million people worldwide, according to the research paper. A quarter of Parkinson's diagnoses are incorrect because other movement disorders may have similar symptoms, according to the report.
Schwab and Karlen used a smartphone app to collect data. Participants input demographic and medical information, and then were asked to perform four tests using their phones. The tests, which were conducted as many as three times a day, include walking, voice, tapping and memory.
The walking test gives instructions, such as walk 20 steps forward and turn around.
The voice test asks the user to say "aaaah" for up to 10 seconds.
The tapping test requires the user to put the phone on a flat surface and alternatively tap two buttons on the screen for 20 seconds.
The memory test asks the user to remember the pattern of flowers lighting up on the screen, and then touch the flowers in the same order.
Schwab and Karlen found that comparing multiple walking tests was the most informative method for diagnosing Parkinson's. Doctors could also access long-term observational data without requiring the patients to be present, according to the report.
The researchers didn't immediately respond to a request for comment.