Analyzing typing patterns may help diagnose the disease
By the time Parkinson’s disease is diagnosed, the damage is likely already done.
But what if there was a way to check for early signs of this progressive disorder of the nervous system? What if a smartphone app or computer keyboard could pick up Parkinson’s just from the way someone swiped or typed? What if it could determine how well people with Parkinson’s are responding to their medications?
Using a combination of digital health applications and artificial intelligence (AI), Luca Giancardo, Ph.D., and other scientists in the Center for Precision Health at The University of Texas Health Science Center at Houston (UTHealth) have created a program that can isolate differences in the typing signatures of people with and without Parkinson’s disease.
As the population ages, the number of people diagnosed with the disorder is expected to rise. Warning signs of Parkinson’s include rigidity or resting tremors, but with earlier intervention, doctors may be able to slow these symptoms.
Yet Giancardo, an assistant professor at UTHealth, said it’s not practical to test everyone for the onset of Parkinson’s, even if there was an easy way to do it. Rather, it is better to monitor an activity people are already doing regularly to see if there are any clues that could lead to a possible diagnosis of the disorder.
With that in mind, he developed a program that analyzes typing signatures—the keystroke patterns and quirks of individual users. The program can be downloaded onto a computer or smartphone and runs in the background while users go about their day, using their computers and smartphones as they normally would.
Giancardo calls this type of monitoring “passive monitoring,” to denote a test that doesn’t require any kind of special activity and can be done outside of the clinic. The goal, he said, is to find hidden patterns in the movements.
The average person takes 100 milliseconds to press and release a key, he said. Whenever a person types on a smartphone or desktop computer, the pressure and speed of the movements will generate a score that would quantify the patterns in the typing signatures.
“The measurements include the time from when you press a key until the time you release it,” he added.
At the moment, Giancardo is testing the program to see how people with Parkinson’s disease are responding to their medications. Although the program is still in the research phase, results suggest that changes in the typing signature could indicate that a medication is not working.
“We’ve learned from some animal studies that if we could diagnose people seven to 15 years earlier, we could get them using drugs designed to slow the neuron degeneration,” Giancardo said.
Giancardo hopes one day to extend these AI techniques to other conditions, including Alzheimer’s, and to integrate them with imaging and other tools.
Mya Schiess, M.D., has been working with Giancardo for the past year and said this technology could be used to measure the mobility of individuals afflicted by Parkinson’s.
“We could profile their mobility and, once we get their measurement, show change over time or rate a disease that is progressive,” said Schiess, vice chair of neurology and the Adriana Blood Endowed Chair in Neurology at McGovern Medical School at UTHealth, and a member of the medical staff at Memorial Hermann-Texas Medical Center.
She is also working to get patients who have had deep brain stimulation to treat their Parkinson’s involved with Giancardo’s program.
A typing profile could show how a person might benefit from deep brain stimulation, which implants a stimulator—similar to a pacemaker—that sends electrical impulses to electrodes implanted in the brain. After this treatment, she said, researchers would be able to show improvement in some tasks and neuroplasticity.
“This is on the verge of becoming a very important investigative tool and biomarker,” Schiess said.
No comments:
Post a Comment