Innovation

Solutions: What Do Your Keystrokes Say About Parkinson’s?

Analyzing typing patterns may help diagnose the disease


By Christine Hall | December 5, 2017

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.

Part of this work was possible through collaborations with faculty, clinicians and researchers at the Massachusetts Institute of Technology, HM CINAC and the Madrid-MIT M+Vision Consortium. Namely: Prof. Martha Gray, Prof. Jose Obeso, Teresa Arroyo Gallego, Dr. Alvaro Sanchez Ferro, Ian Butterworth, Michele Matarazzo, Paloma Montero and Veronica Puertas-Martin. The work was supported, in part, by the Michael J. Fox Foundation for Parkinson’s Research.

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.




Social Posts

profile_image

MD Anderson Cancer Center

@MDAndersonNews

The three focus areas of our Prostate #CancerMoonshot: - overcoming treatment resistance - exploring #immunotherapy - understanding the tumor environment Learn more: https://t.co/0ZQCm8tMdR #PCSM #endcancer

8 hours ago
profile_image

Rice University

@RiceUniversity

The @GlasscockSchool's fall course catalog is chock-full of offerings for ever-curious learners, including classes in the humanities and sciences, foreign languages and personal and professional development. https://t.co/A793BrcLfz https://t.co/8a1K8LbHE2

9 hours ago
profile_image

Veterans Affairs

@DeptVetAffairs

Last Shasta County Navy Veteran who survived Pearl Harbor attack laid to rest https://t.co/lc5ODzbk9l via @JimSchultz_RS

9 hours ago
profile_image

BCMHouston

@bcmhouston

One internal medicine resident looks back at a time where a patient helped him realize the balance of overly identifying with patients and complete detachment. https://t.co/7xsFWEhVc4

9 hours ago
profile_image

Veterans Affairs

@DeptVetAffairs

Santa Clarita Air Force Veteran Launches Campaign For Web Series To Depict Modern Soldier Life https://t.co/Td0ds3jxue via @KHTSRADIO

10 hours ago
profile_image

Veterans Affairs

@DeptVetAffairs

Today’s #VeteranOfTheDay is @USArmy Veteran Stanley Nelson. https://t.co/SHml8yEysj

11 hours ago
profile_image

UTHealth

@UTHealth

Thank you @MBThewoodlands for supporting student scholarships at @MDA_UTHGrad! It also is hosting our next House Calls web chat on sports medicine this Thursday, July 19, at 6:30 p.m. CST at https://t.co/N5UU1Jx4pq. Submit a question for our experts by using #UTHealthHouseCalls.

11 hours ago
profile_image

UTHealth

@UTHealth

RT @abc13houston: The most common sports injuries are strains and sprains, but do you know when you might need to see a doctor? The experts…

11 hours ago
profile_image

BCMHouston

@bcmhouston

RT @BCMHouston_News: Tune in to @FOX26Houston tomorrow in the 8 am hour to hear @bcmhouston's Dr. El-Serag discuss the recent CDC report on…

12 hours ago
profile_image

MD Anderson Cancer Center

@MDAndersonNews

#Cancerpain can be related to the disease and treatment. Hear from Dr. Salahadin Abdi on how patients can find relief. #endcancer https://t.co/snFVjDUjNe

12 hours ago
profile_image

TexasHeartInstitute

@Texas_Heart

RT @HealthyWomen: 10 Sneaky Ways to Get Fruits and Veggies in Your #Diet: https://t.co/wJ9x39147k #health https://t.co/EWktCmct2J

12 hours ago
profile_image

BCMHouston

@bcmhouston

RT @BCMHouston_News: Certain nail products can cause allergic reactions or irritations. @bcmhouston's Dr. Katta shares what to look out for…

13 hours ago
profile_image

MD Anderson Cancer Center

@MDAndersonNews

@ParrotsMatter @TargetedOnc @michaelwangmd Hi, yes we are. Here's a list of current clinical trials: https://t.co/JgHpaIKBbX. Best wishes to you.

13 hours ago
profile_image

Rice University

@RiceUniversity

RT @RiceAthletics: To celebrate #WorldEmojiDay, check out our favorite mascot @SammyTheOwl! 🦉👐 https://t.co/ebIKcsu2Z9

13 hours ago
profile_image

UTHealth

@UTHealth

RT @abc13houston: A workout injury can happen to anyone, no matter your experience or fitness level. Learn about treatment options and how…

13 hours ago