Selfies may help identify Parkinson's disease
A new software program may help identify Parkinson’s disease, researchers have found.
As described in a paper published in Nature Digital Medicine, researchers at the University of Rochester crafted computerized vision software that can analyze videos of patients for subtle clues in their facial muscles and facial expressions normally undetectable to the eye.
The program could predict if a patient who takes a selfie would be more likely to develop PD. Existing technology involves wearable digital markers that measure body motion.
“Parkinson’s is the fastest growing neurological disorder,” Ehsan Hoque, PhD, associate professor of computer science at the university, said in a press release. “What if, with people’s permission, we could analyze those selfies and give them a referral in case they are showing early signs?”
Funding for the project arrived this month through the Gordon and Betty Moore Foundation via a $500,000 grant. However, certain ethical and technological questions need to be answered through continued research for the study to reach its full potential.
“The foundation wants us to validate the feedback that we would give people if they did, indeed, show early signs of Parkinson’s — especially if they are performing the test at home,” Hoque added.
“The challenge is not only validating the accuracy of our algorithms but also translating the raw machine-generated output in a language that is humane, assuring, understandable, and empowering to the patients.”
Through collaborations with a leading in-house PD expert, as well as the university’s PD research center, a test with five separate criteria was created for neurologists to administer to patients seeking telehealth consultations. According to Hoque, the test could provide a necessary breakthrough in the barriers to accurate and timely health care for underserved populations.
Patients involved in the study were asked to complete the following tasks: making the biggest smile they can three times and alternating it with a neutral expression three times; reading a complex written sentence out loud; touching their index finger to their thumb 10 times as quickly as possible; making as much of an ugly face as possible then alternating with a neutral expression three times; and raising both eyebrows as high as possible, then lowering them as far as they can, three times slowly.
While these tasks were being completed, the program was able to assess whether a patient displayed less control over their facial muscles, a signifier of PD.
Within a matter of minutes, the program can produce a quantifiable percentage of likelihood, based on each of the above-mentioned tests, that a patient demonstrates symptoms of PD or associated issues.
“One thing about Parkinson’s is that you don’t show all the symptoms all the time, and not every symptom is shown in every part of your body,” Rafayet Ali, lead author of the paper, said in the release. “For example, you may not have hand tremors, but you may show a significant level of deviation in your smile.”
Although progress has been significant, Hoque and colleagues need further guidance from the FDA before being granted permission to analyze patients’ pictures and proceed to the five-step neurological test.
Additionally, several other motion disorders aside from PD could be indicated through future research, not limited to ataxia, Huntington’s disease, progressive supranuclear palsy and multiple dystrophy.