AI detects subtle movement changes in video tapping fingers and reveals hidden signs of Parkinson's disease

Early detection of even slight motor changes is important to slow the progression of Parkinson's disease. However, these subtle signs are often unaware. Currently, UF researcher Dr. Diego L. Guarín uses AI to find these subtle changes from video recordings before clinical symptoms are revealed to the clinician's eyes.

Guarín, assistant professor at the Department of Applied Physiology at UF Health & Human Performance, recently published the results of his research NPJ Parkinson's Disease.

“Video analysis allows us to see changes in movement that the clinician cannot see,” Guarine said. “Identifying these changes in movement early is important for disease management.”

In his study, Guarín analyzed videos of finger tapping movements from 66 participants, including healthy individuals. People with idiopathic REM sleep behavior disorders or IRBD. And people with early Parkinson's disease.

Idiopathic REM sleep behavior disorders include people who are acting out their dreams, such as talking, moving, or assaulting while sleeping. Over 80% of people with IRBD develop Parkinson or related brain damage, which becomes an important subset of the population to study early changes in motor function.

Importantly, trial participants were required to show no brain condition with visible signs of parkinsonism, i.e. slow motion in the video tapping fingers.

“Expert clinicians watched the video and showed that those participants were healthy,” Guarine said. “We received analysis that made everyone look healthy to an outside observer.”

Video recordings were analyzed using VisionMD, an open source machine learning software developed by the Guarín team.

“When I process this video of a healthy looking person with VisionMD, I say, “No, this person is moving more slowly than you would expect from a healthy person.” So there are movement changes in videos that cannot be detected with the naked eye,” Guarín said. “Our video analysis techniques are so sensitive that we can identify things that clinicians can't.”

This study found that even when clinicians thought the finger tapping test looked normal, video analysis using AI showed that people with Parkinson's disease had smaller and slower movements than other groups, indicating the importance of this approach.

Furthermore, AI and video analyses can detect sequencing effects in IRBD and Parkinson's disease patients. The sequence effect is a gradual reduction in movement amplitude and/or velocity during repeated movements, such as tapping a finger. The origin and mechanism of this motor label is not well understood, but its presence in both IRBD and Parkinson's disease may indicate that the sequence effect is an early indicator of brain damage.

“Through standard video recordings, these simple and effective screenings can open the door to faster brain diagnosis and help people at higher risk of disease progression,” Guarín said.