New AI Tool Allows Remote Assessment of Parkinson’s Symptoms in Minutes
Researchers at the University of Rochester have developed an artificial intelligence (AI) tool that enables individuals with Parkinson’s disease to remotely assess the severity of their symptoms within minutes. This groundbreaking tool, described in a study published in npj Digital Medicine, requires users to tap their fingers ten times in front of a webcam, which helps evaluate their motor performance on a scale of 0-4.
Traditionally, doctors have patients perform simple motor tasks to assess movement disorders and determine their severity, using guidelines such as the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS). However, the new AI model developed by the University of Rochester provides a rapid assessment based on these guidelines. It generates computational metrics, including speed, amplitude, frequency, and period, which are interpretable, standardized, repeatable, and consistent with medical guidebooks. It further uses these attributes to classify the severity of tremors.
To validate the tool’s accuracy, the research team had 250 global participants with Parkinson’s disease perform the finger-tapping task. The AI system’s ratings were then compared with those of three neurologists and three primary care physicians. Although the expert neurologists performed slightly better than the AI model, the AI model outperformed the primary care physicians who had UPDRS certification.
One of the significant implications highlighted by Ehsan Hoque, an associate professor at Rochester’s Department of Computer Science and co-director of the Rochester Human-Computer Interaction Laboratory, is the potential to improve healthcare accessibility and equity for patients who face challenges in accessing neurologists, securing appointments, and traveling to hospitals.
The lead researcher of the study, Md. Saiful Islam, a Google PhD fellow and a graduate student in computer science advised by Hoque, emphasized that this AI method can be applied to other motor tasks, opening the door to evaluating different types of movement disorders such as ataxia and Huntington’s disease. Although the new Parkinson’s assessment tool is available online, the researchers caution that it should not be considered as a definitive measure of the presence or severity of the disease without a physician’s input, as it is still an emerging technology.
The findings of this study mark a significant step towards integrating AI into healthcare, enabling remote assessments and expanding access to necessary medical evaluations. By leveraging technology, individuals with Parkinson’s disease can conveniently and efficiently monitor their symptoms, eliminating the need for frequent visits to a clinic. As AI continues to advance, it holds great promise for improving health equity and accessibility across various medical conditions.
