For millions of people suffering from Parkinson’s, monitoring the progression of the disease can be challenging and time-consuming. Now, researchers at the University of Florida have developed a new video processing system that could revolutionize this process and make it easier to monitor the disease from home with unprecedented precision.

The research published in the journal IEEE Transactions on Neural Networksdescribes a video processing system developed by Diego Guarin, an assistant professor in the University of Florida’s College of Health and Human Performance. This system uses machine learning to analyze videos of patients performing a simple hand movement test, revealing tiny, often imperceptible changes in motor function that could indicate the progression of Parkinson’s disease.

“The beauty of this technology,” Guarin said, “is that a patient can record themselves performing the test and the software analyzes the data and tells the doctor how the patient is moving so the doctor can make decisions.”

Parkinson’s is a complex neurological condition that gradually affects a person’s ability to control their movements. There is currently no cure, and treatments focus on controlling symptoms rather than stopping the disease from progressing. One of the biggest challenges in treating Parkinson’s disease is closely monitoring its progression over time, particularly in the early stages when changes can be subtle and easily missed during standard clinical examinations.

Traditional methods of assessing Parkinson’s disease rely heavily on the Movement Disorder Society’s Unified Parkinson’s Disease Rating Scale, a system that, while widely used, has its limitations. This scale is based on a doctor observing a patient’s movements and scoring them on a 5-point scale that can be subjective and lacks the fine granularity needed to detect small but significant changes. In addition, this method requires patients to attend a clinic, which may not always be feasible, especially for patients with limited mobility.

Recognizing these limitations, Guarin and his team sought to develop a more objective and sensitive method for monitoring motor symptoms in Parkinson’s patients. Their goal was to create a tool that could be easily used at home and provide continuous monitoring and more precise information about the disease progression.

To test their new system, the researchers analyzed video data from 66 Parkinson’s patients and 24 healthy individuals. All participants performed a standardized finger tapping test, which involves tapping the thumb and index finger together 10 times quickly. This test is often used to detect bradykinesia, a slowing of movement that is a hallmark of Parkinson’s disease.

The videos were recorded under controlled conditions at the University of Florida Health Facility using a standard camera. Each participant sat in front of a camera while a trained clinician guided them through the finger-tapping test.

The videos were then processed using a custom machine learning pipeline developed by Guarin’s team. This pipeline uses Google’s MediaPipe, software that can track hand movements by identifying 21 key points on each hand. From these points, the system calculates various movement metrics, such as the speed and amplitude of finger taps, as well as more complex metrics such as movement variability and the time it takes to complete each tap cycle.

The study focused on comparing three different machine learning approaches to predict Parkinson’s disease severity based on these movement features extracted from videos. These approaches included a traditional multi-class classification model, an ordinal binary classification model, and a novel graded binary classification model. The latter was specifically designed by the researchers to account for the fact that different movement features may be more or less important at different stages of the disease.

The results of the study were promising. The new graded binary classification model outperformed the other methods, achieving 85% accuracy in distinguishing between healthy individuals and Parkinson’s patients and 86% overall accuracy in classifying disease severity. This represents a significant improvement over conventional methods and underscores the potential of this approach for clinical use.

“We found that by using a camera and a computer, we can observe the same features that doctors are trying to identify,” Guarin said. “Using AI, the same examination becomes easier and less time-consuming for everyone involved.”

One of the study’s key findings was that certain movement characteristics were more predictive of disease severity at different stages. For example, in the early stages of the disease, characteristics such as the speed and amplitude of finger movements were most predictive of severity. However, as the disease progressed, measures of movement variability became more important. This finding suggests that a one-size-fits-all approach to assessing Parkinson’s disease may not be effective and that different aspects of motor function should be emphasized depending on the stage of the disease.

“We found that in Parkinson’s disease, the opening movement is delayed compared to the same movement in healthy individuals,” Guarin said. “This is new information that is difficult to measure without video and computers. It shows us that technology can help better characterize the effects of Parkinson’s disease on movement and provide new markers to assess the effectiveness of therapies.”

The study also found that the new system can detect very subtle movement changes that may not be noticed by doctors using the traditional 5-point rating scale. This could be particularly useful in detecting the early signs of Parkinson’s, potentially allowing earlier intervention and better treatment of the disease.

Although the results are encouraging, the study has several limitations that need to be addressed in future research. A major limitation is that the video recordings were all made under controlled conditions with a doctor present to guide participants. This may not reflect the real-world conditions in which patients would record themselves at home without professional guidance. The researchers plan to address this by testing their system in more natural settings, where differences in camera angle, lighting, and patient positioning could affect the accuracy of the motion analysis.

Michael S. Okun, director of the Norman Fixel Institute and medical advisor to the Parkinson’s Foundation, called the automated video-based assessments a potential “game changer” for clinical trials and healthcare.

“The finger tap test is one of the most important elements for diagnosing and measuring disease progression in Parkinson’s,” Okun said. “Today you need an expert to interpret the results, but what is revolutionary is the way Diego and three Parkinson’s neurologists at the Fixel Institute were able to objectify disease progression using AI.”

Authors of the study “Characterizing Disease Progression in Parkinson’s Disease from Videos of the Finger Tapping Test” are Diego L. Guarín, Joshua K. Wong, Nikolaus R. McFarland and Adolfo Ramirez-Zamora.