Photo credit: www.sciencedaily.com
Researchers at the University of Florida have made significant strides in the field of movement disorders through the development of an innovative open-source program, dubbed VisionMD. This artificial intelligence-driven tool is designed to analyze videos of patients suffering from Parkinson’s disease and similar conditions, facilitating more precise monitoring of subtle motor changes that can enhance patient management and pave the way for advancements in clinical research.
Diego Guarin, Ph.D., who occupies a position as an assistant professor in the College of Health and Human Performance, created this software in response to issues related to inconsistency and subjectivity often encountered in conventional clinical evaluations.
“Our research has consistently demonstrated that analyzing videos of patients engaged in activities like finger-tapping yields critical insights into disease progression and responses to treatments, such as medications or deep brain stimulation,” Guarin explained. “However, the reality is that most clinicians lack the time and resources to scrutinize these videos. Hence, we developed a user-friendly software solution that generates valuable results with minimal effort.”
Guarin is affiliated with the Fixel Institute for Neurological Disease at UF Health, where he collaborated extensively with neurologists and other clinician-researchers to perfect VisionMD.
The software is capable of analyzing standard video recordings—from those captured on smartphones to those streamed via Zoom—automatically deriving detailed motion metrics. A key feature of VisionMD is its operation on local machines, thus prioritizing data privacy and security.
“Because it’s not cloud-based, there’s no risk of data being transmitted outside the local network. You can even operate it offline,” Guarin noted. The tool is already in practice worldwide, with researchers across Germany, Spain, and Italy using it to scrutinize thousands of patient videos, exploring the potential of computer vision technology in improving care for movement disorders.
Florian Lange, a neurologist from University Hospital Würzburg, praised the program for its capacity to deliver consistent and objective measurements. Along with Martin Reich, a neuroimaging expert at the same institution, Lange has adapted VisionMD to assist in optimizing treatment plans for patients with tremors, particularly those who have received deep brain stimulation (DBS) implants.
“A significant hurdle in current medical practice is obtaining objective data, especially for conditions like Parkinson’s disease or tremors,” Lange remarked from his office in Germany. “If three clinicians were to assess the same video, we could end up with three different severity ratings. In contrast, the software provides precise and unbiased metrics.”
Through the analysis of patient videos captured under various stimulation settings, VisionMD can identify the DBS configuration that results in the most effective symptom relief.
“With countless programming possibilities, this tool streamlines the process of determining optimal settings with accuracy and speed,” Reich added.
As an open-source platform, VisionMD is freely accessible, allowing enhancements and customizations by the broader research community.
The development team is also actively working to broaden the program’s functionalities by incorporating additional motor assessment tasks commonly utilized in clinical environments.
Initial users of VisionMD have lauded its straightforwardness and accessibility, suggesting it holds the potential to significantly enhance both research and care in the domain of movement disorders.
“Processing each video takes only seconds,” Guarin concluded. “We believe that most clinicians will find it manageable, regardless of their level of technical expertise.”
Source
www.sciencedaily.com