Diagnosing Parkinson’s disease can be challenging, especially in its early stages. Studies show that doctors get it right only 55% to 78% of the time in the first five years.
This is because Parkinson’s shares symptoms with other movement disorders, making it hard to tell them apart.
Now, researchers at the University of Florida and the UF Health Norman Fixel Institute for Neurological Diseases have developed a powerful AI-driven tool that can diagnose Parkinson’s and similar conditions with over 96% accuracy.
Their findings were recently published in JAMA Neurology.
How Does the AI Software Work?
The new software, called Automated Imaging Differentiation for Parkinsonism (AIDP), uses MRI scans and machine learning to detect signs of neurodegeneration in the brain. It relies on diffusion-weighted MRI, a type of scan that tracks how water molecules move in brain tissue. By analyzing these patterns, the AI can identify different types of Parkinson’s and related disorders.
“This software works with different MRI machines, regardless of the manufacturer,” said David Vaillancourt, Ph.D., a professor at UF. “That’s a major advantage, as MRI companies typically use different software that doesn’t communicate with each other.”
A Big Step Toward Better Diagnoses
Doctors specializing in movement disorders often struggle to diagnose Parkinson’s accurately. Up to 50% of patients may receive an incorrect diagnosis. This AI tool can help doctors make more precise decisions, reducing the risk of misdiagnosis and ensuring patients receive the right treatment sooner.
While AI won’t replace doctors, it serves as a powerful tool to assist them. “Even the most experienced specialists can benefit from this technology,” Vaillancourt said.
The study was conducted at 21 medical centers—19 in the U.S. and two in Canada—showing that the technology works across different locations and medical settings.
What’s Next?
The next step for the research team is obtaining approval from the U.S. Food and Drug Administration (FDA) so the software can be used in hospitals and clinics nationwide.
“This is a great example of how technology and AI can improve medical care,” said Dr. Michael Okun, a leading expert on Parkinson’s disease. “It will allow us to provide better treatment for patients and improve outcomes for people living with Parkinson’s.”
The researchers behind this breakthrough, including Vaillancourt and Angelos Barmpoutis, Ph.D., are working through their company, Neuropacs, to bring the software to doctors and clinical trials.
With this innovation, diagnosing Parkinson’s could become faster and more accurate, helping countless patients receive the right care at the right time.
