In a significant breakthrough for Parkinson’s disease assessment, researchers at the University of Rochester have developed a remote AI tool that can assess the severity of Parkinson’s disease with just ten taps of the finger.1
Image Credit: PopTika/Shutterstock.com
Traditionally, doctors have relied on simple motor tasks to assess movement disorders and rate the severity using guidelines such as the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS).
However, this new research, published in the journal npj Digital Medicine, could potentially allow individuals suffering from Parkinson’s and other motor neurological conditions to be “remotely” and “objectively” monitored.1 The relevancy of such a breakthrough could support those dealing with neurological disorders, which are the leading source of disability globally, with Parkinson’s being the fastest-growing neurological disorder in the world.2
Advanced Assessment for Improving Healthcare
The AI-powered tool operates by monitoring a patient’s motor performance through video captured by a webcam, during which the patient performs a relatively simple hand gesture. The AI utilizes object tracking features that identify specific points on the patient’s hand. These points are then swiftly cross-referenced with the MDS-UPDRS guidelines, assessing various criteria such as speed, amplitude, and frequency.
Once the assessment is complete, the AI model then scores the patient’s motor performance between 0-4, which offers rapid insights into the severity of symptoms such as tremors.1 With easy access to an advanced AI assessment platform that provides almost instant results, Parkinson’s patients would have improved access and better control of their healthcare.
These findings could have huge implications for patients who have difficulty gaining access to neurologists, getting appointments, and traveling to the hospital… It’s an example of how AI is being gradually introduced into health care to serve people outside of the clinic and improve health equity and access.3
Ehsan Hoque, Associate Professor, Department of Computer Science and Co-Director of Human-Computer Interaction Laboratory, University of Rochester
Results and Insights
Recognizing the transformative potential of AI in the healthcare field, ensuring access to computer learning technology hinges on configuring systems with the right data and precise outcomes. To this end, the Rochester team enlisted 250 volunteers with Parkinson’s disease to participate in the finger-tapping exercise, which involved both hands and was captured through motion tracking using a web-based application and a computer camera.
The team gathered 489 videos for assessment, excluding those that did not meet data quality standards. The results were then evaluated using the AI-powered tool and compared to assessments conducted on the same footage by neurologists and primary care physicians.
While the neurologists achieved the highest level of accuracy, slightly outperforming the algorithm, the AI-powered tool provided more precise and comprehensive insights compared to the primary care physicians with UPDRS certification.
A notable feature of the AI-powered assessment tool is that, unlike human evaluators who can only calculate the number of taps performed within a specific time frame, the computer model has the capability to continuously and accurately track the finger-tapping speed throughout the task.
Several statistical measures of continuous speed are significantly correlated with PD severity. These granular computations are only attainable using automated video analysis, which, to our knowledge, was missing in prior literature.
Saiful Islam, Rochester Ph.D.
Future Application Potential
In this study, the team focused exclusively on evaluating motor-related symptoms associated with Parkinson’s disease. However, they also recognize the platform’s potential for assessing movement and motor responses in patients with various neurological conditions, including Huntington’s disease and conditions like ataxia.
The system is currently accessible online but still requires further refinement, such as reducing its susceptibility to interference from systematic noise, improving its current low-quality cameras, and other intermediary factors.
As the technology is still in its developmental phase, it is important to emphasize that this system is no replacement for an expert neurologist, and proper consultation from a medical professional is still strongly advised to ensure conclusive results when assessing the presence and severity of Parkinson’s.
Improving Healthcare Together
According to the WEF, bridging healthcare gaps around the world can be brought about by improving AI predictive models. The mentioned improvements can also assist in uncovering and identifying patterns and trends in illnesses, potentially facilitating advancements in disease treatment and prevention and providing personalized treatment plans for patients.4
Overall, the Rochester team has devised a potentially game-changing AI-powered tool that could grant patients better access to key care and assessment appointments while streamlining the evaluation criteria for healthcare professionals.
References and Further Reading
- Islam, M.S. et al. (2023) ‘Using AI to measure Parkinson’s disease severity at home’, npj Digital Medicine, 6(1). Available at: https://www.nature.com/articles/s41746-023-00905-9.
- Dorsey, E.R. et al. (2018) ‘The Emerging Evidence of the Parkinson Pandemic’, Journal of Parkinson’s Disease, 8(s1). Available at: https://pubmed.ncbi.nlm.nih.gov/30584159/.
- Auburn, L. (2023) Online AI-based test for Parkinson’s disease severity shows promising results, News Center. Available at: https://www.rochester.edu/newscenter/ai-test-for-parkinsons-disease-severity-566772/
- Yoon, S. and Amadiegwu, A. (2023) AI can make healthcare more accurate, accessible, and sustainable, World Economic Forum. Available at: https://www.weforum.org/agenda/2023/06/emerging-tech-like-ai-are-poised-to-make-healthcare-more-accurate-accessible-and-sustainable/#:~:text=AI%20algorithms%20can%20catalyse%20the,privacy%20and%20security%20are%20essential.
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.