New Machine Learning Method Identify Subtle Movement Deficits in Early Parkinson's Disease, Study Reveals

A new research published in the journal Parkinsonism and Related Disorders, suggests that the use of machine learning (ML) can detect movement deficits in early-stage Parkinson’s disease from the videos. This technique measures motor symptoms in the early stages of Parkinson's disease and could uncover signs of the condition and other movement disorders sooner, potentially leading to improved treatment outcomes.

A research team from the University of Florida and the Fixel Institute for Neurological Diseases has demonstrated that using video analysis can aid in identifying early signs of Parkinsonism in individuals. Their method involves comparing the movement of the left and right sides of a person's body. According to the researchers, this technique leverages the characteristic asymmetry of Parkinson’s disease, which typically begins with more pronounced symptoms on one side of the body compared to the other in the initial stages. This asymmetrical onset allows the video-based approach to detect subtle differences in movement that may indicate the early presence of the disease.

The researchers employed machine learning techniques to analyze videos of individuals executing basic hand and leg movements, which are routinely assessed by neurologists. By examining these videos, the team sought to identify subtle variations between healthy individuals and those in the early stages of Parkinson’s disease. Their innovative approach successfully distinguished between the two groups with an accuracy rate of 86%. This level of precision highlights the effectiveness of their method in detecting early Parkinson’s disease by analyzing nuanced differences in movement patterns.

This technology is user-friendly, easily scalable, and offers significant potential to enhance the monitoring and measurement of motor symptoms in the early stages of Parkinson’s disease.

Reference: Guarín, D. L., Wong, J. K., McFarland, N. R., Ramirez-Zamora, A., & Vaillancourt, D. E. (2024). What the trained eye cannot see: Quantitative kinematics and machine learning detect movement deficits in early-stage Parkinson’s disease from videos. Parkinsonism & Related Disorders, 107, 107104. https://doi.org/10.1016/j.parkreldis.2024.107104