Computer vision uncovers three fundamental dimensions of levodopa-responsive motor improvement in Parkinson’s disease
We developed VisionMD, an AI computer vision platform, analyzing over 1200 clinical videos of Parkinson's patients' hand movements across 13 years. This large-scale, markerless analysis identified three kinematic domains (speed, consistency, timing/scale) reliably improved by levodopa. Our method offers objective, quantitative motor assessment, reducing subjectivity and enhancing reproducibility compared to traditional scales.
Authors
- Florian Lange (ORCID: https://orcid.org/0000-0001-8497-7780)
- Diego L. Guarín (ORCID: https://orcid.org/0000-0001-6077-7063)
- Martin M. Reich (ORCID: https://orcid.org/0000-0001-9115-5136)
- Jens Volkmann (ORCID: https://orcid.org/0000-0002-9570-593X)
- Robert L. Peach (ORCID: https://orcid.org/0000-0002-8738-5825)
- Thorsten Odorfer (ORCID: https://orcid.org/0000-0001-9042-7689)
- Joshua K. Wong (ORCID: https://orcid.org/0000-0002-4632-3417)
- Gabriela Acevedo
- Esther Ademola
- Dalia Mahdy
Institutions
- University of Würzburg (DE)
- University of Florida (US)
Publication Details
- Journal
- npj Parkinson s Disease
- Published
- 2025-05-28
- DOI
- https://doi.org/10.1038/s41531-025-00999-w
- Citations
- 7
- Primary Topic
- Parkinson's Disease Mechanisms and Treatments
- Type
- article
- Field-Weighted Citation Impact
- 9.72
Funders
- Deutsche Forschungsgemeinschaft
- Interdisziplinäres Zentrum für Klinische Forschung, Universitätsklinikum Würzburg