AI-Based Quantification of Parkinson Disease Motor Characteristics from Video Data
Parkinson Disease (PD) is a progressive neurodegenerative disorder that affects movement control, leading to symptoms such as tremor, rigidity, and bradykinesia. Accurate and objective assessment of motor symptoms is crucial for monitoring disease progression and evaluating treatment efficacy. Currently, the Movement Disorder Society – Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) is the gold standard for clinical assessment of motor function in PD. This scale includes several standardized motor tasks, such as finger tapping, hand movements, pronation-supination, and toe tapping, which are evaluated and assessed by trained clinicians. While these assessments are widely used, they are inherently subjective and may suffer from inter-rater variability.
The goal of this project is to develop an automated and objective video-based method for analyzing UPDRS videos. Using pose estimation techniques, kinematic features such as movement amplitude, speed, and rhythm will be extracted.
The intern will work with a large-scale video dataset from the Personalized Parkinson Project (PPP), which includes recordings of 520 individuals with PD collected by the neurology department at Radboudumc. Tasks will involve data preprocessing, pose extraction, feature engineering and data analysis. This interdisciplinary project combines artificial intelligence with clinical neuroscience, contributing to the development of digital biomarkers that can support clinicians in treatment planning and long-term monitoring of PD.
References
[1] Güney, G., Jansen, T.S., Dill, S., Schulz, J.B., Dafotakis, M., Hoog Antink, C. and Braczynski, A.K., 2022. Video-based hand movement analysis of Parkinson patients before and after medication using high-frame-rate videos and MediaPipe. Sensors, 22(20), p.7992.
[2] Vignoud, G., Desjardins, C., Salardaine, Q., Mongin, M., Garcin, B., Venance, L. and Degos, B., 2022. Video-based automated assessment of movement parameters consistent with MDS-UPDRS III in Parkinson’s disease. Journal of Parkinson's Disease, 12(7), pp.2211-2222.
[3] Guo, R., Li, H., Zhang, C. and Qian, X., 2022. A tree-structure-guided graph convolutional network with contrastive learning for the assessment of parkinsonian hand movements. Medical Image Analysis, 81, p.102560.
Contact: Tahereh Zarrat Ehsan (Radboudumc); Luc Evers (Radboudumc); Twan van Laarhoven (RU).