Exploring Parkinson's Disease Datasets: Key Findings, Challenges, and Recommendations for Motor Symptom Analysis

Parkinson's Disease (PD) is a progressive neurodegenerative disorder characterized by motor symptoms such as bradykinesia, tremor, or freezing of gait (FoG). Over the past decade, the rise of low-cost sensing technology has facilitated data collection, leading to the creation of datasets capturing motor symptoms and enabling advancements through machine (ML) and deep (DL) learning techniques for early diagnosis, precise symptom monitoring, and personalized treatment strategies in PD management. However, limited patient accessibility and dataset availability continue to pose challenges for widespread implementation and cross-dataset studies. This paper surveys the 17 (seventeen) most widely used PD motor symptom analysis datasets, examining their features, modalities, and data sources while addressing the variability challenges across datasets.