Correlation of Clinical Scores with Remote Assessments of Accelerometry Output of a Low-Cost Quantitative Continuous Measurement of Movements
Description
This dataset accompanies the research poster titled “Correlation of clinical scores with remote assessments of accelerometry output of a low-cost quantitative continuous measurement of movements.” It focuses on the reliability and validity of using wearable accelerometers to remotely evaluate motor function in individuals with Parkinson’s disease (PD) and compares those results to standard clinical scores obtained through in-person assessment. The study collected accelerometry data from participants performing predefined motor tasks. These data were then processed to extract movement features such as amplitude, frequency, and signal stability. Clinical motor scores were concurrently collected using established neurological scales. The dataset includes both raw and summarized data from individuals with PD and age-matched typically developing (TD) controls. The primary analysis involved assessing the internal consistency and test-retest reliability of the accelerometer-derived features using Cronbach’s Alpha and Intraclass Correlation Coefficients (ICC). High internal consistency (e.g., Cronbach’s Alpha = 0.918 for PD group) and strong reliability in average measures (ICC = 0.918 for PD group) support the use of remote motor assessments as a valid tool for clinical or research purposes. This description style is in line with other published Mendeley Data entries such as: Yoon et al. (2021) – Wearable-based gait and balance features dataset for fall risk prediction in elderly (https://data.mendeley.com/datasets/dyhgxdkf5y/1) Giggins et al. (2022) – Accelerometer-derived physical activity and balance metrics in Parkinson’s disease (https://data.mendeley.com/datasets/xmzz5mfxv9/2) The dataset is intended for researchers interested in neurodegenerative disease monitoring, remote assessment technologies, movement disorders, and wearable sensor applications. It can be reused for meta-analyses, validation studies, and the development of digital health tools, particularly in resource-limited or telehealth contexts.
Files
Steps to reproduce
This section outlines how to replicate the core findings of the study using the provided dataset. It guides researchers or clinicians in verifying the statistical relationships between clinical scores and accelerometry-derived movement metrics in individuals with Parkinson’s disease (PD) and typically developing (TD) participants. No programming knowledge is required—standard statistical tools (e.g., Excel, SPSS, R, or Jamovi) can be used. 1. Access and Explore the Dataset Download the dataset files, which include: A spreadsheet (.xlsx or .csv) with participant IDs, group labels (PD or TD), clinical assessment scores, and extracted accelerometry features. Familiarize yourself with the dataset structure. Common columns include: Participant ID Group (PD or TD) Clinical Score(s) (e.g., UPDRS or similar) Accelerometry Metrics (e.g., mean amplitude, standard deviation, number of peaks) 2. Review Clinical Scores and Group Labels Identify which columns represent the clinical scores used in the study. Ensure participants are correctly grouped into PD and TD categories. Consider filtering or flagging incomplete or missing data points, if any. 3. Analyze Internal Consistency Use a statistical software package (e.g., SPSS or Jamovi) to calculate Cronbach’s Alpha for the accelerometry-based features. This test measures how consistently the features describe similar aspects of motor behavior. Perform the analysis separately for PD and TD groups. Expected result for PD group: Cronbach’s Alpha ≈ 0.918 Expected result for TD group: Cronbach’s Alpha ≈ 0.895 4. Assess Measurement Reliability Calculate the Intraclass Correlation Coefficient (ICC) to evaluate the agreement and reliability across repeated measures or averaged accelerometry metrics. Perform both: Single Measures ICC Average Measures ICC Compare the ICC values for PD and TD groups. Example findings: PD – Average Measures ICC: 0.918 (p = 0.000) TD – Average Measures ICC: 0.704 (p = 0.000) 5. Compare Group Differences Stratify the dataset by PD and TD groups. Use descriptive statistics (mean, SD) to explore differences in accelerometry metrics and clinical scores. Perform group comparison tests (e.g., t-test or ANOVA) to assess statistical significance between groups. 6. Validate Against Clinical Scores Evaluate how accelerometry-derived metrics relate to clinical scores. Use correlation analysis (e.g., Pearson or Spearman) to determine how well these remote measurements reflect clinical assessments. Interpret the strength and direction of correlations in both PD and TD groups. 7. Interpret and Compare Match your results with those provided in the dataset documentation and poster: Are the reliability coefficients consistent? Do the group differences align? Is the correlation with clinical scores strong in the PD group?
Institutions
- New York University
- Johns Hopkins University
- Misr University for Science and Technology
- University of Illinois at Chicago