Markerless Gait (OpenSim .mot) Dataset for Pediatric CP and Controls

## 📘 Dataset Overview This dataset contains fully anonymized, time-normalized gait kinematics from a study of pediatric cerebral palsy (CP) and age-matched control participants. All motion files are formatted for direct use in [OpenSim](https://opensim.stanford.edu/) and compatible with standard biomechanical analyses. Each file is derived from markerless motion capture data (OpenCap) and preprocessed through inverse kinematics using OpenSim 4.5. The dataset includes both the original `.mot` files and processed analysis results (e.g., PCA, clustering, SPM comparison). --- ## 📂 File Structure - `data/cp/P001/V1/gait1.mot`: CP subject joint angle data (10 subjects total)- `data/control/P101/V1/gait1.mot`: Control subject joint angle data (10 subjects total)- `processed_pickle/`: Normalized data, SPM stats, deviation ratios- `results/`: Group mean ±SD graphs, PCA scatter plots, heatmaps- `code/`: Python loaders and full Colab analysis pipeline *(recommended)*- `metadata_subjects.csv`: Subject-level metadata (ID, GMFCS, gait speed, etc.)- `README.md`: Dataset usage and structure- `LICENSE.txt`: Shared under CC BY 4.0 --- ## 🧠 Use Cases - Gait phenotyping in pediatric neurorehabilitation- Benchmarking markerless motion capture pipelines- OpenSim-based musculoskeletal simulation- Statistical comparison (SPM), PCA, clustering --- ## 📊 Data Format Each `.mot` file contains normalized gait cycle joint angles for the lower limb (hip, knee, ankle). Units are in degrees, with sampling over 101 points (0–100% gait cycle). | Column | Description | Unit ||------------------|--------------------------------|----------|| `time` | Gait cycle percentage | % || `hip_flexion_r` | Right hip flexion | degrees || `knee_angle_r` | Right knee flexion | degrees || `ankle_angle_r` | Right ankle dorsiflexion | degrees | --- ## 🔒 Ethical & Privacy Statement - All subjects provided informed consent.- All data are fully anonymized and contain no video or identifiable metadata.- Only numerical motion trajectories are shared in OpenSim `.mot` format. --- ## 🧾 Citation If you use this dataset, please cite: > Min YS, Park E, Jung TD, Lee Y, Choi S, et al. > *Data-Driven Clustering of Gait Patterns in Pediatric Neurological Disorders Using Markerless Motion Capture.* > IEEE Trans Neural Syst Rehabil Eng. (2025, in review) --- ## 🔗 License This dataset is shared under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/) license. You are free to use, modify, and redistribute the data with appropriate attribution.