{"success":true,"data":{"author":"Koo S., Boo J., Seo D., Kim M. (Gait120); redistributed by Mathis lab, EPFL","bucketAccessType":"public","bucketName":"amathislab/Gait120-EMG","bucketPrefix":"","category":"Physiology","cloudProvider":"huggingface","createdAt":"2026-11-16T02:38:32.142Z","description":"Level-walking subset of Gait120, the Comprehensive Human Locomotion and Electromyography Dataset of Koo, Boo, Seo and Kim (figshare doi 10.6084/m9.figshare.27677016), repackaged by the Mathis lab at EPFL for MuscleMimic musculoskeletal motor learning. The original release recorded 120 healthy men aged 20 to 59 with optical motion capture, two force plates and surface electromyography on the right lower limb across seven daily tasks. This repository keeps level walking for all 120 participants (60 young, 20 to 25 years; 60 older, 40 to 59): interpolated EMG per trial and step (ConvertedData.mat), OpenSim inverse-kinematics joint angles (.mot), marker trajectories (.trc and .c3d), MoSh++ SMPL body fits, and ground reaction forces for subjects S011 to S120. Measured: 4,203 files, about 8.9 GB, grouped in blocks of ten subjects, with demographics and anthropometrics tables and the original MATLAB processing code. Raw EMG, stairs, slopes and sit-to-stand are in the full figshare release. Useful for gait biomechanics, EMG-driven musculoskeletal models and motion imitation. CC BY 4.0 as stated by the original figshare record; cite Koo et al. 2025.","displayTitle":"Gait120 Level-Walking EMG, Motion Capture and SMPL Fits for 120 Adults","doi":"10.6084/m9.figshare.27677016.v1","id":"f4f45ff2-a173-58d4-847d-508a151b065c","license":"CC-BY-4.0","licenseEvidenceUrl":"https://api.figshare.com/v2/articles/27677016","objectCountEstimate":4203,"probeStatus":"ok","referenceUrl":"https://doi.org/10.6084/m9.figshare.27677016.v1","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"amathislab/Gait120-EMG\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"amathislab/Gait120-EMG\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=amathislab, license:cc-by-4.0, size_categories:n<1K, format:csv, downloads=1745"],"format":"mixed","generatedAt":"2026-11-16T02:38:37.757Z","layout":{"example":"Gait120_001_to_010/S001/MotionCapture/LevelWalking/TRC/Trial01/.smpl_cache/AllSteps_smpl_v1_modelc1c0f828_smplx_male_c3d927f29af3a44_s1shapejoint_dogleg_jax_headcorr51a26b41_s1100.npz","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"Gait120_001_to_010/S001/MotionCapture/LevelWalking/TRC/Trial01/.smpl_cache/AllSteps_smpl_v1_modelc1c0f828_smplx_male_c3d927f29af3a44_s1shapejoint_dogleg_jax_headcorr51a26b41_s1100.npz","truncatedAt":512}],"sampleObjects":["Anthropometric_Data_Participants.xlsx","Gait120_001_to_010/S001/EMG/ConvertedData.mat","Gait120_001_to_010/S001/JointAngle/LevelWalking/Trial01/step01.mot","Gait120_001_to_010/S001/JointAngle/LevelWalking/Trial01/step02.mot","Gait120_001_to_010/S001/JointAngle/LevelWalking/Trial02/step01.mot"],"sidecars":[{"bytes":"Gait120 EMG and Kinematic Data Processing Code\r\n\r\nThis repository contains MATLAB scripts developed for processing and visualizing electromyography (EMG) signals and kinematic data from the Gait120 Dataset. The scripts facilitate EMG data preprocessing, maximum voluntary contraction (MVC) calculations, and kinematic visualization.\r\n\r\nIncluded Files and Descriptions\r\n\r\n1. calculateMVC.m\r\n\r\nCalculates maximum voluntary contraction (MVC) values from processed EMG signals across multiple tasks and trials.\r\n\r\n2. dataVisualizer.m\r\n\r\nVisualizes joint angle and EMG signal data from the Gait120 dataset.\r\n\r\nUtilizes kinematicsVisualizer and EMGSignalVisualizer functions for visualization.\r\n\r\nIncludes the function varycolor originally created by Daniel Helmick (2008), provided under a BSD-style license (see function definition within the file for details).\r\n\r\n3. emgProcessedDataConstruction.m\r\n\r\nProcesses raw EMG signals, performing filtering, rectification, MVC normalization, and interpolation.\r\n\r\nGenerates processed EMG data suitable for further analysis.\r\n\r\n4. signalProcess.m\r\n\r\nApplies bandpass filtering, rectification, outlier removal, and RMS envelope extraction to raw EMG signals.\r\n\r\nDataset Information\r\n\r\nDataset: Comprehensive Human Locomotion and Electromyography Dataset: Gait120\r\n\r\nAuthors: Junyo Boo, Dongwook Seo, Minseung Kim, Seungbum Koo (corresponding author: skoo@kaist.ac.kr)\r\n\r\nDataset DOI: 10.6084/m9.figshare.27677016\r\n\r\nPublication Year: 2025\r\n\r\nLicense Information\r\n\r\nDataset and Original Scripts\r\n\r\nThe dataset and associated MATLAB scripts (calculateMVC.m, dataVisualizer.m, emgProcessedDataConstruction.m, and signalProcess.m) are provided under the Creative Commons Attribution 4.0 International License (CC-BY 4.0).\r\n\r\nYou are free to:\r\n\r\nShare - copy and redistribute the material in any medium or format.\r\n\r\nAdapt - remix, transform, and build upon the material for any purpose, even commercially.\r\n\r\nUnder the following terms:\r\n\r\nAttribution - you must give appropriate credit, provide a link to the license, and indicate if changes were made.\r\n\r\nIncluded Third-party Code\r\n\r\nThe varycolor function, included within dataVisualizer.m, was originally created by Daniel Helmick (2008) and is distributed under a BSD-style license as detailed within the function definition.\r\n\r\nOriginal Source URL: https://kr.mathworks.com/matlabcentral/fileexchange/21050-varycolor\r\n\r\nContact\r\n\r\nFor any questions or additional information, please contact:\r\n\r\nSeungbum Koo, Department of Mechanical Engineering, KAIST, Republic of Korea, Email: skoo@kaist.ac.kr","encoding":"utf8","key":"Matlab_Codes_for_Processing/README.md","name":"README.md","size":2584,"truncated":false},{"bytes":"---\nlicense: cc-by-4.0\npretty_name: \"Gait120 — Human Locomotion & EMG (cleaned subset)\"\ntags:\n  - arxiv:2603.25544\n  - biomechanics\n  - electromyography\n  - emg\n  - gait\n  - motion-capture\n  - musculoskeletal\nsize_categories:\n  - 1B<n<10B\n---\n\n# Gait120 — Human Locomotion & EMG (cleaned subset)\n\n**MuscleMimic paper:** [Towards Embodied AI with MuscleMimic: Unlocking full-body musculoskeletal motor learning at scale](https://arxiv.org/abs/2603.25544).\n\nPer-subject electromyography (EMG), joint kinematics, motion-capture marker data and\nSMPL motion fits for **120 participants** performing **level walking**.\n\n> ⚠️ **This is a modified / trimmed redistribution**, not the complete original archive.\n> See [Relationship to the original dataset](#relationship-to-the-original-dataset).\n\n## Attribution & license\n\nThis dataset is derived from and redistributed under **CC BY 4.0**:\n\n> Koo, Seungbum; Boo, Junyo; Seo, Dongwook; Kim, Minseung (2025).\n> *Comprehensive Human Locomotion and Electromyography Dataset: Gait120.* figshare. Dataset.\n> https://doi.org/10.6084/m9.figshare.27677016.v1\n\nOriginal source: https://springernature.figshare.com/articles/dataset/27677016 ·\nLicense: [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).\nIf you use this data, please cite the original authors above.\n\n## Participants\n\n120 participants, **all male**, organised as two age cohorts (subject ID is confounded with age):\n\n| Cohort | Subjects | N | Mean age | Age range |\n|--------|----------|---|----------|-----------|\n| Young (YG) | S001–S060 | 60 | ~23 yr | 20–25 |\n| Older (OG) | S061–S120 | 60 | ~47 yr | 40–59 |\n\nAggregate (n=120): age 35.4 ± 12.9 yr · height 1.73 ± 0.06 m · body mass 74.2 ± 12.5 kg.\nPer-subject values are in `subject_metadata.csv` (adds an `age_group` column) and in the\noriginal `Subject_Demographics.xlsx` / `Anthropometric_Data_Participants.xlsx`.\n\n## Directory structure\n\nSubjects are grouped in blocks of ten:\n\n```\nGait120_001_to_010/\n  S001/\n    EMG/ConvertedData.mat                         # interpolated EMG per trial/step (LevelWalking)\n    JointAngle/LevelWalking/TrialXX/stepYY.mot    # OpenSim inverse-kinematics joint angles (deg)\n    MotionCapture/LevelWalking/\n      TRC/TrialXX/{Step01.trc, Step02.trc, AllSteps.c3d,\n                   mosh_results/.../AllSteps_stageii.pkl}   # markers, C3D, SMPL (MoSh++) fit\n      MOT/TrialXX/{Step01.mot, Step02.mot}         # ground reaction forces (2 force plates)\n  ...\nSubject_Demographics.xlsx\nAnthropometric_Data_Participants.xlsx\nsubject_metadata.csv\nMatlab_Codes_for_Processing/\n```\n\n### Coverage notes\n- **EMG `ConvertedData.mat`**: all 120 subjects.\n- **JointAngle/LevelWalking** (`.mot`): all 120 subjects.\n- **MotionCapture TRC + SMPL fit** (`AllSteps_stageii.pkl`): all 120 subjects (5 trials each).\n- **Ground reaction forces** (`MotionCapture/.../MOT/`): **S011–S120 only** (110 subjects).\n  S001–S010 have no processed GRF in the source dataset.\n\n## Relationship to the original dataset\n\nTo keep the archive compact, the following were **removed** relative to the figshare release:\n- `EMG/RawData.mat` and `EMG/ProcessedData.mat` (only `ConvertedData.mat` retained);\n- all preview `*.mp4` animations;\n- all non-`LevelWalking` movements (StairAscent/Descent, SlopeAscent/Descent, SitToStand,\n  StandToSit) under `JointAngle/`.\n\n`JointAngle/LevelWalking` was (re)fetched for all 120 subjects so that branch is complete.\nFor the full multi-activity dataset (raw EMG, stairs, slopes, sit-to-stand, etc.),\nuse the original figshare release.\n","encoding":"utf8","key":"README.md","name":"README.md","size":3558,"truncated":false}],"sourceObjectKey":"Gait120_001_to_010/S001/MotionCapture/LevelWalking/TRC/Trial01/.smpl_cache/AllSteps_smpl_v1_modelc1c0f828_smplx_male_c3d927f29af3a44_s1shapejoint_dogleg_jax_headcorr51a26b41_s1100.npz","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"amathislab/Gait120-EMG\", filename=\"Gait120_001_to_010/S001/MotionCapture/LevelWalking/TRC/Trial01/.smpl_cache/AllSteps_smpl_v1_modelc1c0f828_smplx_male_c3d927f29af3a44_s1shapejoint_dogleg_jax_headcorr51a26b41_s1100.npz\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"Anthropometric_Data_Participants.xlsx","totalObjects":4203},"summary":"Level-walking subset of Gait120, the Comprehensive Human Locomotion and Electromyography Dataset of Koo, Boo, Seo and Kim (figshare doi 10.6084/m9.figshare.27677016), repackaged…","title":"Gait120 Level-Walking EMG, Motion Capture and SMPL Fits for 120 Adults","updatedAt":"2026-11-16T15:37:42.918Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/amathislab/Gait120-EMG"}}