{"success":true,"database":"eegdash","data":{"_id":"69d16e04897a7725c66f4c82","dataset_id":"nm000158","associated_paper_doi":null,"authors":["Haijie Liu","Penghu Wei","Haochong Wang","Xiaodong Lv","Wei Duan","Meijie Li","Yan Zhao","Qingmei Wang","Xinyuan Chen","Gaige Shi","Bo Han","Junwei Hao"],"bids_version":"1.9.0","contact_info":null,"contributing_labs":null,"data_processed":true,"dataset_doi":null,"datatypes":["eeg"],"demographics":{"subjects_count":50,"ages":[71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71,71],"age_min":71,"age_max":71,"age_mean":71.0,"species":null,"sex_distribution":null,"handedness_distribution":null},"experimental_modalities":null,"external_links":{"source_url":"https://nemar.org/dataexplorer/detail/nm000158","osf_url":null,"github_url":null,"paper_url":null},"funding":["National Natural Science Foundation of China (grant nos. 82090043 and 81825008)"],"ingestion_fingerprint":"dcd56753ca03011813c3a0180543fc43773a5b5373853c50f25151a7cbe33f04","license":"CC-BY-4.0","n_contributing_labs":null,"name":"Liu, Lv et al. 2023 — EEG datasets of stroke patients (motor imagery)","readme":"# Dataset [1]_ from the study on motor imagery [2]_\nDataset [1]_ from the study on motor imagery [2]_.\n## Dataset Overview\n- **Code**: Liu2024\n- **Paradigm**: imagery\n- **DOI**: 10.1038/s41597-023-02787-8\n- **Subjects**: 50\n- **Sessions per subject**: 1\n- **Events**: left_hand=1, right_hand=2\n- **Trial interval**: (0, 4) s\n- **File format**: MAT and EDF\n- **Data preprocessed**: True\n- **Contributing labs**: Xuanwu Hospital Capital Medical University\n## Acquisition\n- **Sampling rate**: 500.0 Hz\n- **Number of channels**: 29\n- **Channel types**: eeg=29, eog=2\n- **Channel names**: C3, C4, CP3, CP4, Cz, F3, F4, F7, F8, FC3, FC4, FCz, FP1, FP2, FT7, FT8, Fz, HEOL, O1, O2, Oz, P3, P4, Pz, T3, T4, T5, T6, TP7, TP8, VEOR\n- **Montage**: 10-10\n- **Hardware**: ZhenTec NT1 wireless multichannel EEG acquisition system\n- **Reference**: CPz\n- **Ground**: FPz\n- **Sensor type**: semi-dry Ag/AgCl\n- **Line frequency**: 50.0 Hz\n- **Impedance threshold**: 20 kOhm\n- **Cap manufacturer**: Xi'an ZhenTec Intelligence Technology Co., Ltd.\n- **Cap model**: ZhenTec NT1\n- **Electrode type**: semi-dry\n- **Electrode material**: Ag/AgCl semi-dry electrodes based on highly absorbable porous sponges dampened with 3% NaCl solution\n- **Auxiliary channels**: EOG (2 ch, horizontal, vertical)\n## Participants\n- **Number of subjects**: 50\n- **Health status**: acute stroke patients\n- **Clinical population**: acute stroke patients (1-30 days post-stroke)\n- **Age**: mean=56.7, std=10.57, min=31.0, max=77.0\n- **Gender distribution**: male=39, female=11\n## Experimental Protocol\n- **Paradigm**: imagery\n- **Number of classes**: 2\n- **Class labels**: left_hand, right_hand\n- **Trial duration**: 8.0 s\n- **Trials per class**: left_hand=20, right_hand=20\n- **Study design**: Imagining grasping a spherical object with left or right hand while watching a video of gripping motion. Each trial: instruction stage (prompt), MI stage (4s video-guided imagery), break stage (rest).\n- **Feedback type**: none\n- **Stimulus type**: video and audio\n- **Stimulus modalities**: visual, audio\n- **Synchronicity**: cue-based\n- **Mode**: offline\n- **Training/test split**: True\n- **Instructions**: Subject sat approximately 80 cm from computer screen. Computer played audio instructions. Patients imagined grasping spherical object with prompted hand during 4s video playback.\n## HED Event Annotations\nSchema: HED 8.4.0 | Browse: https://www.hedtags.org/hed-schema-browser\n```\n  left_hand\n    ├─ Sensory-event, Experimental-stimulus, Visual-presentation\n    └─ Agent-action\n       └─ Imagine\n          ├─ Move\n          └─ Left, Hand\n  right_hand\n    ├─ Sensory-event, Experimental-stimulus, Visual-presentation\n    └─ Agent-action\n       └─ Imagine\n          ├─ Move\n          └─ Right, Hand\n```\n## Paradigm-Specific Parameters\n- **Detected paradigm**: motor_imagery\n- **Imagery tasks**: left_hand, right_hand\n- **Cue duration**: 2.0 s\n- **Imagery duration**: 4.0 s\n## Data Structure\n- **Trials**: 40\n- **Trials per class**: left_hand=20, right_hand=20\n- **Trials context**: 40 trials per subject total (20 left-hand, 20 right-hand), alternating. Each trial: 8s total (instruction + 4s MI + break). Training/test split: 60%/40%.\n## Preprocessing\n- **Data state**: preprocessed\n- **Preprocessing applied**: True\n- **Steps**: baseline removal (mean removal method), FIR filtering (0.5-40 Hz)\n- **Highpass filter**: 0.5 Hz\n- **Lowpass filter**: 40.0 Hz\n- **Bandpass filter**: [0.5, 40.0]\n- **Filter type**: FIR\n- **Epoch window**: [0.0, 8.0]\n- **Notes**: Preprocessed with EEGLAB toolbox in MATLAB R2019b. Filtered data split into trials x channels x time-samples format by marker '1'. Some motion artifacts present in subjects 4, 5, 13, 14, 18, 24, 28, 33, 42, 43, 47, 48, 49.\n## Signal Processing\n- **Classifiers**: CSP+LDA, FBCSP+SVM, TSLDA+DGFMDRM, TWFB+DGFMDM\n- **Feature extraction**: CSP, FBCSP, ERD/ERS, Riemannian geometry (SCMs on SPD manifolds), Tangent Space, Time-Frequency (Morlet wavelet), TWFB (Time Window Filter Bank)\n- **Frequency bands**: alpha=[8.0, 15.0] Hz; beta=[15.0, 30.0] Hz; analyzed=[8.0, 30.0] Hz\n- **Spatial filters**: CSP, FBCSP, Discriminant Geodesic Filtering\n## Cross-Validation\n- **Method**: 10-fold cross-validation\n- **Folds**: 10\n- **Evaluation type**: within_subject\n## Performance (Original Study)\n- **Csp+Lda Accuracy**: 55.57\n- **Fbcsp+Svm Accuracy**: 57.57\n- **Tslda+Dgfmdrm Accuracy**: 61.2\n- **Twfb+Dgfmdm Accuracy**: 72.21\n- **Twfb+Dgfmdm Kappa**: 0.4442\n- **Twfb+Dgfmdm Precision**: 0.7543\n- **Twfb+Dgfmdm Sensitivity**: 0.7845\n## BCI Application\n- **Applications**: rehabilitation\n- **Environment**: hospital\n- **Online feedback**: False\n## Tags\n- **Pathology**: Stroke\n- **Modality**: Motor\n- **Type**: Motor Imagery\n## Documentation\n- **Description**: EEG motor imagery dataset from 50 acute stroke patients performing left- and right-handed hand-grip imagination tasks. First open dataset addressing left- and right-handed motor imagery in acute stroke patients.\n- **DOI**: 10.1038/s41597-023-02787-8\n- **License**: CC-BY-4.0\n- **Investigators**: Haijie Liu, Penghu Wei, Haochong Wang, Xiaodong Lv, Wei Duan, Meijie Li, Yan Zhao, Qingmei Wang, Xinyuan Chen, Gaige Shi, Bo Han, Junwei Hao\n- **Senior author**: Junwei Hao\n- **Contact**: haojunwei@vip.163.com\n- **Institution**: Xuanwu Hospital Capital Medical University\n- **Department**: Department of Neurology\n- **Address**: Beijing, 100053, China\n- **Country**: CN\n- **Repository**: Figshare\n- **Data URL**: https://doi.org/10.6084/m9.figshare.21679035.v5\n- **Publication year**: 2024\n- **Funding**: National Natural Science Foundation of China (grant nos. 82090043 and 81825008)\n- **Ethics approval**: Ethics Committee of Xuanwu Hospital of Capital Medical University (No. 2021-236)\n- **Keywords**: motor imagery, BCI, brain-computer interface, stroke patients, EEG, rehabilitation, acute stroke, hand-grip imagery, databases, scientific data\n## Abstract\nThe brain-computer interface (BCI) is a technology that involves direct communication with parts of the brain and has evolved rapidly in recent years; it has begun to be used in clinical practice, such as for patient rehabilitation. Patient electroencephalography (EEG) datasets are critical for algorithm optimization and clinical applications of BCIs but are rare at present. We collected data from 50 acute stroke patients with wireless portable saline EEG devices during the performance of two tasks: 1) imagining right-handed movements and 2) imagining left-handed movements. The dataset consists of four types of data: 1) the motor imagery instructions, 2) raw recording data, 3) pre-processed data after removing artefacts and other manipulations, and 4) patient characteristics. This is the first open dataset to address left- and right-handed motor imagery in acute stroke patients.\n## Methodology\n50 acute stroke patients (1-30 days post-stroke) performed 40 trials of hand-grip motor imagery (20 left, 20 right). Each 8s trial included instruction, 4s video-guided imagery, and rest phases. EEG recorded with ZhenTec NT1 wireless system (29 EEG + 2 EOG channels) at 500 Hz. Data organized in EEG-BIDS format with raw (.mat) and preprocessed (.edf) versions. Clinical assessments: NIHSS (mean=4.16, SD=2.85), MBI (mean=70.94, SD=18.22), mRS (mean=2.66, SD=1.44). 23 patients right hemiplegia, 27 left hemiplegia.\n## References\nLiu, Haijie; Lv, Xiaodong (2022). EEG datasets of stroke patients. figshare. Dataset. DOI: https://doi.org/10.6084/m9.figshare.21679035.v5\nLiu, Haijie, Wei, P., Wang, H. et al. An EEG motor imagery dataset for brain computer interface in acute stroke patients. Sci Data 11, 131 (2024). DOI: https://doi.org/10.1038/s41597-023-02787-8\nNotes\nTo add the break and instruction events, set the `break_events` and `instr_events` parameters to True while instantiating the class.\n.. versionadded:: 1.1.1\nAppelhoff, S., Sanderson, M., Brooks, T., Vliet, M., Quentin, R., Holdgraf, C., Chaumon, M., Mikulan, E., Tavabi, K., Hochenberger, R., Welke, D., Brunner, C., Rockhill, A., Larson, E., Gramfort, A. and Jas, M. (2019). MNE-BIDS: Organizing electrophysiological data into the BIDS format and facilitating their analysis. Journal of Open Source Software 4: (1896). https://doi.org/10.21105/joss.01896\nPernet, C. R., Appelhoff, S., Gorgolewski, K. J., Flandin, G., Phillips, C., Delorme, A., Oostenveld, R. (2019). EEG-BIDS, an extension to the brain imaging data structure for electroencephalography. Scientific Data, 6, 103. https://doi.org/10.1038/s41597-019-0104-8\n---\nGenerated by MOABB 1.5.0 (Mother of All BCI Benchmarks)\nhttps://github.com/NeuroTechX/moabb","recording_modality":["eeg"],"senior_author":null,"sessions":["0"],"size_bytes":706372005,"source":"nemar","storage":{"backend":"nemar","base":"s3://nemar/nm000158","raw_key":"dataset_description.json","dep_keys":["README.md","participants.json","participants.tsv"]},"study_design":null,"study_domain":null,"tasks":["imagery"],"timestamps":{"digested_at":"2026-04-30T14:08:43.861312+00:00","dataset_created_at":null,"dataset_modified_at":"2026-04-11T21:11:02Z"},"total_files":50,"computed_title":"Liu, Lv et al. 2023 — EEG datasets of stroke patients (motor imagery)","nchans_counts":[{"val":29,"count":50}],"sfreq_counts":[{"val":500.0,"count":50}],"stats_computed_at":"2026-05-01T13:49:34.645240+00:00","total_duration_s":15999.9,"tagger_meta":{"config_hash":"3557b68bca409f28","metadata_hash":"02b82d22c1d629a6","model":"openai/gpt-5.2","tagged_at":"2026-04-07T09:32:40.872789+00:00"},"tags":{"pathology":["Other"],"modality":["Multisensory"],"type":["Motor"],"confidence":{"pathology":0.85,"modality":0.8,"type":0.85},"reasoning":{"few_shot_analysis":"Closest few-shot match by paradigm is the \"EEG Motor Movement/Imagery Dataset\" example (Healthy / Visual / Motor): it labels motor execution/imagery studies as Type=Motor (construct is movement/imagery rather than e.g., attention/memory). This guides selecting Type=Motor here as well, since the current dataset is explicitly a left-vs-right hand motor imagery paradigm. For Modality conventions, the few-shot set shows that modality is based on stimulus input channel (e.g., tactile braille => Tactile; music/speech listening => Auditory). Here the dataset explicitly includes both video and audio cues, so by the same convention the best fit is Multisensory rather than Motor for modality. For Pathology conventions, few-shot datasets use specific clinical labels when available (e.g., Parkinson's, TBI, Epilepsy). Stroke is not an allowed pathology label, so per catalog constraints the closest allowed label is Other, despite explicit stroke recruitment.","metadata_analysis":"Key participant/pathology facts: \"Health status: acute stroke patients\" and \"Clinical population: acute stroke patients (1-30 days post-stroke)\". Key task/type facts: \"Paradigm: imagery\" and \"Events: left_hand=1, right_hand=2\" with protocol \"Imagining grasping a spherical object with left or right hand\". Key stimulus/modality facts: \"Stimulus type: video and audio\" and \"Stimulus modalities: visual, audio\" plus \"watching a video of gripping motion\" and \"Computer played audio instructions\".","paper_abstract_analysis":"Useful paper information is embedded in the dataset README/Abstract: \"We collected data from 50 acute stroke patients ... during the performance of two tasks: 1) imagining right-handed movements and 2) imagining left-handed movements\" and notes BCI/rehabilitation motivation (\"used in clinical practice, such as for patient rehabilitation\"). This supports Pathology=stroke clinical cohort and Type=Motor (motor imagery BCI).","evidence_alignment_check":"Pathology: Metadata says \"acute stroke patients\" and \"(1-30 days post-stroke)\". Few-shot pattern suggests using a specific disorder label when it exists; however \"Stroke\" is not in the Allowed Labels, so we map to the closest allowed category \"Other\". ALIGN in facts (clinical cohort) but constrained by label set.\nModality: Metadata says stimuli are \"video and audio\" and modalities are \"visual, audio\". Few-shot pattern labels modality by stimulus channel (e.g., auditory listening => Auditory; tactile braille => Tactile). This aligns with choosing \"Multisensory\" here (visual+auditory).\nType: Metadata says \"motor imagery\" and describes imagining left/right hand grasping during a guided video. Few-shot motor/imagery example maps such paradigms to Type=Motor. ALIGN (both indicate motor imagery as the primary construct).","decision_summary":"Top-2 candidates and selection:\n1) Pathology candidates: (a) Other — supported by explicit recruitment of a clinical population \"acute stroke patients\" but no Stroke label exists in Allowed Labels; (b) Unknown — would be used if pathology were unclear, but it is explicitly stated. Winner: Other. Evidence/quotes: \"Health status: acute stroke patients\"; \"Clinical population: acute stroke patients (1-30 days post-stroke)\"; abstract: \"50 acute stroke patients\".\n2) Modality candidates: (a) Multisensory — explicit \"Stimulus type: video and audio\" and \"Stimulus modalities: visual, audio\"; (b) Visual — could be argued as dominant because it is \"video-guided imagery\" and \"watching a video of gripping motion\". Winner: Multisensory because both sensory channels are explicitly part of the stimulus protocol. Evidence/quotes: \"Stimulus type: video and audio\"; \"Stimulus modalities: visual, audio\"; \"Computer played audio instructions\".\n3) Type candidates: (a) Motor — explicit motor imagery left vs right hand; (b) Clinical/Intervention — clinical/rehabilitation context (stroke, BCI for rehab) could justify a clinical-purpose label, but the experimental construct is motor imagery classification. Winner: Motor, guided by the motor imagery few-shot convention. Evidence/quotes: \"Paradigm: imagery\"; \"Imagining grasping a spherical object with left or right hand\"; \"imagining right-handed movements ... imagining left-handed movements\"."}},"canonical_name":null,"name_confidence":0.55,"name_meta":{"suggested_at":"2026-04-14T10:18:35.343Z","model":"openai/gpt-5.2 + openai/gpt-5.4-mini + deterministic_fallback"},"name_source":"author_year","author_year":"Liu2024"}}