EEGdash›NeMAR›NM000402
Iss. 402 · 29 subjects · 78 recordings · CC-BY-4.0
Dataset Brief · Multi-Expert Seizure Annotation

NM000402: ieeg dataset, 29 subjects#

Multi-Expert Seizure Annotation

Access recordings and metadata through EEGDash.

Citation: William K. Ojemann, Daniel Zhou, Caren Armstrong, Nishant Sinha, Brian Litt, Erin Conrad (2019). Multi-Expert Seizure Annotation. 10.82901/nemar.nm000402

Modality: ieeg Subjects: 29 Recordings: 78 License: CC-BY-4.0 Source: nemar

Metadata: Complete (100%)

29-participant iEEG dataset — Multi-Expert Seizure Annotation.

Data & curation William K. Ojemann · Daniel Zhou · Caren Armstrong · Nishant Sinha · Brian Litt · Erin Conrad
Senior author Erin Conrad
Year 2019 · Distributed via NeMAR
Funding National Science Foundation Grant Graduate Research Fellowship DGE-1845298 · National Institue of Neurological Disorders and Stroke of the National Institutes of Health K99NS138680 · + 9 more
iEEG · 178 (6), 115 (6), 118 (5), 223 (4), 170 (4), 108 (4), 94 (4), 230 (4), 102 (4), 58 (4), 131 (4), 169 (4), 124 (3), 192 (3), 146 (3), 122 (2), 133 (2), 207 (2), 162 (2), 154 (2), 234 (2), 182, 156, 120, 132 ch512, 1024, 2048 HzBIDS 1.7.078 tasks
Layer 01Study
What was asked
Hypothesis, independent & dependent variables, paradigm, cohort, and the editorial caveats around what the recordings can and cannot answer.
Layer 02Signal · BIDS
What was recorded
Sidecars, channels & electrodes, coordinate system, event semantics, and quality stats from the NEMAR pipeline when available.
Layer 03Training · ML
What you can train on
Recommended access modes — MNE Raw, braindecode windows, PyTorch DataLoader — plus the targets the metadata makes addressable.
§ 01Access · Get started

Quickstart#

Install

pip install eegdash

Access the data

from eegdash.dataset import NM000402

dataset = NM000402(cache_dir="./data")
# Get the raw object of the first recording
raw = dataset.datasets[0].raw
print(raw.info)

Filter by subject

dataset = NM000402(cache_dir="./data", subject="01")

Advanced query

dataset = NM000402(
    cache_dir="./data",
    query={"subject": {"$in": ["01", "02"]}},
)

Iterate recordings

for rec in dataset:
    print(rec.subject, rec.raw.info['sfreq'])

If you use this dataset in your research, please cite the original authors.

BibTeX

@dataset{nm000402,
  title = {Multi-Expert Seizure Annotation},
  author = {William K. Ojemann and Daniel Zhou and Caren Armstrong and Nishant Sinha and Brian Litt and Erin Conrad},
  doi = {10.82901/nemar.nm000402},
  url = {https://doi.org/10.82901/nemar.nm000402},
}
§ 02Study · The README

About This Dataset#

Automated seizure detection and localization from intracranial EEG requires validated benchmark datasets with expert annotations, yet existing open datasets lack multi-expert consensus annotations and exclude stimulation-induced seizures. We present stereotactic EEG recordings from 78 seizures (44 spontaneous, 34 stimulation-induced) across 29 patients (19 from the University of Pennsylvania, 10 from the Children’s Hospital of Philadelphia) with drug-resistant epilepsy. Three or five board-certified epileptologists independently annotated each seizure for onset time, onset channels, and channels seizing at 10 seconds post-onset using a standardized protocol. All data follow Brain Imaging Data Structure (BIDS) standards and include electrode localizations, patient demographics, and clinical outcomes. This dataset enables the validation of seizure onset and spread detection and localization against human expert performance and supports comparative analysis of seizure networks across spontaneous and stimulation-induced seizures.

This NEMAR dataset redistributes Pennsieve Discover dataset 530, version 3 (doi:10.26275/dely-fzwx, published 2026-07-29 by

Penn CNT; licence on the record: “Creative Commons Attribution”; licence in the authors’ dataset_description.json: “CC-BY 4.0”). The release was already organised in BIDS by the authors (MNE-BIDS 0.17.0). Contents: 29 participants, 78 SEEG seizure recordings (EDF), 6.8 h in total, sampled at 512-2048 Hz.

DOI

Multi-Expert Seizure Annotation (spontaneous and stimulation-induced seizures in SEEG)

Overview

Contents

  • sub-<HUP|CHOP>###/ses-postimplant/ieeg/: one EDF per seizure, with _channels.tsv, _events.tsv/json and _ieeg.json.

View full README

DOI

Multi-Expert Seizure Annotation (spontaneous and stimulation-induced seizures in SEEG)

Overview

Contents

  • sub-<HUP|CHOP>###/ses-postimplant/ieeg/: one EDF per seizure, with _channels.tsv, _events.tsv/json and _ieeg.json. The task label encodes the seizure type and the approximate onset time in seconds from the start of the implant recording, as named by the authors (task-ictal<onset>: spontaneous seizure; task-stim<onset>: stimulation-induced seizure).

  • participants.tsv/json: age, sex, mesial-temporal epilepsy, unifocal, lesional, Engel outcome, follow-up, disease duration, age at onset, number of seizures and number of stimulation-induced seizures (definitions in participants.json).

  • annotations.tsv/json: the multi-expert annotations, one row per seizure: per-clinician unequivocal electrographic onset (UEO) time, UEO channels, channels seizing 10 s after onset, the consensus values, stimulation parameters, semiology, LVFA at onset (column definitions in annotations.json). Listed in .bidsignore because BIDS has no slot for dataset-level annotation tables.

  • sub-*/ses-postimplant/ieeg/*_space-Other_electrodes.tsv/.json + _coordsystem.json: BIDS electrode files generated from the authors’ tables (HUP: native-space mm coordinates; CHOP: voxel indices only, units n/a). Column label renamed name, size = n/a, other columns kept.

  • derivatives/electrode-localization/: the authors’ per-subject electrode localisation tables (HUP: native mm, tkrRAS and voxel coordinates with DK ROI; CHOP: voxel coordinates, tissue class, DK ROI), moved byte-identically from the release’s non-BIDS sub-*/derivatives/ folders (mapping in sourcedata/pennsieve-530-v3/moved_files.tsv).

  • sourcedata/pennsieve-530-v3/: the Pennsieve record files (readme.md, manifest.json, changelog.md, banner.jpg), the release’s original README and dataset_description.json, the Pennsieve file listing with checksums, and our download verification (every file matched the Pennsieve SHA-256).

Changes made for NEMAR

No signal file (EDF), channels, events or ieeg sidecar was modified. Changes: - the authors’ electrode tables were moved byte-identically to derivatives/electrode-localization/ (see above), and BIDS

_space-Other_electrodes.tsv/.json + _coordsystem.json were generated from them (HUP: native-space mm coordinates; CHOP: voxel indices only, units n/a; label renamed name, size = n/a; the authors’ matter Levels map moved into the column Description because the values use other spellings such as ‘grey’);

  • participants.json outcome: the integer Levels map was moved into the Description (values are Engel subclasses such as 1.1, which the validator rejects against integer levels); participants.tsv values unchanged;

  • dataset_description.json: License written as SPDX CC-BY-4.0, Authors taken from the Pennsieve record’s contributor list, Pennsieve DOI added to ReferencesAndLinks and HowToAcknowledge;

  • this README replaces the release README (kept verbatim below and in sourcedata).

Privacy

EDF headers carry no patient information (patient field X X X X, start date 01.01.85 set by the authors); scans.tsv acq_time is n/a. Subject labels are the authors’ study codes. Clinician names are replaced by Clin # in the release.

Ethics approval

From the authors’ dataset_description.json (EthicsApprovals): “University of Pennsylvania Human Research Protections Program, Institutional Review Boards (Protocol 703979, 811097, and/or 821778)”.

Funding

As listed by the authors in dataset_description.json (Funding).

How to cite

Cite the dataset (doi:10.26275/dely-fzwx) and the manuscript https://doi.org/10.64898/2026.01.15.26344025.

Licence

Creative Commons Attribution 4.0 (CC-BY-4.0), as released by the authors.

Original release README

References

Appelhoff, S., Sanderson, M., Brooks, T., Vliet, M., Quentin, R., Holdgraf, C., Chaumon, M., Mikulan, E., Tavabi, K., Höchenberger, 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 Holdgraf, C., Appelhoff, S., Bickel, S., Bouchard, K., D’Ambrosio, S., David, O., … Hermes, D. (2019). iEEG-BIDS, extending the Brain Imaging Data Structure specification to human intracranial electrophysiology. Scientific Data, 6, 102. https://doi.org/10.1038/s41597-019-0105-7

Additional metadata and localisation (added 2026-10-08)

Compiled after the upload from the article, its supplement and the source deposit (each statement names its source). Text and sidecar metadata only; no data file was changed. Stimulation. HUP: bipolar, biphasic stimulation at 1 Hz, 3 mA, 300 µs (first 14 patients) or 500 µs pulse width; CHOP: 1–8 mA, 1–2 Hz, 300–500 µs (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods). Per-seizure stim channels/parameters are in annotations.tsv (stim_channels, stim_frequency, stim_amplitude, stim_pulse_width). Recording system.**Natus Quantum system, sampling rate 512–2048 Hz (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods); the deposited EDFs are at 1024 Hz with channels.tsv low_cutoff 0.0 / high_cutoff 512.0 (deposit *_ieeg.json SamplingFrequency, *_channels.tsv). PowerLineFrequency, SoftwareFilters and Manufacturer are n/a in the sidecars. **Reference scheme. Referenced to a contact hypothesised to be in non-epileptogenic tissue, typically medullary bone (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods); iEEGReference is n/a in the deposit sidecars. Electrode types. sEEG depth electrodes; HUP: Ad-Tech (Oak Creek, WI); CHOP: PMT (MN, USA) or DIXI Medical (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods). Localisation method.**CHOP: post-implant CT co-registered to pre-implant MRI with Gardel, Desikan–Killiany (DK) atlas segmentation in FreeSurfer. HUP: iEEG-recon, contacts localised in pre-implant T1 space and MNI152 space, FreeSurfer DK parcellation (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods “Electrode Localization”). The deposit gives per-subject ``derivatives/electrodes.tsv``: CHOP files have voxel x/y/z + matter + DK brain_area; HUP files have mm_x/y/z, surfmm_x/y/z, vox_x/y/z, roi (FreeSurfer aseg/DK name) and roiNum. The MNI coordinates mentioned in the paper are not** in the deposit; the mm coordinates are subject-native (deposit does not name the space).

NEMAR Metadata#

[![DOI](https://img.shields.io/badge/DOI-10.82901%2Fnemar.nm000402-blue)](https://doi.org/10.82901/nemar.nm000402) # Multi-Expert Seizure Annotation (spontaneous and stimulation-induced seizures in SEEG) ## Overview Automated seizure detection and localization from intracranial EEG requires validated benchmark datasets with expert annotations, yet existing open datasets lack multi-expert consensus annotations and exclude stimulation-induced seizures. We present stereotactic EEG recordings from 78 seizures (44 spontaneous, 34 stimulation-induced) across 29 patients (19 from the University of Pennsylvania, 10 from the Children’s Hospital of Philadelphia) with drug-resistant epilepsy. Three or five board-certified epileptologists independently annotated each seizure for onset time, onset channels, and channels seizing at 10 seconds post-onset using a standardized protocol. All data follow Brain Imaging Data Structure (BIDS) standards and include electrode localizations, patient demographics, and clinical outcomes. This dataset enables the validation of seizure onset and spread detection and localization against human expert performance and supports comparative analysis of seizure networks across spontaneous and stimulation-induced seizures. This NEMAR dataset redistributes Pennsieve Discover dataset 530, version 3 (doi:10.26275/dely-fzwx, published 2026-07-29 by Penn CNT; licence on the record: “Creative Commons Attribution”; licence in the authors’ dataset_description.json: “CC-BY 4.0”). The release was already organised in BIDS by the authors (MNE-BIDS 0.17.0). Contents: 29 participants, 78 SEEG seizure recordings (EDF), 6.8 h in total, sampled at 512-2048 Hz. ## Contents - sub-<HUP|CHOP>###/ses-postimplant/ieeg/: one EDF per seizure, with _channels.tsv, _events.tsv/json and _ieeg.json.

The task label encodes the seizure type and the approximate onset time in seconds from the start of the implant recording, as named by the authors (task-ictal<onset>: spontaneous seizure; task-stim<onset>: stimulation-induced seizure).

  • participants.tsv/json: age, sex, mesial-temporal epilepsy, unifocal, lesional, Engel outcome, follow-up, disease duration, age at onset, number of seizures and number of stimulation-induced seizures (definitions in participants.json).

  • annotations.tsv/json: the multi-expert annotations, one row per seizure: per-clinician unequivocal electrographic onset (UEO) time, UEO channels, channels seizing 10 s after onset, the consensus values, stimulation parameters, semiology, LVFA at onset (column definitions in annotations.json). Listed in .bidsignore because BIDS has no slot for dataset-level annotation tables.

  • sub-*/ses-postimplant/ieeg/*_space-Other_electrodes.tsv/.json + _coordsystem.json: BIDS electrode files generated from the authors’ tables (HUP: native-space mm coordinates; CHOP: voxel indices only, units n/a). Column label renamed name, size = n/a, other columns kept.

  • derivatives/electrode-localization/: the authors’ per-subject electrode localisation tables (HUP: native mm, tkrRAS and voxel coordinates with DK ROI; CHOP: voxel coordinates, tissue class, DK ROI), moved byte-identically from the release’s non-BIDS sub-*/derivatives/ folders (mapping in sourcedata/pennsieve-530-v3/moved_files.tsv).

  • sourcedata/pennsieve-530-v3/: the Pennsieve record files (readme.md, manifest.json, changelog.md, banner.jpg), the release’s original README and dataset_description.json, the Pennsieve file listing with checksums, and our download verification (every file matched the Pennsieve SHA-256).

## Changes made for NEMAR No signal file (EDF), channels, events or ieeg sidecar was modified. Changes: - the authors’ electrode tables were moved byte-identically to derivatives/electrode-localization/ (see above), and BIDS

_space-Other_electrodes.tsv/.json + _coordsystem.json were generated from them (HUP: native-space mm coordinates; CHOP: voxel indices only, units n/a; label renamed name, size = n/a; the authors’ matter Levels map moved into the column Description because the values use other spellings such as ‘grey’);

  • participants.json outcome: the integer Levels map was moved into the Description (values are Engel subclasses such as 1.1, which the validator rejects against integer levels); participants.tsv values unchanged;

  • dataset_description.json: License written as SPDX CC-BY-4.0, Authors taken from the Pennsieve record’s contributor list, Pennsieve DOI added to ReferencesAndLinks and HowToAcknowledge;

  • this README replaces the release README (kept verbatim below and in sourcedata).

## Privacy EDF headers carry no patient information (patient field X X X X, start date 01.01.85 set by the authors); scans.tsv acq_time is n/a. Subject labels are the authors’ study codes. Clinician names are replaced by Clin # in the release. ## Ethics approval From the authors’ dataset_description.json (EthicsApprovals): “University of Pennsylvania Human Research Protections Program, Institutional Review Boards (Protocol 703979, 811097, and/or 821778)”. ## Funding As listed by the authors in dataset_description.json (Funding). ## How to cite Cite the dataset (doi:10.26275/dely-fzwx) and the manuscript https://doi.org/10.64898/2026.01.15.26344025. ## Licence Creative Commons Attribution 4.0 (CC-BY-4.0), as released by the authors. ## Original release README References ———- Appelhoff, S., Sanderson, M., Brooks, T., Vliet, M., Quentin, R., Holdgraf, C., Chaumon, M., Mikulan, E., Tavabi, K., Höchenberger, 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 Holdgraf, C., Appelhoff, S., Bickel, S., Bouchard, K., D’Ambrosio, S., David, O., … Hermes, D. (2019). iEEG-BIDS, extending the Brain Imaging Data Structure specification to human intracranial electrophysiology. Scientific Data, 6, 102. https://doi.org/10.1038/s41597-019-0105-7 ## Additional metadata and localisation (added 2026-10-08) Compiled after the upload from the article, its supplement and the source deposit (each statement names its source). Text and sidecar metadata only; no data file was changed. Stimulation. HUP: bipolar, biphasic stimulation at 1 Hz, 3 mA, 300 µs (first 14 patients) or 500 µs pulse width; CHOP: 1–8 mA, 1–2 Hz, 300–500 µs (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods). Per-seizure stim channels/parameters are in annotations.tsv (stim_channels, stim_frequency, stim_amplitude, stim_pulse_width). Recording system. Natus Quantum system, sampling rate 512–2048 Hz (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods); the deposited EDFs are at 1024 Hz with channels.tsv low_cutoff 0.0 / high_cutoff 512.0 (deposit _ieeg.json SamplingFrequency, *_channels.tsv). PowerLineFrequency, SoftwareFilters and Manufacturer are n/a in the sidecars. **Reference scheme.* Referenced to a contact hypothesised to be in non-epileptogenic tissue, typically medullary bone (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods); iEEGReference is n/a in the deposit sidecars. Electrode types. sEEG depth electrodes; HUP: Ad-Tech (Oak Creek, WI); CHOP: PMT (MN, USA) or DIXI Medical (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods). Localisation method. CHOP: post-implant CT co-registered to pre-implant MRI with Gardel, Desikan–Killiany (DK) atlas segmentation in FreeSurfer. HUP: iEEG-recon, contacts localised in pre-implant T1 space and MNI152 space, FreeSurfer DK parcellation (doi:10.64898/2026.01.15.26344025 (medRxiv, Ojemann et al. 2026), Methods “Electrode Localization”). The deposit gives per-subject derivatives/electrodes.tsv: CHOP files have voxel x/y/z + matter + DK brain_area; HUP files have mm_x/y/z, surfmm_x/y/z, vox_x/y/z, roi (FreeSurfer aseg/DK name) and roiNum. The MNI coordinates mentioned in the paper are not in the deposit; the mm coordinates are subject-native (deposit does not name the space).

License: CC-BY-4.0

Authors:

  • William K. Ojemann

  • Daniel Zhou

  • Caren Armstrong

  • Nishant Sinha

  • Brian Litt

  • … and 1 more

Versions:

Version

DOI

Released

current

10.82901/nemar.nm000402
§ 03Cohort · Participants

Cohort#

Dataset Statistics#

Age distribution by gender (n=28, range 4–58 yr, mean 29.8 yr)

0510152025303540455055
Female · 17Male · 11

Sex composition

29
subjects
Female
17
Male
12
F : M ratio
1.42 : 1
59% female · n = 29 subjects with reported sex.

Channel counts (ch)

5894102108115118120122124131132133146154156162169170178182192207223230234

Sampling frequencies (Hz)

51210242048

Total recording duration: 6 h 47 min

§ 04Signal · Electrodes & trace

Signal · Electrodes & live trace#

Fig. 01 Signal & montage 178 (6), 115 (6), 118 (5), 223 (4), 170 (4), 108 (4), 94 (4), 230 (4), 102 (4), 58 (4), 131 (4), 169 (4), 124 (3), 192 (3), 146 (3), 122 (2), 133 (2), 207 (2), 162 (2), 154 (2), 234 (2), 182, 156, 120, 132 ch · iEEG · 512, 1024, 2048 Hz · 29 subjects, 78 recordings
Live trace viewer — sub-CHOP005 · ses-postimplant · task-stim68881 · run-00

Showing one representative recording out of 29 subjects and 78 recordings in this dataset. Browse the full set on OpenNeuro; drop any other _ieeg.{set,edf,bdf,vhdr} file onto the viewer (or pass ?ieeg=<url>) to inspect it.

Electrode layout — iEEG · 60 sensors — 60 channels

NEMAR Processing Statistics#

The plots below are generated by NEMAR’s automated EEG pipeline. The histogram shows pipeline success for data cleaning and ICA decomposition, the percentage of data frames and EEG channels retained after artefact removal, line noise per channel (RMS, dB), and the age/gender distribution of participants.

NEMAR pipeline statistics — NM000402
HED event descriptors word cloud HED event descriptors word cloud — NM000402
§ 05Manifest · BIDS tree

Manifest#

File Explorer#

Browse the BIDS file structure of this dataset. Records are fetched on demand from the EEGDash catalog the first time you open the explorer.

Recordings—
Files—
Subjects—
Modalities—
Click to load file structure…
Full dataset metadata table

Dataset ID

NM000402

Title

Multi-Expert Seizure Annotation

Author (year)

—

Canonical

—

Importable as

NM000402

Year

2019

Authors

William K. Ojemann, Daniel Zhou, Caren Armstrong, Nishant Sinha, Brian Litt, Erin Conrad

License

CC-BY-4.0

Citation / DOI

10.82901/nemar.nm000402

Source links

OpenNeuro | NeMAR | Source URL
Copy-paste BibTeX
@dataset{nm000402,
  title = {Multi-Expert Seizure Annotation},
  author = {William K. Ojemann and Daniel Zhou and Caren Armstrong and Nishant Sinha and Brian Litt and Erin Conrad},
  doi = {10.82901/nemar.nm000402},
  url = {https://doi.org/10.82901/nemar.nm000402},
}
§ 06API · Programmatic access

API Reference#

Signature
eegdash.dataset
class
eegdash.dataset.NM000402(cache_dir, query=None, s3_bucket=None, **kwargs)
Bases: EEGDashDataset
Author (year)—
Canonical—
Importable asNM000402
Sourceeegdash/dataset/registry.py · [source ↗]
class eegdash.dataset.NM000402(cache_dir: str, query: dict | None = None, s3_bucket: str | None = None, **kwargs)[source]#

Multi-Expert Seizure Annotation

Study:

nm000402 (NeMAR)

Author (year):

—

Canonical:

—

Also importable as: NM000402.

Modality: ieeg; Subject type: Unknown. Subjects: 29; recordings: 78; tasks: 78.

Parameters:

  • cache_dir (str | Path) – Directory where data are cached locally.

  • query (dict | None) – Additional MongoDB-style filters to AND with the dataset selection. Must not contain the key dataset.

  • s3_bucket (str | None) – Base S3 bucket used to locate the data.

  • **kwargs (dict) – Additional keyword arguments forwarded to EEGDashDataset.

data_dir#

Local dataset cache directory (cache_dir / dataset_id).

Type:

Path

query#

Merged query with the dataset filter applied.

Type:

dict

records#

Metadata records used to build the dataset, if pre-fetched.

Type:

list[dict] | None

Notes

Each item is a recording; recording-level metadata are available via dataset.description. query supports MongoDB-style filters on fields in ALLOWED_QUERY_FIELDS and is combined with the dataset filter. Dataset-specific caveats are not provided in the summary metadata.

References

OpenNeuro dataset: https://openneuro.org/datasets/nm000402 NeMAR dataset: https://nemar.org/dataexplorer/detail?dataset_id=nm000402 DOI: https://doi.org/10.82901/nemar.nm000402

Examples

>>> from eegdash.dataset import NM000402
>>> dataset = NM000402(cache_dir="./data")
>>> recording = dataset[0]
>>> raw = recording.load()
__init__(cache_dir: str, query: dict | None = None, s3_bucket: str | None = None, **kwargs)[source]#
save(path: str, overwrite: bool = False, offset: int = 0)[source]#

Save datasets to files by creating one subdirectory for each dataset:

path/
    0/
        0-raw.fif | 0-epo.fif
        description.json
        raw_preproc_kwargs.json (if raws were preprocessed)
        window_kwargs.json (if this is a windowed dataset)
        window_preproc_kwargs.json  (if windows were preprocessed)
        target_name.json (if target_name is not None and dataset is raw)
    1/
        1-raw.fif | 1-epo.fif
        description.json
        raw_preproc_kwargs.json (if raws were preprocessed)
        window_kwargs.json (if this is a windowed dataset)
        window_preproc_kwargs.json  (if windows were preprocessed)
        target_name.json (if target_name is not None and dataset is raw)
Parameters:

  • path (str) –

    Directory in which subdirectories are created to store

    -raw.fif | -epo.fif and .json files to.

  • overwrite (bool) – Whether to delete old subdirectories that will be saved to in this call.

  • offset (int) – If provided, the integer is added to the id of the dataset in the concat. This is useful in the setting of very large datasets, where one dataset has to be processed and saved at a time to account for its original position.

Access modesMNE → braindecode → PyTorch → ML
.rawMNE Raw object — standard tools (filter, epoch, ICA, plot_psd).mne
BaseConcatDatasetEach record is a lazy BaseDataset from braindecode — windowed via create_windows_from_events.braindecode
DataLoaderWraps the windowed dataset into a PyTorch DataLoader; supports parallel workers and on-the-fly augmentations.pytorch
Zarr cacheOptional braindecode Zarr mirror for fast resume; persisted to cache_dir.zarr
Hugging FaceNo per-dataset mirror published yet — browse the EEGDash org listing for sibling datasets. See the datasets loader API.huggingface
Croissant 1.0Machine-readable JSON-LD descriptor — NM000402.croissant.json (MLCommons schema, ingestible by PyTorch / TensorFlow / JAX).mlcommons
Examples using EEGDashcurated · start here
Find datasets with the EEGDash APIQuery the catalogue, filter by task or modality, list candidates.Load one EEG recordingResolve a single record to an MNE Raw with channels and events.EEG recording to PyTorch DataLoaderWrap braindecode windows in a DataLoader for model training.Preprocess EEG and create windowsFilter, resample, epoch — and persist the windowed dataset.Save and reload prepared dataCache a windowed dataset to disk and reattach it without recompute.Download a dataset locallyPrefetch BIDS files to a local cache and validate the layout.

Swap any load_dataset(...) call for nm000402 to reproduce the tutorial on this dataset.

Citation

William K. Ojemann, Daniel Zhou, Caren Armstrong, Nishant Sinha, Brian Litt, … (2019). Multi-Expert Seizure Annotation. 10.82901/nemar.nm000402

Provenance

¹Contributed to nemar in BIDS format.

²Curated & ingested by the EEGDash catalog; see CITATION.cff for canonical reference.

³Persistent identifier: 10.82901/nemar.nm000402.

BIDS
BIDS 1.7.0
Sidecars
events · events.json · channels · eeg.json
Provenance
CC-BY-4.0 · 10.82901/nemar.nm000402
Machine-readable
schema.org/Dataset · Croissant
Mirrors
OpenNeuro · NEMAR · HF org

See Also#