EEGdash›NeMAR›ON008257

Iss. 8257 · 6 subjects · 768 recordings · CC0

Dataset Brief · EEG Moments Dataset (EMD)

ON008257: eeg dataset, 6 subjects#

EEG Moments Dataset (EMD)

Access recordings and metadata through EEGDash.

Citation: Alessandro T. Gifford, Pablo Oyarzo, Anne W. Zonneveld, Christina Sartzetaki, Iris I.A. Groen, Radoslaw M. Cichy (20). EEG Moments Dataset (EMD). 10.82901/nemar.on008257

Modality: eeg Subjects: 6 Recordings: 768 License: CC0 Source: nemar

Metadata: Complete (100%)

6-participant EEG dataset — EEG Moments Dataset (EMD).

Data & curation Alessandro T. Gifford · Pablo Oyarzo · Anne W. Zonneveld · Christina Sartzetaki · Iris I.A. Groen · Radoslaw M. Cichy
Senior author Radoslaw M. Cichy
Year 20 · Distributed via NeMAR

EEG · 127 ch1000 HzBIDS 1.9.0Task · video8 sessionsHealthyMultisensoryPerception

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#

Get Started

Install

pip install eegdash

Access the data

from eegdash.dataset import ON008257

dataset = ON008257(cache_dir="./data")
# Get the raw object of the first recording
raw = dataset.datasets[0].raw
print(raw.info)

Query & Filter

Filter by subject

dataset = ON008257(cache_dir="./data", subject="01")

Advanced query

dataset = ON008257(
    cache_dir="./data",
    query={"subject": {"$in": ["01", "02"]}},
)

Iterate recordings

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

Cite This Dataset

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

BibTeX

@dataset{on008257,
  title = {EEG Moments Dataset (EMD)},
  author = {Alessandro T. Gifford and Pablo Oyarzo and Anne W. Zonneveld and Christina Sartzetaki and Iris I.A. Groen and Radoslaw M. Cichy},
  doi = {10.82901/nemar.on008257},
  url = {https://doi.org/10.82901/nemar.on008257},
}

§ 02Study · The README

About This Dataset#

This is the data repository for the EEG Moments Dataset (EMD). EMD contains EEG responses to 1,102 3-second videos across 6 human subjects. During the EMD experiment, the video stimuli were presented with the corresponding audio track, so as to enable analyses of visual and/or auditory processing of naturalistic dynamic events. Each subject saw the 1,000 video training set 6 times, and the 102 video testing set 24 times. Each video is additionally human-annotated with 15 object labels, 5 scene labels, 5 action labels, 5 sentence text descriptions, 1 spoken transcription, 1 memorability score, and 1 memorability decay rate.

EMD is the EEG companion dataset of the BOLD Moments Dataset (BMD), which consists of fMRI responses for the same 1,102 3-second videos. EMD additionally contains eye tracking data (gaze and pupil size) collected during the EEG experiment. Note that subjects were instructed to maintain central fixation during the stimulus video presentation.

The home folder (everything except the ./stimuli and the ./derivatives/ folders) contains the raw EEG and eye-tracking data in BIDS format before any preprocessing. The eye-tracking data is denoted as physio in the corresponding file names. Download this folder if you want to run your own preprocessing pipeline.

DOI

EEG Moments Dataset (EMD)

The ./stimuli/ folder contains a .txt file with instructions to access the video stimuli which, due to copyright permission considerations, must be downloaded separately. The ./derivatives/ folder contains all data derivatives, including the stimulus metadata (./derivatives/stimuli_metadata/), preprocessed EEG data (./derivatives/eeg/), and preprocessed eye-tracking data (./derivatives/eyetracking/).

📝 Data collection notes

View full README

DOI

EEG Moments Dataset (EMD)

The ./stimuli/ folder contains a .txt file with instructions to access the video stimuli which, due to copyright permission considerations, must be downloaded separately. The ./derivatives/ folder contains all data derivatives, including the stimulus metadata (./derivatives/stimuli_metadata/), preprocessed EEG data (./derivatives/eeg/), and preprocessed eye-tracking data (./derivatives/eyetracking/).

📝 Data collection notes

🧠 Missing EEG data

Subject 1

Session 3:
- Run 10: The first EEG video trial is missing, as the EEG recording only starts ~400ms after the onset of the first video trial.

Subject 5

Session 7:
- Run 16: The EEG recording only contains the first 64 (out of 66) video trials.

👁️ Missing/Bad eye tracking data

Subject 3

Subject 3 exhibited more substantial eye movements, which the source authors attribute as likely to calibration difficulties associated with wearing glasses rather than actual eye movements per se.

Subject 4

Session 5:
- Run 5: The eye tracking recording only contains the first 39 (out of 66) video trials.
Session 6:
- Run 10: The eye tracking recording only contains the first 23 (out of 66) video trials.
- Run 12: The eye tracking recording only contains the first 33 (out of 66) video trials.

Subject 5

Session 6:
- Run 11: The eye tracking recording only contains the first 50 (out of 66) video trials.

💻 Code

The code we used for collecting, preprocessing and analyzing the EEG Moments Dataset (EMD) is available on GitHub.

If you wish to familiarize with EMD’s preprocessed EEG and eye tracking data, check out this interactive Colab tutorial.

📧 Contact

For any question regarding the EEG Moments Dataset, you can get in touch with Ale Gifford (alessandro.gifford@gmail.com).

📜 Citation

If you use EMD’s data, please cite the paper:

Gifford AT, Oyarzo P, Zonneveld AW, Sartzetaki C, Groen IIA, Cichy RM. 2026. A large dataset of human EEG responses to short naturalistic videos for studying dynamic visual event processing. _arXiv_. DOI: https://doi.org/10.48550/arXiv.2608.28768

If you use EMD’s stimuli or stimulus metadata, please also cite the paper:

Lahner B, Dwivedi K, Iamshchinina P, Graumann M, Lascelles A, Roig G, Gifford AT, Pan B, Jin S, Murty AR, Kay K, Oliva A, Cichy RM. 2024. Modeling short visual events through the BOLD moments video fMRI dataset and metadata. _Nature Communications_. DOI: https://doi.org/10.1038/s41467-024-50310-3

Metadata reuse and source-version notes

See REUSE_METADATA.md for stimulus joins, annotation provenance, known metadata caveats, and stimulus-access restrictions. This metadata-only enrichment does not change the archived recordings or their source release.

Stimuli (metadata only)

No video is included in this dataset. The 1,102 videos are the stimulus set of the BOLD Moments Dataset (BMD; Lahner et al. 2024, https://doi.org/10.1038/s41467-024-50310-3), 3-second clips that derive from Moments in Time. They belong to third parties. Obtain them from the official BMD page https://boldmomentsdataset.csail.mit.edu/stimuli_metadata after accepting its terms (non-commercial research and education only; no redistribution of videos, images, tags or text). See also stimuli/stimuli_download.txt. stimuli/stimuli.tsv has one row per stim_id (the integer in every *_events.tsv; 1-1000 train, 1001-1102 test): duration, frame rate, size, frame count, audio track and SHA-256 of the BMD stimulus_set MP4 (H.264) file with that number, so you can check your own copy. Over the 1,102 files: 268 x 268 pixels; 3.00-3.11 s; 30 fps (823 files), 24 (146), 25 (84), 15 (30), other rates (19).

Known issue: none of the 1,102 BMD MP4 files has an audio track, while EMD presented the videos with their audio (see above). The audio heard in the experiment therefore most likely came from another version of each clip (an inference from the missing audio tracks, not stated by the authors), and the SHA-256 identifies the BMD file, not necessarily the exact presentation file. 34 videos have no audio at all (derivatives/stimuli_metadata/videos_with_no_audio.csv).

Join: stim_id in the events = stim_id in stimuli/stimuli.tsv = BMD video number = the zero-padded keys of derivatives/stimuli_metadata/annotations.json.

Added 2026-10-08 (metadata enrichment, see CHANGES). Properties were measured on the files obtained under the BMD terms.

§ 03Cohort · Participants

Cohort#

Dataset Statistics#

Age distribution by gender (n=6, range 21–32 yr, mean 24.8 yr)

202530

Female · 3Male · 3

Sex composition

Female
3
Male
3
F : M ratio
1.00 : 1

50% female · n = 6 subjects with reported sex.

Channel counts: 127 ch (n=768 recordings)

Sampling frequencies: 1000.0 Hz (n=768 recordings)

Total recording duration: 97 h

§ 04Signal · Electrodes & trace

Signal · Electrodes & live trace#

Fig. 01 Signal & montage 127 ch · EEG · 1000 Hz · 6 subjects, 768 recordings

Live trace viewer — sub-05 · ses-01 · task-video · run-13

Showing one representative recording out of 6 subjects and 768 recordings in this dataset. Browse the full set on OpenNeuro; drop any other _eeg.{set,edf,bdf,vhdr} file onto the viewer (or pass ?eeg=<url>) to inspect it.

Electrode layout — EEG · 125 sensors — 125 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.

HED event descriptors word cloud

§ 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…

§ 06API · Programmatic access

API Reference#

Signature

eegdash.dataset

class

eegdash.dataset.ON008257(cache_dir, query=None, s3_bucket=None, **kwargs)

Bases: EEGDashDataset

Author (year)—

Canonical—

Importable asON008257

Sourceeegdash/dataset/registry.py · [source ↗]

class eegdash.dataset.ON008257(cache_dir: str, query: dict | None = None, s3_bucket: str | None = None, **kwargs)[source]#

EEG Moments Dataset (EMD)

Study:: on008257 (NeMAR)
Author (year):: —
Canonical:: —

Also importable as: ON008257.

Modality: eeg; Experiment type: Perception; Subject type: Healthy. Subjects: 6; recordings: 768; tasks: 1.

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/on008257 NeMAR dataset: https://nemar.org/dataexplorer/detail?dataset_id=on008257 DOI: https://doi.org/10.82901/nemar.on008257

Examples

>>> from eegdash.dataset import ON008257
>>> dataset = ON008257(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 — ON008257.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 on008257 to reproduce the tutorial on this dataset.

Citation

Alessandro T. Gifford, Pablo Oyarzo, Anne W. Zonneveld, Christina Sartzetaki, Iris I.A. Groen, … (20). EEG Moments Dataset (EMD). 10.82901/nemar.on008257

Provenance

¹Contributed to nemar in BIDS format.

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

³Persistent identifier: 10.82901/nemar.on008257.

Related & sibling datasets

ON004306EEG · 12 subj ON005672EEG · 3 subj ON005815EEG · 20 subj ON003846EEG · 19 subj ON008083EEG · 43 subj

+ 1 more — see See Also below →

BIDS

BIDS 1.9.0

Sidecars

events · channels

Provenance

CC0 · 10.82901/nemar.on008257

Machine-readable

schema.org/Dataset · Croissant

Mirrors

OpenNeuro · NEMAR · HF org

Dataset ID	`ON008257`
Title	EEG Moments Dataset (EMD)
Author (year)	—
Canonical	—
Importable as	`ON008257`
Year	20
Authors	Alessandro T. Gifford, Pablo Oyarzo, Anne W. Zonneveld, Christina Sartzetaki, Iris I.A. Groen, Radoslaw M. Cichy
License	CC0
Citation / DOI	10.82901/nemar.on008257
Source links	OpenNeuro \| NeMAR \| Source URL

ON008257: eeg dataset, 6 subjects#

Quickstart#

About This Dataset#

Cohort#

Dataset Statistics#

Signal · Electrodes & live trace#

NEMAR Processing Statistics#

Manifest#

File Explorer#

API Reference#

Citation

Provenance

Related & sibling datasets

See Also#