EEGdash›OpenNeuro›DS006104

Iss. 6104 · 24 subjects · 56 recordings · CC0

Dataset Brief · EEG dataset for speech decoding

DS006104: eeg dataset, 24 subjects#

EEG dataset for speech decoding

Citation: João Pedro Carvalho Moreira, Vinícius Rezende Carvalho, Eduardo Mazoni Andrade Marçal Mendes, Ariah Fallah, Terrence J. Sejnowski, Claudia Lainscsek, Lindy Comstock (20). EEG dataset for speech decoding. 10.18112/openneuro.ds006104.v1.0.1

24-participant EEG dataset — EEG dataset for speech decoding.

Data & curation João Pedro Carvalho Moreira · Vinícius Rezende Carvalho · Eduardo Mazoni Andrade Marçal Mendes · Ariah Fallah · Terrence J. Sejnowski · Claudia Lainscsek · …
Year 20 · Distributed via OpenNeuro

EEG · 61 (53), 83 (3) ch2000 HzBIDS 1.6.03 tasks2 sessionsHealthyAuditoryPerception

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 DS006104

dataset = DS006104(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 = DS006104(cache_dir="./data", subject="01")

Advanced query

dataset = DS006104(
    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{ds006104,
  title = {EEG dataset for speech decoding},
  author = {João Pedro Carvalho Moreira and Vinícius Rezende Carvalho and Eduardo Mazoni Andrade Marçal Mendes and Ariah Fallah and Terrence J. Sejnowski and Claudia Lainscsek and Lindy Comstock},
  doi = {10.18112/openneuro.ds006104.v1.0.1},
  url = {https://doi.org/10.18112/openneuro.ds006104.v1.0.1},
}

§ 02Study · The README

About This Dataset#

This dataset contains EEG recordings from a phoneme discrimination task with TMS.

The data were collected during two related studies in 2019 and 2021.

Study 1 (2019, Session 01): - 8 participants (P01-P08) - Focus on CV and VC phoneme pairs - 2 blocks: CV pairs and VC pairs - TMS targeted to LipM1 (-56, -8, 46) and TongueM1 (-60, -10, 25)

EEG dataset for speech decoding

Dataset Overview

Study 2 (2021, Session 02): - 16 participants (S01-S16) - Expanded to include single phonemes and phoneme triplets - 4 blocks: single phonemes, CV pairs, real words, and pseudowords - Additional TMS targets included Broca’s area (BA 44: -51, 7, 23) and verbal memory region (BA 6: -46, 1, 41)

View full README

EEG dataset for speech decoding

Dataset Overview

Study 2 (2021, Session 02): - 16 participants (S01-S16) - Expanded to include single phonemes and phoneme triplets - 4 blocks: single phonemes, CV pairs, real words, and pseudowords - Additional TMS targets included Broca’s area (BA 44: -51, 7, 23) and verbal memory region (BA 6: -46, 1, 41)

Task Description

Participants listened to speech sounds and identified stimuli with a button-press response.

The stimuli included: 1. Single phonemes - Consonants (/b/, /p/, /d/, /t/, /s/, /z/) and vowels (/i/, /E/, /A/, /u/, /oU/) 2. Phoneme pairs - CV and VC combinations of the phonemes 3. Phoneme triplets - Real and pseudowords constructed of CVC sequences

TMS Methodology

Detailed information about TMS parameters can be found in the sourcedata/tms_metadata/tms_parameters.json file. TMS was applied using a Magstim Super Rapid Plus1 stimulator with a figure-of-eight 40 mm coil.

Stimulation was delivered at 110% of resting motor threshold as paired pulses with 50ms interpulse interval. Detailed information about the methodology and results can be found in the associated publication:

Moreira et al. “An open-access EEG dataset for speech decoding: Exploring the role of articulation and coarticulation”

Directory Structure

The dataset follows BIDS convention with the following structure: /sub-[subject]/ses-[session]/eeg/ Where subject is P01-P08 for Study 1 and S01-S16 for Study 2.

Session is 01 for Study 1 and 02 for Study 2.

Contact Information

For questions about this dataset, please contact Lindy Comstock at lbcomstock@ucla.edu

§ 03Cohort · Participants

Cohort#

Dataset Statistics#

Channel counts (ch)

6183

Sampling frequencies: 2000.0 Hz (n=56 recordings)

Total recording duration: 50 h

§ 04Signal · Electrodes & trace

Signal · Electrodes & live trace#

Fig. 01 Signal & montage 61 (53), 83 (3) ch · EEG · 2000 Hz · 24 subjects, 56 recordings

Live trace viewer — sub-S07 · ses-02 · task-singlephoneme

Showing one representative recording out of 24 subjects and 56 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 · 61 sensors — 61 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.DS006104(cache_dir, query=None, s3_bucket=None, **kwargs)

Bases: EEGDashDataset

Author (year)Moreira2025

Canonical—

Importable asDS006104 · Moreira2025

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

class eegdash.dataset.DS006104(cache_dir: str, query: dict | None = None, s3_bucket: str | None = None, **kwargs)[source]#

EEG dataset for speech decoding

Study:: ds006104 (OpenNeuro)
Author (year):: Moreira2025
Canonical:: —

Also importable as: DS006104, Moreira2025.

Modality: eeg; Experiment type: Perception; Subject type: Healthy. Subjects: 24; recordings: 56; tasks: 3.

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/ds006104 NeMAR dataset: https://nemar.org/dataexplorer/detail?dataset_id=ds006104 DOI: https://doi.org/10.18112/openneuro.ds006104.v1.0.1

Examples

>>> from eegdash.dataset import DS006104
>>> dataset = DS006104(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 FacePre-bundled mirror at EEGDash/ds006104 · pull with datasets.load_dataset("EEGDash/ds006104").huggingface

Croissant 1.0Machine-readable JSON-LD descriptor — DS006104.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 ds006104 to reproduce the tutorial on this dataset.

Citation

João Pedro Carvalho Moreira, Vinícius Rezende Carvalho, Eduardo Mazoni Andrade Marçal Mendes, Ariah Fallah, Terrence J. Sejnowski, … (20). EEG dataset for speech decoding. 10.18112/openneuro.ds006104.v1.0.1

Provenance

¹Contributed to openneuro in BIDS format.

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

³Persistent identifier: 10.18112/openneuro.ds006104.v1.0.1.

Related & sibling datasets

DS007615EEG · 69 subj DS004256EEG · 53 subj DS007175EEG · 41 subj DS007069EEG · 281 subj DS006735EEG · 27 subj

+ 1 more — see See Also below →

BIDS

BIDS 1.6.0

Sidecars

events · events.json · channels · coordsystem · eeg.json

Provenance

CC0 · 10.18112/openneuro.ds006104.v1.0.1

Machine-readable

schema.org/Dataset · Croissant

Mirrors

OpenNeuro · NEMAR · HuggingFace

Dataset ID	`DS006104`
Title	EEG dataset for speech decoding
Author (year)	`Moreira2025`
Canonical	—
Importable as	`DS006104`, `Moreira2025`
Year	20
Authors	João Pedro Carvalho Moreira, Vinícius Rezende Carvalho, Eduardo Mazoni Andrade Marçal Mendes, Ariah Fallah, Terrence J. Sejnowski, Claudia Lainscsek, Lindy Comstock
License	CC0
Citation / DOI	doi:10.18112/openneuro.ds006104.v1.0.1
Source links	OpenNeuro \| NeMAR \| Source URL

DS006104: eeg dataset, 24 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#