EEGDash Corpus v0#

EEGDash Corpus v0 is the public part of REVE’s pretraining corpus, rebuilt from the archives and converted: 77 datasets, 14,074 hours, 7,272 subjects. Every number on this page is measured in the converted Zarr store, as of 2026-11-03; Download says how to get it.

In numbers#

77datasets
14,074hours
7,272subjects
41,133sessions
632channel sets (buckets)
641electrode positions
10native sampling rates
10paradigms
Measured in the converted Zarr store: 77 datasets, 14,074 h from 7,272 subjects, in 632 channel sets.

Download#

EEGDash Corpus v0 comes in two forms: the converted Zarr store, ready for training, and the raw datasets it was converted from, fetched through EEGDash. The corpus table, eegdash-corpus-v0.csv, lists the 77 datasets, one row each: the dataset id, EEGDash’s id for it, the name, paradigm and archive, subjects, sessions, hours, mean channel count, native sampling rate, size of the raw copy, licence and the terms it sets. This table ships with EEGDash: use the name "eegdash-corpus-v0" in Python, with no separate CSV download. The object fetches the raw archive files, not the converted Zarr store.

The converted store#

The store is published on Hugging Face as a gated dataset repository, EEGDash/eegdash-corpus-v0, at tag v1.0.

Access. A Hugging Face account that accepted the repository’s terms, and a token on the machine that downloads (hf auth login). Every dataset keeps its own licence and the collection has none of its own; LICENSES.csv in the repository lists them: CC0-1.0 (52 datasets), CC-BY-SA-4.0 (12), CC-BY-4.0 (11), CC-BY-NC-4.0 (1: SparrKULee) and GPL-3.0 (1: Shin2017A).

What it holds. One Zarr v3 store, eegdash-corpus-v0.zarr, with one group per bucket (one dataset and one exact channel set; 632 buckets): 5,045,111 samples of 10 seconds at 200 Hz (2,000 points), in float32 microvolts, band-passed 0.5–99.5 Hz (0.5–90 Hz for the three datasets recorded at 200 Hz) and not normalised, with each session’s per-channel mean and standard deviation stored beside them and the electrode positions in metres. In all 14,074 hours from 7,272 subjects, 3.60 TB.

How it was built. The code that converted the raw datasets into the store and trained REVE on it is published in eegdash/eegdash-corpus (BSD-3-Clause), with the records of the conversion: the sessions and hours kept per dataset, the recordings skipped and the channels dropped.

Download everything, or a part with --include:

hf download EEGDash/eegdash-corpus-v0 --repo-type dataset --local-dir eegdash-corpus-v0          # everything, 3.6 TB
hf download EEGDash/eegdash-corpus-v0 --repo-type dataset --local-dir eegdash-corpus-v0 \
    --include "eegdash-corpus-v0.zarr/zarr.json" "eegdash-corpus-v0.zarr/0/**"                    # the root and bucket 0

Load with zarr (version 3 or later):

import numpy as np
import zarr  # zarr >= 3

root = zarr.open_group("eegdash-corpus-v0/eegdash-corpus-v0.zarr", mode="r")
print(root.attrs["n_samples"], root.attrs["licences"]["ds004395"])        # 5045111 CC0-1.0

b = root["0"]
print(b.attrs["dataset"], b.attrs["licence"], b.attrs["source"], len(b.attrs["channels"]))
x = b["data"][0]                            # (2000, C) float32, µV: one 10-s sample
mean, sd = b["stats"][int(b["session"][0])] # its session's per-channel mean and SD
z = np.clip((x - mean) / sd, -15, 15)       # REVE's normalisation, applied at load
pos = b["positions"][:]                     # (C, 3), metres

# the subset without NC/GPL licences (CC0, CC BY, CC BY-SA), from the root's index
without_nc_gpl = {"CC0-1.0", "CC-BY-4.0", "CC-BY-SA-4.0"}
keep = [str(r["bucket"]) for r in root.attrs["buckets"] if root.attrs["licences"][r["dataset"]] in without_nc_gpl]

CC BY still requires attribution and CC BY-SA retains its share-alike terms; this filter does not replace any dataset’s licence.

Integrity. eegdash-corpus-v0-SHA256SUMS holds the SHA-256 of every file (44,801 files). From the download root, sha256sum -c eegdash-corpus-v0-SHA256SUMS checks them (add --ignore-missing after a partial download). The list’s own SHA-256, at tag v1.0, is 0cba6a7b2106d8f9853c04d07f647256c32aa0c51ae4f8bc0953eaa882e32e54.

The raw datasets#

The raw copy is what the converter turned into the store: the archives’ own files, fetched through EEGDash into a cache that mirrors each archive’s layout. The ids of the table are EEGDash’s: OpenNeuro datasets ds…, NEMAR datasets nm…. EEGDash serves most OpenNeuro datasets from their re-host on NEMAR, on…, and caches them under that id; the corpus object follows it, offline too, so the table’s ds… ids work as they are. One object, EEGDashCorpus, plans, downloads, verifies and reports the whole list, from the command line or from Python. A path to a CSV with a dataset_id column works too, for a custom corpus.

The Python interface accepts the shipped name directly; the current command-line interface accepts a CSV path or comma-separated dataset ids, not a corpus name. For the CLI commands below, first save the optional corpus table linked above as eegdash-corpus-v0.csv. No separate table is needed for the Python example.

Plan. The files each dataset needs, from EEGDash’s catalogue, without downloading:

python -m eegdash.corpus plan eegdash-corpus-v0.csv --cache-dir eegdash-cache

One row per dataset: the id it is served under, its state, the files expected and how many are present, missing or corrupt, the bytes present and expected, the subjects. A dataset the catalogue does not serve, or whose catalogue lists fewer subjects than the table, is reported as a problem. Each dataset’s plan is written to <cache>/<dataset>/.eegdash/manifest.jsonl; planned again later, the report’s drift (--json report.json) lists the files the catalogue added or dropped since.

Download. A few datasets at a time, smallest first:

python -m eegdash.corpus download eegdash-corpus-v0.csv --cache-dir eegdash-cache \
    --n-jobs 16 --n-datasets 4 --time-budget 21600     # 16 workers over 4 datasets; no new dataset after 6 h

A dataset already complete is skipped after a size check, a file already at its size is not fetched again, and the datasets the time budget left out are reported as deferred: the same line, run again, moves only what is missing.

Verify. Offline, every expected file against its size (--level size) or also against the archive’s checksum (--level hash, which reads every file); a dataset whose files all pass gets a .eegdash/COMPLETE marker, removed when one fails:

python -m eegdash.corpus verify eegdash-corpus-v0.csv --cache-dir eegdash-cache --level hash

Status. Offline and without reading the files, from each dataset’s .eegdash/manifest.jsonl and COMPLETE:

python -m eegdash.corpus status eegdash-corpus-v0.csv --cache-dir eegdash-cache

Every command exits with status 1 when a dataset has a problem, and download, verify and status also when a file is missing. In Python:

from eegdash import EEGDashCorpus

corpus = EEGDashCorpus("eegdash-corpus-v0", cache_dir="eegdash-cache")
print(corpus.plan())                                           # the catalogue: files, bytes
corpus.download(n_jobs=16, n_datasets=4, time_budget=6 * 3600)
print(corpus.verify(level="hash"))                             # sizes, then checksums
print(corpus.status())                                         # offline, from the manifests
for rec in corpus.records():                                   # every recording, offline
    raw = rec.raw                                              # an MNE Raw, read on access

For a custom list, pass dataset ids or a CSV you maintain:

subset = EEGDashCorpus(["ds004395", "nm000138"], cache_dir="eegdash-cache")
custom = EEGDashCorpus("my-corpus.csv", cache_dir="eegdash-cache")

The CSV needs a dataset_id column; an optional n_subjects column is checked against the catalogue during planning.

Electrode coverage#

One mark per electrode position of the Zarr store (641 positions in 632 channel sets), coloured and sized by the hours recorded there or by the number of datasets that include it.

CzPzC3C4O1O2T7T8Fp1FzFp2Ozhours of data at the position (log scale)1 h10 h100 h1,000 h10,000 hdatasets (of 77) that include the position1102565

Channel set (bucket): one exact set of channel names of one dataset in the Zarr store.

641 electrode positions; Cz carries the most data (11,182 h over 65 datasets); 22 positions appear in a single dataset. 12 electrodes of two datasets were placed from their own coordinates and are not drawn.

Drag, or focus the view and use the arrow keys, to rotate. Channel set (bucket): one exact set of channel names of one dataset in the Zarr store.

641 electrode positions; Cz carries the most data (11,182 h over 65 datasets); 22 positions appear in a single dataset. 12 electrodes of two datasets were placed from their own coordinates and are not drawn.

Hours per dataset#

Order by
1 h10 h100 h1,000 h10,000 h
PEERS (ds004395)9,110.8 hNMT (nm000181)489.0 hSpatial memory (ds004706)470.6 hHBN R7 (ds005511)279.7 hHBN R4 (ds005508)261.8 hHBN R5 (ds005509)255.3 hHBN R9 (ds005514)210.8 hLee2019-MI (nm000338)183.1 hHBN R8 (ds005512)179.1 hHBN R3 (ds005507)158.8 hSparrKULee (nm000238)155.7 hSiena (nm000385)141.0 hHBN R2 (ds005506)127.5 hDreyer2023 (nm000250)126.9 hHBN R1 (ds005505)117.5 hLee2019ERP (nm000323)116.2 hHBN R6 (ds005510)103.5 hTHINGS-EEG2 (nm000232)87.3 hTest-retest rest (ds004148)75.0 hPerceive/imagine (ds005697)69.5 hMeditation vs think (ds003969)66.5 hSelective attention (ds005089)62.8 hSpanish sentences (ds004279)53.7 hHypnosis (ds004572)53.2 hSeen vs unseen (ds005273)52.5 hImagined emotion (ds003004)48.5 hRest & auditory (ds003690)47.7 hTHINGS-EEG1 (ds003825)46.3 hMusic vs speech (ds004356)46.3 hLee2019SSVEP (nm000273)40.6 hTask not stated (ds005121)40.5 hFake faces (ds004582)34.2 hNumerosity (ds005586)33.5 hContext trajectory (ds004603)30.7 hShin2017A (nm000267)29.0 hSchirrmeister2017 (nm000172)28.7 hMeditation (ds001787)27.6 hObject motion 1 (ds003885)27.1 hOfner2017 (nm000173)27.1 hObject motion 2 (ds003887)26.8 hAffective music (ds002721)26.3 hBI2015b (nm000217)26.1 hBraille letters (ds004951)25.8 hArabic inner speech (ds005262)25.0 hPrediction error (ds003846)22.7 hMood induction (ds004315)21.1 hGo/no-go (ds002680)20.8 hNatural speech (ds004408)20.6 hCho2017 (nm000245)20.5 hRecollection (ds005189)20.1 hSosulski2019 (nm000266)19.6 hToon faces (ds004324)19.2 hFeature attention (ds004043)18.2 hSleep deprivation (ds004902)18.0 hAnaesthesia (ds005620)17.9 hSternberg (ds005095)16.9 hPandemic scenario (ds004477)13.6 hWeibo2014 (nm000146)13.1 hBI2014a (nm000244)12.4 hAlphabetic decision (ds005594)12.3 hRSVP attention 2 (ds004817)11.9 hBI2015a (nm000216)11.7 hDrum trainer (ds004152)11.5 hAgency illusion (ds004561)11.4 hFeatures-EEG (ds004357)11.3 hMusic therapy (ds004840)10.9 hRSVP attention 1 (ds004816)10.7 hSRM rest (ds003775)10.1 hTask value (ds004147)9.6 hNeuroforecasting (ds004284)9.5 hLiu2024 (stroke) (nm000158)8.9 hFeedback (ds004262)8.3 hSteer the ship (ds004264)7.5 hKalunga2016 (nm000123)2.6 hParkinson rest (ds002778)2.5 hBI2014b (nm000215)2.4 hAlexMI (nm000138)1.1 h
The largest dataset (PEERS) holds 65 % of the 14,074 h shown; the median dataset has 27.1 h.

Channel count per dataset#

Both bar lists follow the same order; switch it on either.

Order by
0326496128160192224
PEERS (ds004395)126 chNMT (nm000181)21 chSpatial memory (ds004706)128 chHBN R7 (ds005511)129 chHBN R4 (ds005508)129 chHBN R5 (ds005509)129 chHBN R9 (ds005514)129 chLee2019-MI (nm000338)62 chHBN R8 (ds005512)129 chHBN R3 (ds005507)129 chSparrKULee (nm000238)64 chSiena (nm000385)30 chHBN R2 (ds005506)129 chDreyer2023 (nm000250)27 chHBN R1 (ds005505)129 chLee2019ERP (nm000323)62 chHBN R6 (ds005510)129 chTHINGS-EEG2 (nm000232)63 chTest-retest rest (ds004148)61 chPerceive/imagine (ds005697)62 chMeditation vs think (ds003969)64 chSelective attention (ds005089)63 chSpanish sentences (ds004279)62 chHypnosis (ds004572)58 chSeen vs unseen (ds005273)61 chImagined emotion (ds003004)208 chRest & auditory (ds003690)61 chTHINGS-EEG1 (ds003825)66 chMusic vs speech (ds004356)32 chLee2019SSVEP (nm000273)62 chTask not stated (ds005121)58 chFake faces (ds004582)59 chNumerosity (ds005586)60 chContext trajectory (ds004603)64 chShin2017A (nm000267)30 chSchirrmeister2017 (nm000172)128 chMeditation (ds001787)64 chObject motion 1 (ds003885)123 chOfner2017 (nm000173)61 chObject motion 2 (ds003887)123 chAffective music (ds002721)19 chBI2015b (nm000217)32 chBraille letters (ds004951)63 chArabic inner speech (ds005262)8 chPrediction error (ds003846)63 chMood induction (ds004315)60 chGo/no-go (ds002680)31 chNatural speech (ds004408)128 chCho2017 (nm000245)64 chRecollection (ds005189)62 chSosulski2019 (nm000266)31 chToon faces (ds004324)28 chFeature attention (ds004043)63 chSleep deprivation (ds004902)61 chAnaesthesia (ds005620)62 chSternberg (ds005095)63 chPandemic scenario (ds004477)64 chWeibo2014 (nm000146)60 chBI2014a (nm000244)16 chAlphabetic decision (ds005594)64 chRSVP attention 2 (ds004817)63 chBI2015a (nm000216)32 chDrum trainer (ds004152)31 chAgency illusion (ds004561)62 chFeatures-EEG (ds004357)123 chMusic therapy (ds004840)8 chRSVP attention 1 (ds004816)63 chSRM rest (ds003775)64 chTask value (ds004147)31 chNeuroforecasting (ds004284)129 chLiu2024 (stroke) (nm000158)30 chFeedback (ds004262)31 chSteer the ship (ds004264)31 chKalunga2016 (nm000123)8 chParkinson rest (ds002778)32 chBI2014b (nm000215)32 chAlexMI (nm000138)16 ch
Median 62 channels; 25 % of the 77 datasets have more than 64 channels and 23 have 32 or fewer.

Distributions#

Channel-count bands#

1–16 channels6 datasets · 53 h (0 %)17–32 channels18 datasets · 1,007 h (7 %)33–64 channels36 datasets · 1,566 h (11 %)65–128 channels8 datasets · 9,695 h (69 %)>128 channels12 datasets · 1,754 h (12 %)
69 % of the 14,075 converted hours come from 65–128-channel recordings (8 datasets); a dataset recorded with several montages counts in each band it uses.

Subjects#

1–10 subjects5 datasets · 126 h11–30 subjects30 datasets · 704 h31–100 subjects30 datasets · 1,941 h101–300 subjects7 datasets · 907 h>300 subjects5 datasets · 10,397 h
Most datasets (30 of 77) have 11–30 subjects; the 5 with more than 300 subjects hold 74 % of the hours.

Native sampling rates#

200 Hz3 datasets · 531 h250 Hz1 dataset · 25 h256 Hz3 datasets · 79 h500 Hz18 datasets · 11,046 h512 Hz12 datasets · 433 h1,000 Hz29 datasets · 1,093 h1,024 Hz5 datasets · 274 h2,048 Hz2 datasets · 484 h5,000 Hz1 dataset · 18 h10,000 Hz3 datasets · 92 h
The converter resampled 74 of the 77 converted datasets to 200 Hz from 9 native rates, from 250 to 10,000 Hz; 3 were recorded at 200 Hz.

Paradigms#

The datasets fall into 10 paradigm groups.

memory & cognition5 datasets · 9,659 hdevelopment (rest & tasks)9 datasets · 1,694 hperception & emotion20 datasets · 671 hBCI (MI & P300)15 datasets · 627 hclinical routine EEG1 dataset · 489 hresting & consciousness9 datasets · 318 hauditory & language6 datasets · 312 hseizure1 dataset · 141 hlearning & reward9 datasets · 120 hSSVEP2 datasets · 43 h
10 paradigms; memory & cognition holds 69 % of the hours with 5 datasets; perception & emotion has the most datasets (20).

Paradigm summary#

The full breakdown behind the chart: dataset and subject counts, hours, tokens per epoch and their shares. Hours are summed from the channel-set buckets before rounding.

ParadigmDatasetsSubjectsHoursHours (%)Tokens (B)Tokens (%)
memory & cognition55109,65968.64.7977.2
development (rest & tasks)92,1921,69412.00.8513.7
perception & emotion205986714.80.203.2
BCI (MI & P300)155846274.50.132.1
clinical routine EEG12,4134893.50.040.7
resting & consciousness95423192.30.081.2
auditory & language61983122.20.071.1
seizure1141411.00.020.3
learning & reward91551200.90.030.4
SSVEP266430.30.010.2
Total777,27214,075100.06.21100.0
The corpus by paradigm, ordered by hours. Hours and token shares use unrounded totals; subjects are summed across datasets. Tokens per epoch: 11 patches per channel in each 10-second window (1-second patches, 0.9-second stride). B means billion.

Subjects against hours#

1101001,00010,0001 h10 h100 h1,000 h10,000 hsubjectsPEERS (ds004395)NMT (nm000181)Spatial memory (ds004706)HBN ×2
Median 0.8 h per subject over 77 datasets; PEERS records the most, 25.0 h per subject.

Recordings#

Recording length#

How long one recording lasts in each dataset, summarised by its quantiles (the per-recording lengths themselves are not shipped), on a log axis from minutes to days.

0.1 min1 min10 min1 h10 h100 h
Siena (nm000385)2.5 hMusic vs speech (ds004356)2.0 hSelective attention (ds005089)1.9 hSeen vs unseen (ds005273)1.7 hSpatial memory (ds004706)1.5 hPandemic scenario (ds004477)1.5 hPerceive/imagine (ds005697)90 minPEERS (ds004395)85 minNumerosity (ds005586)84 minImagined emotion (ds003004)82 minWeibo2014 (nm000146)79 minBraille letters (ds004951)74 minObject motion 2 (ds003887)68 minObject motion 1 (ds003885)67 minTask not stated (ds005121)66 minTHINGS-EEG1 (ds003825)55 minSpanish sentences (ds004279)54 minFeature attention (ds004043)53 minContext trajectory (ds004603)49 minTask value (ds004147)47 minAlphabetic decision (ds005594)45 minMeditation (ds001787)45 minFeatures-EEG (ds004357)42 minToon faces (ds004324)41 minRecollection (ds005189)40 minRSVP attention 2 (ds004817)35 minNeuroforecasting (ds004284)32 minRSVP attention 1 (ds004816)32 minDrum trainer (ds004152)31 minPrediction error (ds003846)31 minSchirrmeister2017 (nm000172)31 minAgency illusion (ds004561)29 minFake faces (ds004582)28 minLee2019-MI (nm000338)26 minMood induction (ds004315)25 minFeedback (ds004262)24 minCho2017 (nm000245)23 minLee2019SSVEP (nm000273)22 minSteer the ship (ds004264)21 minSternberg (ds005095)21 minTHINGS-EEG2 (nm000232)18 minLee2019ERP (nm000323)16 minSparrKULee (nm000238)15 minNMT (nm000181)12 minBI2014a (nm000244)12 minMeditation vs think (ds003969)10 minShin2017A (nm000267)10 minMusic therapy (ds004840)10 minAffective music (ds002721)9.3 minBI2015b (nm000217)8.8 minRest & auditory (ds003690)8.3 minAlexMI (nm000138)8.3 minArabic inner speech (ds005262)8.1 minDreyer2023 (nm000250)7.5 minHypnosis (ds004572)6.0 minOfner2017 (nm000173)5.4 minLiu2024 (stroke) (nm000158)5.3 minTest-retest rest (ds004148)5.0 minSleep deprivation (ds004902)5.0 minAnaesthesia (ds005620)5.0 minKalunga2016 (nm000123)4.9 minBI2015a (nm000216)4.9 minHBN R2 (ds005506)4.4 minHBN R3 (ds005507)4.2 minHBN R1 (ds005505)4.2 minHBN R7 (ds005511)4.2 minHBN R6 (ds005510)4.2 minHBN R4 (ds005508)4.1 minHBN R5 (ds005509)4.1 minHBN R9 (ds005514)4.1 minHBN R8 (ds005512)4.0 minSRM rest (ds003775)4.0 minBI2014b (nm000215)3.7 minGo/no-go (ds002680)3.5 minParkinson rest (ds002778)3.2 minNatural speech (ds004408)3.2 minSosulski2019 (nm000266)0.5 min
The median recording lasts 4.2 min over 41,133 recordings of 77 datasets (per-dataset medians weighted by recordings); medians run from 0.5 min (Sosulski2019) to 2.5 h (Siena), and Siena has the longest recordings, up to 13.5 h. Bars span the 10th to 90th percentile, boxes the middle half, the tick is the median and the dot the longest recording.

Recordings per subject#

How much each subject contributes: recordings and hours per subject, the median subject and the subject with the most.

recordings per subject1101001,000hours per subject0.01 h0.1 h1 h10 h100 h
PEERS (ds004395)24.6 hSpatial memory (ds004706)12.7 hTHINGS-EEG2 (nm000232)8.8 hSiena (nm000385)7.8 hLee2019-MI (nm000338)3.4 hBraille letters (ds004951)2.4 hSchirrmeister2017 (nm000172)2.2 hLee2019ERP (nm000323)2.1 hMusic vs speech (ds004356)2.1 hArabic inner speech (ds005262)2.0 hSparrKULee (nm000238)1.9 hSelective attention (ds005089)1.9 hOfner2017 (nm000173)1.8 hSeen vs unseen (ds005273)1.7 hPandemic scenario (ds004477)1.5 hDreyer2023 (nm000250)1.5 hPerceive/imagine (ds005697)1.5 hGo/no-go (ds002680)1.5 hSosulski2019 (nm000266)1.4 hNumerosity (ds005586)1.4 hImagined emotion (ds003004)1.4 hWeibo2014 (nm000146)1.3 hMusic therapy (ds004840)1.3 hTest-retest rest (ds004148)1.2 hTask not stated (ds005121)1.2 hPrediction error (ds003846)1.2 hObject motion 2 (ds003887)1.1 hObject motion 1 (ds003885)1.1 hNatural speech (ds004408)1.1 hMeditation (ds001787)1.1 hHypnosis (ds004572)1.0 hShin2017A (nm000267)1.0 hHBN R3 (ds005507)1.0 hHBN R1 (ds005505)0.9 hHBN R2 (ds005506)0.9 hTHINGS-EEG1 (ds003825)0.9 hSpanish sentences (ds004279)0.9 hFeature attention (ds004043)0.9 hAnaesthesia (ds005620)0.9 hAffective music (ds002721)0.9 hContext trajectory (ds004603)0.8 hHBN R6 (ds005510)0.8 hHBN R5 (ds005509)0.8 hHBN R4 (ds005508)0.8 hTask value (ds004147)0.8 hHBN R8 (ds005512)0.8 hHBN R9 (ds005514)0.8 hHBN R7 (ds005511)0.8 hAlphabetic decision (ds005594)0.8 hLee2019SSVEP (nm000273)0.7 hFeatures-EEG (ds004357)0.7 hToon faces (ds004324)0.7 hMeditation vs think (ds003969)0.7 hRecollection (ds005189)0.7 hRest & auditory (ds003690)0.6 hRSVP attention 2 (ds004817)0.6 hBI2015b (nm000217)0.6 hNeuroforecasting (ds004284)0.5 hRSVP attention 1 (ds004816)0.5 hDrum trainer (ds004152)0.5 hAgency illusion (ds004561)0.5 hFake faces (ds004582)0.5 hMood induction (ds004315)0.4 hFeedback (ds004262)0.4 hCho2017 (nm000245)0.4 hSteer the ship (ds004264)0.4 hSternberg (ds005095)0.3 hSleep deprivation (ds004902)0.3 hBI2015a (nm000216)0.2 hNMT (nm000181)0.2 hBI2014a (nm000244)0.2 hLiu2024 (stroke) (nm000158)0.2 hKalunga2016 (nm000123)0.2 hAlexMI (nm000138)0.1 hSRM rest (ds003775)0.1 hBI2014b (nm000215)0.1 hParkinson rest (ds002778)0.1 h
The median dataset records 2 recordings and 0.9 h per subject; PEERS records the most per subject (24.6 h median), Parkinson rest the least (0.06 h), and Sosulski2019 has the most recordings for one subject (200). Dots mark the median subject, lines reach the subject with the most.

People#

Ages and sex of the subjects, for the datasets whose records state them.

age (years)020406080100femalemaleshare of subjects
Liu2024 (stroke) (nm000158)Parkinson rest (ds002778)17 F · 14 MRest & auditory (ds003690)60 F · 15 MSiena (nm000385)6 F · 8 MBraille letters (ds004951)8 F · 3 MMeditation (ds001787)12 F · 12 MMeditation vs think (ds003969)26 F · 72 MTask value (ds004147)7 F · 5 MMood induction (ds004315)28 F · 22 MAffective music (ds002721)18 F · 13 MSRM rest (ds003775)69 F · 42 MPandemic scenario (ds004477)6 F · 3 MDreyer2023 (nm000250)42 F · 45 MLee2019ERP (nm000323)Music therapy (ds004840)1 F · 8 MTHINGS-EEG2 (nm000232)8 F · 2 MShin2017A (nm000267)Schirrmeister2017 (nm000172)Anaesthesia (ds005620)12 F · 8 MPrediction error (ds003846)Drum trainer (ds004152)16 F · 5 MFeedback (ds004262)16 F · 5 MPerceive/imagine (ds005697)25 F · 25 MSteer the ship (ds004264)16 F · 5 MFake faces (ds004582)39 F · 37 MAlphabetic decision (ds005594)12 F · 4 MBI2014b (nm000215)Cho2017 (nm000245)Weibo2014 (nm000146)Imagined emotion (ds003004)20 F · 14 MBI2014a (nm000244)BI2015a (nm000216)BI2015b (nm000217)Numerosity (ds005586)15 F · 8 MRecollection (ds005189)20 F · 10 MRSVP attention 1 (ds004816)9 F · 6 MSosulski2019 (nm000266)Spanish sentences (ds004279)31 F · 25 MPEERS (ds004395)172 F · 143 MRSVP attention 2 (ds004817)17 F · 3 MSelective attention (ds005089)18 F · 18 MSternberg (ds005095)33 F · 15 MMusic vs speech (ds004356)11 F · 11 MSleep deprivation (ds004902)34 F · 37 MTest-retest rest (ds004148)32 F · 28 MTHINGS-EEG1 (ds003825)36 F · 14 MAgency illusion (ds004561)17 F · 6 MContext trajectory (ds004603)30 F · 7 MFeatures-EEG (ds004357)11 F · 5 MObject motion 1 (ds003885)15 F · 9 MObject motion 2 (ds003887)16 F · 7 MHBN R1 (ds005505)46 F · 90 MHBN R2 (ds005506)54 F · 96 MHBN R3 (ds005507)63 F · 121 MHBN R4 (ds005508)103 F · 221 MHBN R5 (ds005509)109 F · 221 MHBN R6 (ds005510)52 F · 83 MHBN R7 (ds005511)125 F · 256 MHBN R8 (ds005512)91 F · 166 MHBN R9 (ds005514)103 F · 192 MGo/no-go (ds002680)7 F · 7 MSparrKULee (nm000238)74 F · 11 M
Ages known for 60 of 77 datasets, sex for 50; median ages run from 9 (HBN R9) to 71 (Liu2024 (stroke)); 45 % of the 3,991 subjects with a recorded sex are female. Datasets with neither are not shown.

From archive to licence#

Where the hours come from and under which licence; switch the flow width to datasets to count datasets instead.

Width by
OpenNeuro · 12,531 hNEMAR · 1,543 hCC0-1.0 · 11,211 hCC-BY-SA-4.0 · 2,204 hCC-BY-4.0 · 475 hCC-BY-NC-4.0 · 156 hGPL-3.0 · 29 harchivelicence
OpenNeuro · 56NEMAR · 21CC0-1.0 · 52CC-BY-SA-4.0 · 12CC-BY-4.0 · 11CC-BY-NC-4.0 · 1GPL-3.0 · 1archivelicence
14,074 h over 77 of the 77 datasets: OpenNeuro hosts 89 % and CC0-1.0 covers 80 %.

Training budget#

What one training epoch reads: 6.21 B tokens, per dataset, grouped by paradigm, each tile shaded by its share of the tokens.

PEERS (ds004395)73 %memory & cognitionHBN R7 (ds005511)2 %HBN … (ds005508)2 %HBN R52 %HBN R92 %HBN R81 %HBN R21 %HBN R11 %development (rest & tasks)perception & emot…BCI (MI & P300)share of the tokens< 0.01 %0.01–0.03 %0.03–0.1 %0.1–0.3 %0.3–1 %1–3 %3–10 %≥ 10 %
  • memory & cognition 77 %
  • development (rest & tasks) 14 %
  • perception & emotion 3 %
  • BCI (MI & P300) 2 %
  • resting & consciousness 1 %
  • auditory & language 1 %
  • clinical routine EEG 0.7 %
  • learning & reward 0.4 %
  • seizure 0.3 %
  • SSVEP 0.2 %
6.21 B tokens per epoch over 77 datasets: PEERS alone holds 73 %, the three largest (PEERS, Spatial memory, HBN R7) 79 %; areas are tokens, colour is each dataset's share of them (darker is larger, legend below), outlines group the datasets by the 10 paradigms listed below.

Montages#

A channel is one electrode. The columns are the datasets, by channel count, and the rows the 69 standard names used by at least 10 datasets, so nets with their own naming, such as EGI or BioSemi, show only the electrodes whose names coincide.

Fp1FpzFp2AF7AF3AFzAF4AF8F7F5F3F1FzF2F4F6F8FT9FT7FT8FT10FC5FC3FC1FCzFC2FC4FC6T7T8C5C3C1CzC2C4C6TP9TP7TP8TP10CP5CP3CP1CPzCP2CP4CP6P9P7P5P3P1PzP2P4P6P8P10PO9PO7PO3POzPO4PO10PO8O1OzO2
129–497 channels · 12 datasetsPEERS (ds004395)Imagined emotion (ds003004)HBN R1 (ds005505)HBN R2 (ds005506)HBN R3 (ds005507)HBN R4 (ds005508)HBN R5 (ds005509)HBN R6 (ds005510)HBN R7 (ds005511)HBN R8 (ds005512)HBN R9 (ds005514)Neuroforecasting (ds004284)65–128 · 7Schirrmeister2017 (nm000172)Natural speech (ds004408)Spatial memory (ds004706)THINGS-EEG1 (ds003825)Features-EEG (ds004357)Object motion 1 (ds003885)Object motion 2 (ds003887)33–64 · 35Alphabetic decision (ds005594)Cho2017 (nm000245)Context trajectory (ds004603)Meditation vs think (ds003969)Pandemic scenario (ds004477)SparrKULee (nm000238)SRM rest (ds003775)Meditation (ds001787)Braille letters (ds004951)Feature attention (ds004043)Mood induction (ds004315)Prediction error (ds003846)RSVP attention 1 (ds004816)RSVP attention 2 (ds004817)Selective attention (ds005089)Sternberg (ds005095)THINGS-EEG2 (nm000232)Agency illusion (ds004561)Recollection (ds005189)Anaesthesia (ds005620)Perceive/imagine (ds005697)Spanish sentences (ds004279)Lee2019-MI (nm000338)Lee2019ERP (nm000323)Lee2019SSVEP (nm000273)Seen vs unseen (ds005273)Sleep deprivation (ds004902)Test-retest rest (ds004148)Rest & auditory (ds003690)Ofner2017 (nm000173)Numerosity (ds005586)Weibo2014 (nm000146)Fake faces (ds004582)Task not stated (ds005121)Hypnosis (ds004572)17–32 · 18BI2014b (nm000215)BI2015a (nm000216)BI2015b (nm000217)Music vs speech (ds004356)Parkinson rest (ds002778)Drum trainer (ds004152)Feedback (ds004262)Steer the ship (ds004264)Task value (ds004147)Siena (nm000385)Sosulski2019 (nm000266)Go/no-go (ds002680)Liu2024 (stroke) (nm000158)Shin2017A (nm000267)Toon faces (ds004324)Dreyer2023 (nm000250)NMT (nm000181)Affective music (ds002721)1–16 · 5AlexMI (nm000138)BI2014a (nm000244)Arabic inner speech (ds005262)Kalunga2016 (nm000123)Music therapy (ds004840)
Columns are the 77 datasets, most electrodes first, in channel-count bands; rows are the 69 10-20 / 10-10 names that at least 10 datasets use, front to back. Cz is the most shared (65 datasets); the median dataset carries 32 of the 69, and 16 datasets use their own naming (fewer than half of their electrodes among the rows), so their columns show only the names that coincide. Hover a column for its k of n: k of the dataset's n distinct electrodes are among the rows.

The 77 datasets#

77 datasetsDownload CSV
DatasetNameStatusSourceArchiveParadigmSubjectsSessionsHoursChannelsRate (Hz)Size (GB)LicenceRedistributableListed in
ds004395Penn Electrophysiology of Encoding and Retrieval Study (PEERS)
PEERS
converted (v0)OpenNeuroOpenNeuromemory & cognition3646,4809,110.8125.75009,583.3CC0-1.0yespaper · card · v0
nm000181NMT: Neurodiagnostic Montage Template Scalp EEG
NMT
converted (v0)OtherNEMARclinical routine EEG2,4132,413489.021.020014.8CC-BY-SA-4.0yes (same licence)paper · v0
ds004706Spatial memory and non-invasive closed-loop stimulus timing
Spatial memory
converted (v0)OpenNeuroOpenNeuromemory & cognition34297470.6128.02,0481,426.1CC0-1.0yespaper · card · v0
ds005511Healthy Brain Network (HBN) EEG - Release 7
HBN R7
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)3813,098279.7128.6500262.9CC-BY-SA-4.0yes (same licence)paper · card · v0
ds005508Healthy Brain Network (HBN) EEG - Release 4
HBN R4
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)3243,342261.8129.0500246.8CC-BY-SA-4.0yes (same licence)paper · card · v0
ds005509Healthy Brain Network (HBN) EEG - Release 5
HBN R5
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)3303,324255.3129.0500240.7CC-BY-SA-4.0yes (same licence)paper · card · v0
ds005514Healthy Brain Network (HBN) EEG - Release 9
HBN R9
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)2952,881210.8129.0500198.7CC-BY-SA-4.0yes (same licence)paper · card · v0
nm000338Lee et al. 2019 (Motor Imagery) — EEG dataset and OpenBMI toolbox for
Lee2019-MI
converted (v0)MOABBNEMARBCI (MI & P300)54432183.162.01,00065.3CC0-1.0yespaper · card · v0
ds005512Healthy Brain Network (HBN) EEG - Release 8
HBN R8
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)2572,318179.1129.0500168.8CC-BY-SA-4.0yes (same licence)paper · card · v0
ds005507Healthy Brain Network (HBN) EEG - Release 3
HBN R3
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)1841,812158.8129.0500149.6CC-BY-SA-4.0yes (same licence)paper · card · v0
nm000238SparrKULee: A Speech-Evoked Auditory Response Repository from KU Leuven, Containing the EEG of 85 Participants
SparrKULee
converted (v0)OtherNEMARauditory & language80659155.764.01,024CC-BY-NC-4.0yes (non-commercial)paper · v0
nm000385Siena Scalp EEG Database: video-EEG monitoring of 14 adults with epilepsy (47 annotated seizures)
Siena
converted (v0)PhysioNetNEMARseizure1441141.030.451217.0CC-BY-4.0yes (credit)paper · v0
ds005506Healthy Brain Network (HBN) EEG - Release 2
HBN R2
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)1501,404127.5129.0500120.1CC-BY-SA-4.0yes (same licence)paper · card · v0
nm000250Dreyer et al. 2023 — A large EEG database with users' profile information for motor imagery brain-computer interface research
Dreyer2023
converted (v0)OtherNEMARBCI (MI & P300)871,040126.927.051217.0CC-BY-4.0yes (credit)paper · v0
ds005505Healthy Brain Network (HBN) EEG - Release 1
HBN R1
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)1361,342117.5129.0500110.7CC-BY-SA-4.0yes (same licence)paper · card · v0
nm000323Lee2019ERP
Lee2019ERP
converted (v0)MOABBNEMARBCI (MI & P300)54432116.262.01,000141.4CC0-1.0yespaper · v0
ds005510Healthy Brain Network (HBN) EEG - Release 6
HBN R6
converted (v0)OpenNeuroOpenNeurodevelopment (rest & tasks)1351,227103.5129.050097.5CC-BY-SA-4.0yes (same licence)paper · card · v0
nm000232THINGS-EEG2: A large and rich EEG dataset for modeling human visual object recognition
THINGS-EEG2
converted (v0)OtherNEMARperception & emotion1031987.363.01,000259.3CC-BY-4.0yes (credit)paper · v0
ds004148A test-retest resting and cognitive state EEG dataset
Test-retest rest
converted (v0)OpenNeuroOpenNeuroresting & consciousness6090075.061.050032.9CC0-1.0yespaper · card · v0
ds005697PerceiveImagine
Perceive/imagine
converted (v0)OpenNeuroOpenNeuroperception & emotion454569.562.01,00071.5CC0-1.0yespaper · card · v0
ds003969Meditation vs thinking task
Meditation vs think
converted (v0)OpenNeuroOpenNeuroresting & consciousness9839266.564.01,02458.5CC0-1.0yespaper · card · v0
ds005089Proactive selective attention across competition contexts
Selective attention
converted (v0)OpenNeuroOpenNeuroperception & emotion333362.863.01,00073.0CC0-1.0yespaper · card · v0
ds004279Large Spanish EEG
Spanish sentences
converted (v0)OpenNeuroOpenNeuroauditory & language566053.762.01,00027.1CC0-1.0yespaper · card · v0
ds004572The effects of sham hypnosis techniques
Hypnosis
converted (v0)OpenNeuroOpenNeuroresting & consciousness5251653.258.01,00046.8CC0-1.0yespaper · card · v0
ds005273Neural representation of consciously seen and unseen information
Seen vs unseen
converted (v0)OpenNeuroOpenNeuroperception & emotion323252.561.01,00047.7CC0-1.0yespaper · card · v0
ds003004Imagined Emotion Study
Imagined emotion
converted (v0)OpenNeuroOpenNeuroperception & emotion343448.5208.325638.7CC0-1.0yespaper · card · v0
ds003690EEG, ECG and pupil data from young and older adults: rest and auditory
Rest & auditory
converted (v0)OpenNeuroOpenNeuroresting & consciousness7537547.761.050023.0CC0-1.0yespaper · card · v0
ds003825Human electroencephalography recordings from 50 subjects for 22,248 im
THINGS-EEG1
converted (v0)OpenNeuroOpenNeuroperception & emotion505046.365.71,00044.3CC0-1.0yespaper · card · v0
ds004356Subcortical responses to music and speech are alike while cortical res
Music vs speech
converted (v0)OpenNeuroOpenNeuroauditory & language222446.332.010,000228.8CC0-1.0yespaper · card · v0
nm000273Lee2019SSVEP
Lee2019SSVEP
converted (v0)MOABBNEMARSSVEP5410840.662.01,00077.7CC0-1.0yespaper · v0
ds005121Siefert2024
Task not stated
converted (v0)OpenNeuroOpenNeuromemory & cognition343940.558.05129.7CC0-1.0yespaper · card · v0
ds004582FakeFaceEmo_data
Fake faces
converted (v0)OpenNeuroOpenNeuroperception & emotion737334.259.010,000315.9CC0-1.0yespaper · card · v0
ds005586Electroencephalographic responses to the number of objects in partiall
Numerosity
converted (v0)OpenNeuroOpenNeuroperception & emotion232333.560.01,00030.4CC0-1.0yespaper · card · v0
ds004603Visual Attribute-Specific Contextual Trajectory Paradigm
Context trajectory
converted (v0)OpenNeuroOpenNeuroperception & emotion373730.764.01,02429.4CC0-1.0yespaper · card · v0
nm000267Shin2017A
Shin2017A
converted (v0)MOABBNEMARBCI (MI & P300)2917429.030.02002.0GPL-3.0yes (GPL, same licence)paper · v0
nm000172Schirrmeister2017
Schirrmeister2017
converted (v0)MOABBNEMARBCI (MI & P300)142828.7128.050019.9CC-BY-4.0yes (credit)paper · v0
ds001787EEG meditation study
Meditation
converted (v0)OpenNeuroOpenNeuroresting & consciousness244027.664.02566.1CC0-1.0yespaper · card · v0
ds003885Capacity for movement is an organisational principle in object represe
Object motion 1
converted (v0)OpenNeuroOpenNeuroperception & emotion242427.1123.01,00049.6CC0-1.0yespaper · card · v0
nm000173Motor Imagery ataset from Ofner et al 2017
Ofner2017
converted (v0)MOABBNEMARBCI (MI & P300)1530027.161.05129.2CC-BY-4.0yes (credit)paper · card · v0
ds003887Capacity for movement is an organisational principle in object represe
Object motion 2
converted (v0)OpenNeuroOpenNeuroperception & emotion242426.8123.01,00049.1CC0-1.0yespaper · card · v0
ds002721An EEG dataset recorded during affective music listening
Affective music
converted (v0)OpenNeuroOpenNeuroperception & emotion3118526.319.01,0003.6CC0-1.0yespaper · card · v0
nm000217P300 dataset BI2015b from a "Brain Invaders" experiment
BI2015b
converted (v0)MOABBNEMARBCI (MI & P300)4417626.132.05124.6CC-BY-4.0yes (credit)paper · card · v0
ds004951Braille letters - EEG
Braille letters
converted (v0)OpenNeuroOpenNeurolearning & reward112325.863.01,00023.6CC0-1.0yespaper · card · v0
ds005262ArEEG: Arabic Inner Speech EEG dataset
Arabic inner speech
converted (v0)OpenNeuroOpenNeuroauditory & language1218625.08.02500.7CC0-1.0yespaper · card · v0
ds003846Prediction Error
Prediction error
converted (v0)OpenNeuroOpenNeurolearning & reward195022.763.050010.5CC0-1.0yespaper · card · v0
ds004315Mood Manipulation and PST, Experiment 1
Mood induction
converted (v0)OpenNeuroOpenNeuroperception & emotion505021.159.550010.5CC0-1.0yespaper · card · v0
ds002680Go-nogo categorization and detection task
Go/no-go
converted (v0)OpenNeuroOpenNeuroperception & emotion1435020.831.01,0009.9CC0-1.0yespaper · card · v0
ds004408EEG responses to continuous naturalistic speech
Natural speech
converted (v0)OpenNeuroOpenNeuroauditory & language1938020.6128.051220.1CC0-1.0yespaper · card · v0
nm000245Motor Imagery dataset from Cho et al 2017
Cho2017
converted (v0)MOABBNEMARBCI (MI & P300)525220.564.05127.2CC0-1.0yespaper · card · v0
ds005189Search Superiority Recollection Familiarity
Recollection
converted (v0)OpenNeuroOpenNeuromemory & cognition303020.162.01,00017.3CC0-1.0yespaper · card · v0
nm000266Sosulski2019
Sosulski2019
converted (v0)MOABBNEMARBCI (MI & P300)132,12019.631.01,0003.9CC-BY-SA-4.0yes (same licence)paper · v0
ds004324ToonFaces
Toon faces
converted (v0)OpenNeuroOpenNeuroperception & emotion262619.228.05002.6CC0-1.0yespaper · card · v0
ds004043The time-course of feature-based attention effects dissociated from te
Feature attention
converted (v0)OpenNeuroOpenNeuroperception & emotion202018.263.01,00016.6CC0-1.0yespaper · card · v0
ds004902A Resting-state EEG Dataset for Sleep Deprivation
Sleep deprivation
converted (v0)OpenNeuroOpenNeuroresting & consciousness7121818.061.05008.9CC0-1.0yespaper · card · v0
ds005620A repeated awakening study exploring the capacity of complexity measur
Anaesthesia
converted (v0)OpenNeuroOpenNeuroresting & consciousness2120217.962.05,00083.0CC0-1.0yespaper · card · v0
ds005095STERNBERG DIFFICULT
Sternberg
converted (v0)OpenNeuroOpenNeuromemory & cognition484816.963.01,00015.3CC0-1.0yespaper · card · v0
ds004477PES - Pandemic Emergency Scenario
Pandemic scenario
converted (v0)OpenNeuroOpenNeurolearning & reward9913.664.02,04824.0CC0-1.0yespaper · card · v0
nm000146Motor Imagery dataset from Weibo et al 2014
Weibo2014
converted (v0)MOABBNEMARBCI (MI & P300)101013.160.02001.7CC0-1.0yespaper · card · v0
nm000244P300 dataset BI2014a from a "Brain Invaders" experiment
BI2014a
converted (v0)MOABBNEMARBCI (MI & P300)646412.416.05121.1CC-BY-4.0yes (credit)paper · card · v0
ds005594Alphabetic Decision Task (Arial Light Font)
Alphabetic decision
converted (v0)OpenNeuroOpenNeuroperception & emotion161612.364.01,00011.7CC0-1.0yespaper · card · v0
ds004817EEG-attention-rsvp-exp2
RSVP attention 2
converted (v0)OpenNeuroOpenNeuroperception & emotion202011.963.01,00010.8CC0-1.0yespaper · card · v0
nm000216P300 dataset BI2015a from a "Brain Invaders" experiment
BI2015a
converted (v0)MOABBNEMARBCI (MI & P300)4312911.732.05122.1CC-BY-4.0yes (credit)paper · card · v0
ds004152Drum Trainer
Drum trainer
converted (v0)OpenNeuroOpenNeurolearning & reward212111.531.01,0005.1CC0-1.0yespaper · card · v0
ds004561Illusion of Agency over Electrically-Actuated Movements
Agency illusion
converted (v0)OpenNeuroOpenNeurolearning & reward232311.462.010,000104.9CC0-1.0yespaper · card · v0
ds004357Features-EEG
Features-EEG
converted (v0)OpenNeuroOpenNeuroperception & emotion161611.3123.01,00020.8CC0-1.0yespaper · card · v0
ds004840Dataset of electrophysiological signals (EEG, ECG, EMG) during Music t
Music therapy
converted (v0)OpenNeuroOpenNeuroauditory & language95110.98.01,0240.6CC0-1.0yespaper · card · v0
ds004816EEG-attention-rsvp-exp1
RSVP attention 1
converted (v0)OpenNeuroOpenNeuroperception & emotion202010.763.01,0009.7CC0-1.0yespaper · card · v0
ds003775SRM Resting-state EEG
SRM rest
converted (v0)OpenNeuroOpenNeuroresting & consciousness11015210.164.01,0244.8CC0-1.0yespaper · card · v0
ds004147Average Task Value
Task value
converted (v0)OpenNeuroOpenNeurolearning & reward12129.631.01,0004.3CC0-1.0yespaper · card · v0
ds004284eeg-neuroforecasting
Neuroforecasting
converted (v0)OpenNeuroOpenNeurolearning & reward18189.5129.01,00017.6CC0-1.0yespaper · card · v0
nm000158Liu, Lv et al. 2023 — EEG datasets of stroke patients (motor imagery)
Liu2024 (stroke)
converted (v0)MOABBNEMARBCI (MI & P300)591008.929.55000.7CC-BY-4.0yes (credit)paper · card · v0
ds004262Continuous Feedback Processing
Feedback
converted (v0)OpenNeuroOpenNeurolearning & reward21218.331.01,0003.7CC0-1.0yescard · v0
ds004264Steer the Ship
Steer the ship
converted (v0)OpenNeuroOpenNeurolearning & reward21217.531.01,0003.5CC0-1.0yespaper · card · v0
nm000123Kalunga2016
Kalunga2016
converted (v0)MOABBNEMARSSVEP12302.68.02560.1CC-BY-4.0yes (credit)paper · v0
ds002778UC San Diego Resting State EEG Data from Patients with Parkinson's Dis
Parkinson rest
converted (v0)OpenNeuroOpenNeuroresting & consciousness31462.532.05120.6CC0-1.0yespaper · card · v0
nm000215P300 dataset BI2014b from a "Brain Invaders" experiment
BI2014b
converted (v0)MOABBNEMARBCI (MI & P300)38382.432.05120.4CC-BY-4.0yes (credit)paper · card · v0
nm000138AlexMI
AlexMI
converted (v0)MOABBNEMARBCI (MI & P300)881.116.05120.1CC-BY-SA-4.0yes (same licence)paper · v0
The 77 datasets of eegdash-corpus-v0. Click a header to sort, hover a number for its source.