Discover datasets

EEG recordings from 15 healthy participants performing a visual P300 oddball task while walking on a treadmill at 1.6 m/s, released as BCI Competition 2020 Track 5. The dataset combines 46-channel scalp EEG, 14-channel ear-EEG, 4 EOG channels, and 6 inertial measurement unit channels sampled at 100 Hz. It is intended for evaluating brain-computer interface performance under ambulatory conditions, with trials divided temporally into training, validation, and test sets.

Participants

15

Channels

46 (10-10)

HED

v8.4.0

Size

224 MB

Version

v1.0.4

Updated

Oct 8, 2026

An open-access EEG dataset for inner speech recognition recorded from ten participants with a 136-channel acquisition system. It covers three paradigms (inner speech, pronounced speech, and visualized condition) with four command classes (Up, Down, Right, Left) across 5640 trials. It is intended to support brain-computer interface development and the study of the neural mechanisms of the inner voice.

Participants

10

Citations

78

Size

24.6 GB

Version

v1.0.0

Updated

Jun 22, 2026

Alljoined1 is an EEG dataset of neural responses to natural images presented in a rapid serial visual presentation paradigm, built for EEG-to-image decoding research. Eight healthy adults viewed images from the Natural Scenes Dataset (MS-COCO) while 64-channel BioSemi EEG was recorded across one or two sessions. This BIDS version contains raw BDF recordings of 960 shared images, event tables linked to stimuli, and epoched derivatives from the original authors.

Participants

8

Citations

5

Size

8.04 GB

Version

v1.0.4

Updated

Sep 23, 2026

A sample intracranial EEG (iEEG) dataset prepared for the WIRED ICM course on intracranial recordings in humans, epilepsy, and deep brain stimulation (Paris, March 2024). It contains recordings from patients with epilepsy during naturalistic listening tasks (movie trailers and TIMIT sentences). It also includes illustrative examples of seizure-related activity and vagus nerve stimulator artifact.

Participants

3

Size

348 MB

Version

v1.0.0

Updated

Jun 26, 2026

High-density (256-channel) electrocorticography recorded at UCSF from patients undergoing clinical treatment for epilepsy while they read aloud consonant-vowel syllables from a list. The dataset comprises 31 recording blocks across four subject codes, with hand-marked annotations of syllable identity, onset, consonant-vowel transition and offset, plus rest baseline periods and anatomical electrode labels. It supports study of the functional organization of human ventral sensorimotor cortex during speech articulation and is organized as iEEG-BIDS with source NWB files retained unchanged.

Participants

4

Size

104 GB

Version

v1.0.0

Updated

Oct 8, 2026

EEG recordings from twenty healthy right-handed adults exposed to tendon vibration of the left elbow, designed to separate the percept of illusory movement from plain vibratory sensation. Sixteen sensorimotor electrodes were recorded while biceps and triceps tendons were vibrated at 90/50 Hz (illusion of elbow extension) or both at 70 Hz (control, no illusion). Each participant completed three blocks of 72 randomized three-second stimulations.

Participants

20

Channels

16 (10-10)

Citations

16

Size

664 MB

Version

v1.0.2

Updated

Oct 8, 2026

Motor-imagery EEG recorded from 12 healthy, BCI-naive participants performing cued left- versus right-hand imagery to control a Baxter robot arm reaching toward objects. Each participant completed three sessions with first-person and third-person visual feedback conditions, giving 480 trials per subject. Data were recorded with a 32-channel LiveAmp system at 250 Hz and re-released in BIDS format via MOABB.

Participants

12

Channels

32 (10-10)

HED

v8.4.0

Size

5.69 GB

Version

v1.0.0

Updated

Oct 8, 2026

EEG recordings from 30 healthy participants performing a left-hand grasping motor imagery task versus rest, cued by auditory stimuli, with simultaneous pupillometry. The dataset was recorded with a 32-channel BioSemi ActiveTwo system at 512 Hz across two runs of 25 trials per subject. It supports evaluating whether task-evoked pupil diameter changes can improve EEG-based motor imagery brain-computer interface classification.

Participants

30

Channels

32 (10-10)

Citations

36

HED

v8.4.0

Size

3.22 GB

Version

v1.0.2

Updated

Aug 18, 2026

The EEG Moments Dataset (EMD) contains EEG responses from 6 human subjects viewing 1,102 naturalistic 3-second videos presented with their audio tracks. Each subject viewed the 1,000-video training set 6 times and the 102-video testing set 24 times. Eye tracking data (gaze and pupil size) were recorded during the experiment. The videos carry human annotations for objects, scenes, actions, sentence descriptions, spoken transcriptions, and memorability. EMD is the EEG companion to the BOLD Moments fMRI dataset.

Participants

6

Citations

1

Size

210 GB

Version

v1.0.0

Updated

Sep 8, 2026

This dataset contains EEG recordings collected to study how expectation affects repetition suppression in nociception. It supports investigation of how the brain's responses to repeated painful stimuli are modulated by predictions about those stimuli. The data are organized in BIDS format.

Participants

36

Channels

32 (10-10)

Size

42.7 GB

Version

v1.0.0

Updated

Jun 28, 2026