Discover datasets

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

…Applications include AR/VR text input, mobile computing, and hands-free communication…

Participants

100

Size

45.3 GB

Version

v2.0.0

Updated

Feb 23, 2026

EEG recordings from 54 healthy participants performing a 36-symbol ERP (P300) row-column speller task with face stimuli, collected over two sessions with offline training and online copy-spelling test phases. Data were recorded at 1000 Hz from 62 EEG channels plus 4 additional EMG channels. The dataset is part of a three-paradigm BCI collection (ERP, SSVEP, motor imagery) created to investigate BCI illiteracy and distributed with the OpenBMI toolbox.

Participants

54

Channels

62 (10-05)

Citations

100

HED

v8.4.0

Size

132 GB

Version

v1.0.5

Updated

Oct 8, 2026

[![DOI](https://img.shields.io/badge/DOI-10.82901%2Fnemar.on003688-blue…

Participants

63

Citations

10

Size

15.3 GB

Version

v1.0.0

Updated

Jun 22, 2026

…The mapping of inputs to electrode contacts is not part of the…

Participants

6

Size

996 MB

Version

v1.0.0

Updated

Oct 8, 2026

EEG recordings from 13 healthy participants performing a copy-spelling task with a visual 6x7 matrix ERP speller. The speller was designed for learning from label proportions (LLP), an unsupervised, calibrationless classification approach for ERP-based brain-computer interfaces. Two stimulus sequences with different target/non-target ratios allowed LLP to reconstruct mean target and non-target responses without labeled data. In the original online study, 84.5% of characters were spelled correctly on average without prior calibration.

Participants

13

Channels

31 (10-10)

Citations

46

HED

v8.4.0

Size

5.15 GB

Version

v1.0.3

Updated

Oct 8, 2026

…Performance gap vs traditional inputs ## Use Cases - **Continuous control**: Cursor/pointer movement…

Participants

100

Size

24.9 GB

Version

v2.0.0

Updated

Feb 23, 2026

EEG recordings from an online brain-computer interface study with twelve healthy participants using a visual matrix P300 speller. Three fully unsupervised decoding methods were compared: learning from label proportions (LLP), expectation-maximization (EM), and their mixture (MIX). Participants copy-spelled a German sentence across three blocks, one per algorithm. The dataset supports research on calibration-free ERP-based BCIs.

Participants

12

Channels

31 (10-10)

Citations

27

HED

v8.4.0

Size

4.81 GB

Version

v1.0.3

Updated

Oct 8, 2026

…rate, and `ANIN` (4 analog inputs at 24414.0625 Hz). The iEEG…

Participants

3

Size

9.30 GB

Version

v1.0.0

Updated

Oct 8, 2026

EEG recordings from 13 healthy participants performing an auditory oddball (P300) paradigm with target (1000 Hz) and non-target (500 Hz) tones at stimulus onset asynchronies between 60 and 600 ms. The study evaluated closed-loop Bayesian optimization and random search to find each individual's best stimulation speed for maximizing classification performance of a regularized LDA. The data comprise 31-channel EEG sampled at 1000 Hz, with an optimization part and a validation part.

Participants

13

Channels

31 (10-10)

Citations

2

HED

v8.4.0

Size

3.68 GB

Version

v1.0.7

Updated

Oct 9, 2026