Discover datasets

EEG dataset recorded during rapid serial visual presentation (RSVP) of sentences with flanking words, in which the critical center word was shown between grey flanker words. Sentences contained subject-verb agreement violations, semantic violations, double violations, word-order violations, or no violation (controls). The dataset supports research on sentence processing and parafoveal word processing during reading.

Participants

64

Channels

32 (10-10)

Size

6.22 GB

Version

v1.0.0

Updated

Sep 8, 2026

EEG dataset from the PURSUE project containing an N400 word processing task, recorded from participants at three primarily undergraduate institutions (Southern California, Massachusetts, and Virginia) in 2017 and 2018. The task design follows the ERP CORE paradigm. The dataset supports event-related potential research on semantic word processing.

Participants

288

Channels

32 (10-10)

Size

8.98 GB

Version

v1.0.0

Updated

Jun 3, 2026

This dataset contains raw EEG recordings, in BrainVision format and organized according to BIDS, from participants silently reading word-by-word naturalistic narratives. It accompanies a study of auditory representations of words evoked during silent visual reading. Derivatives include preprocessed word-level EEG epochs, story and run metadata, expected/unexpected epoch labels, and corresponding feature embeddings.

Participants

10

Channels

63 (10-10)

Size

35.3 GB

Version

v1.0.0

Updated

Jun 29, 2026

Raw 64-channel EEG, ECG, PPG and behavioral recordings from 30 healthy young adults (18–32 years) during resting state and a verbal working memory digit-span task with serial recall. The resting session alternated eyes-closed and eyes-open blocks followed by passive cartoon viewing. In the working memory task, seven-digit sequences were presented visually either simultaneously or sequentially at different rates and with different inter-stimulus intervals, allowing study of how presentation mode affects neural oscillations and peripheral physiology.

Participants

30

Channels

63 (10-10)

Citations

3

Size

41.4 GB

Version

v1.0.0

Updated

Jun 29, 2026

High-density (127-channel) EEG recordings from 30 healthy, right-handed, native Chinese-speaking adults performing imagined and executed vocalization and writing tasks. Stimuli were four Chinese strokes (heng, shu, pie, na) and four English letters (a, b, c, d). The factorial design crossed task mode, action, script and token identity, yielding 32 event-coded conditions and 15,360 task events. The release also includes EEGLAB preprocessing derivatives and subject-level machine-learning exports.

Participants

30

Channels

124 (10-05)

Size

181 GB

Version

v1.0.0

Updated

Sep 8, 2026

This dataset pools EOG recordings from two earlier studies (20 participants from Jaeger et al., 2020 and 21 from Holtze et al., 2021) that used a competing-speaker paradigm. Participants attended to either the left or the right of two simultaneously presented audio books. The dataset supports analysis of the relation between eye blinks and speech pauses.

Participants

41

Channels

7 (egi-geodesic)

Citations

1

Size

8.01 GB

Version

v1.0.0

Updated

Jun 24, 2026

The Brain Treebank is a large-scale dataset of intracranial (stereo-EEG) recordings from 10 epilepsy patients who watched Hollywood movies. Signals were sampled at 2048 Hz from 1,688 electrodes in total, across 26 movie trials and about 43 hours of data. The movie audio was transcribed with manually annotated word onsets, and each transcript was parsed into Universal Dependencies syntax trees. This makes the dataset a resource for studying the neural basis of naturalistic language. This record provides the data in BIDS-iEEG format.

Participants

10

Size

257 GB

Version

v1.1.0

Updated

Sep 21, 2026

MULTI-CLARID is a combined fMRI/EEG dataset acquired during an auditory category learning task and resting state. It includes 63-channel EEG with ECG, EOG, facial EMG and skin conductance, plus respiration and finger-pulse recordings, and T1-weighted and UTE anatomical MR scans. Five participants' datasets containing the multi-dimensional category learning experiment are currently shared.

Participants

34

Channels

72 (10-10)

Size

6.43 GB

Version

v1.0.0

Updated

Jun 27, 2026

EEG dataset from 19 participants performing a visual P300 speller-like paradigm to assess how visual stimulus properties affect evoked responses and classification accuracy. Each participant completed three sessions with conditions varying the number of stimuli (4 to 9) and the flash type (standard flash versus a superimposed yellow smiling cartoon face). Recordings comprise 382 files from eight electrodes sampled at 256 Hz, with markers encoding trial phases and activated symbols.

Participants

19

Channels

9 (10-10)

Citations

1

Size

1.27 GB

Version

v1.0.0

Updated

May 13, 2026

EEG and behavioral dataset from 39 Italian-speaking university students performing a picture-matching task in which target pictures matched or mismatched a preceding cue. Cues were physical (picture), imagery (single word), literal sentence, or metaphorical sentence. The data are provided as preprocessed long epochs (filtered, mastoid re-referenced, ocular ICA cleaned). The dataset was collected within the ERC PROMENADE project to test how mental representations generated by different cue types influence target picture processing. Vocabulary and mental imagery measures are included per participant.

Participants

39

Channels

58 (10-10)

Size

14.8 GB

Version

v1.0.0

Updated

Jun 28, 2026