Discover datasets

This dataset contains preprocessed EEG recordings from 6 healthy participants performing imagined speech tasks with three short word conditions (out, in, up). Data were acquired at 256 Hz using 64 channels and analyzed using Riemannian manifold methods and relevance vector machines for brain-computer interface applications. The motor imagery paradigm employed auditory and visual cueing, yielding 5,400 trials suitable for BCI research and benchmarking.

Participants

6

Channels

60 (10-10)

Citations

98

HED

v8.4.0

Size

1.90 GB

Version

v1.0.3

Updated

Aug 18, 2026

Participants

2

Size

1.87 GB

Version

v1.0.0

Updated

Jul 10, 2026

…After collecting sufficient training data, classifier models predict recall in closed-loop…

Participants

34

Channels

128 (biosemi)

Citations

5

Size

1.30 TB

Version

v1.0.0

Updated

Jul 10, 2026

Imported from OpenNeuro ds005262

Participants

12

Size

689 MB

Version

v1.0.0

Updated

Jul 10, 2026

…The primary goal is to advance our understanding of language processing in…

Participants

52

Channels

64

Citations

3

Size

109 GB

Version

v1.0.0

Updated

Aug 19, 2026

This dataset comprises preprocessed EEG recordings from 6 healthy participants performing imagined speech tasks, specifically discriminating between short and long words ('cooperate' vs 'in'). Data were acquired at 256 Hz using 64 EEG channels and processed with bandpass filtering (8-70 Hz), notch filtering (60 Hz), and artifact removal. The dataset contains 1,200 trials. Classification results (mean accuracy 73.3±8.9%) reported in the original Nguyen et al. 2017 study are provided for reference.

Participants

6

Channels

60 (10-10)

Citations

98

HED

v8.4.0

Size

1.15 GB

Version

v1.0.3

Updated

Aug 18, 2026

…or by more advanced transformer models of vision. We propose that the…

Participants

21

Citations

13

Size

71.1 GB

Version

v1.0.0

Updated

Aug 19, 2026

This dataset comprises EEG recordings from 6 healthy participants performing imagined speech tasks involving two conditions: cooperate and independent word imagery. The study employed a motor imagery paradigm with auditory and visual cues, yielding 3,600 preprocessed trials (1,800 per class) recorded at 256 Hz from 64 channels. Data were analyzed using Riemannian manifold methods and relevance vector machines for brain-computer interface applications, achieving mean classification accuracy of 66.2±4.8%.

Participants

6

Channels

60 (10-10)

Citations

97

HED

v8.4.0

Size

1.27 GB

Version

v1.0.3

Updated

Aug 18, 2026

[![DOI](https://img.shields.io/badge/DOI-10.82901%2Fnemar.on007338-blue…

Participants

1

Channels

129 (egi-geodesic)

Citations

8

Size

45.0 MB

Version

v1.0.0

Updated

Aug 18, 2026

MEG-SCANS is a magnetoencephalography (MEG) dataset comprising raw and MaxFiltered recordings from 24 German-speaking participants performing an auditory speech experiment. Stimuli included audiobooks, Oldenburger Matrix Sentence Test sentences at varying intelligibility levels for speech reception threshold assessment, and chirp stimuli for signal quality checks. The dataset also includes structural MRI scans, hearing screening results, empty-room recordings, and audio stimulus materials, organized in BIDS format to support encoding/decoding analyses of speech-evoked brain responses.

Participants

24

Citations

1

Size

204 GB

Version

v1.0.0

Updated

Aug 19, 2026