Discover datasets

A longitudinal MEG dataset investigating brain activity during early skill learning, overnight consolidation, and longer-term retention in 31 healthy right-handed adults. Participants performed a sequential finger tapping task with the non-dominant left hand across up to three MEG sessions (initial training, about 24 hours later, and about 30 days later), with resting-state scans before and after practice. Anatomical MRI, behavioral responses, eye tracking and pupillometry, and forearm EMG were also recorded and synchronized with MEG.

Participants

31

Size

95.8 GB

Version

v1.0.0

Updated

Jun 29, 2026

Beetl2021-B is a preprocessed 32-channel EEG dataset from the NeurIPS 2021 BEETL competition, built to benchmark transfer learning for motor imagery decoding. It contains recordings from 2 healthy subjects performing four-class motor imagery (left hand, right hand, feet, rest) at 200 Hz, totaling 1590 trials. The dataset supports evaluation of cross-subject and cross-dataset generalization and domain adaptation in brain-computer interfaces.

Participants

2

Channels

32 (10-10)

Citations

1

HED

v8.4.0

Size

1.99 GB

Version

v1.0.5

Updated

Aug 20, 2026

An EEG dataset from the NeuroTechs group, collected in the context of STEM (science, technology, engineering and mathematics) skills. Recordings were acquired with a Unicorn wearable EEG system through its Python API, and values are provided in microvolts. The dataset is organized according to BIDS with HED annotations.

Participants

63

Channels

8 (10-10)

Size

1.39 GB

Version

v1.0.0

Updated

Jun 29, 2026

NOD-EEG is an electroencephalography dataset of human responses to naturalistic ImageNet object images. It extends the Natural Object Dataset-fMRI (NOD-fMRI) by recording EEG, and MEG in a companion dataset, from the same subjects viewing the same stimuli. Together the three modalities support study of object recognition with both high spatial and high temporal resolution. The dataset includes raw BIDS data, preprocessed continuous data, epoched data, and detailed trial-level metadata.

Participants

19

Channels

62 (10-10)

Citations

5

Size

24.3 GB

Version

v1.0.0

Updated

Jun 28, 2026

This dataset contains EEG recordings from 34 participants (one session each) who viewed 242 object images, 121 challenge and 121 control images, selected from comparisons between human behavioural performance and AlexNet. The study examines adaptive recruitment of cortex-wide recurrent processing during visual object recognition. The derivatives folder contains outputs of the decoding analyses, including time-resolved decoding accuracy matrices for object identity (.npz).

Participants

34

Channels

63 (10-10)

Citations

2

Size

61.2 GB

Version

v1.0.0

Updated

Jun 30, 2026

Alljoined-1.6M is a large-scale EEG dataset of neural responses to rapid serial visual presentation of natural images from the THINGS database. It was recorded with a consumer-grade 32-channel EMOTIV FLEX2 system in 20 adult participants over four sessions each, yielding more than 1.6 million trials. The dataset was built to test whether deep-learning-based brain-computer interface and semantic decoding research can be done with affordable EEG hardware.

Participants

20

Channels

32 (10-10)

Size

8.78 GB

Version

v1.0.3

Updated

Sep 23, 2026

Raw intracranial EEG (iEEG) recorded from 8 neurosurgical patients with electrodes implanted for seizure monitoring while they passively viewed rapid streams of scene photographs. Streams contained statistical structure at the category level or the exemplar level, alongside a random baseline. Neural entrainment (phase coherence) at the image frequency (2 Hz) and the pair frequency (1 Hz) was used to track online statistical learning across levels of abstraction. The data are converted to iEEG-BIDS without signal modification.

Participants

8

Size

17.5 GB

Version

v1.0.0

Updated

Oct 8, 2026

EEG recordings from 20 healthy participants performing a tactile detection task with confidence ratings, plus ECoG from one patient with intractable epilepsy. The data were collected to study how evidence accumulation relates to perceptual consciousness and metacognitive monitoring. Near-threshold vibrotactile stimuli were delivered to the right wrist, with detection and confidence reports collected over 500 trials per participant.

Participants

19

Channels

69 (10-10)

Size

27.9 GB

Version

v1.0.0

Updated

Jun 20, 2026

EEG dataset collected during an architectural affordances task, in which participants experienced architectural spaces while brain activity was recorded. The dataset is shared in raw form, follows an unofficial BIDS extension, and is mirrored on NEMAR from OpenNeuro (ds004980).

Participants

17

Channels

64 (other)

Size

15.8 GB

Version

v1.0.0

Updated

Jun 26, 2026

Beetl2021-A is a preprocessed four-class motor imagery EEG dataset (rest, left hand, right hand, feet) from Task 2 of the BEETL Competition at NeurIPS 2021. It contains 63-channel recordings at 500 Hz from healthy participants playing an online BCI racing game (Cybathlon2020IC), with 1490 trials of 4 s each in a single session. It serves as a benchmark for transfer learning and domain adaptation across heterogeneous EEG datasets and subjects, and is re-hosted in BIDS format via the MOABB pipeline.

Participants

3

Channels

63 (10-10)

HED

v8.4.0

Size

2.17 GB

Version

v1.0.6

Updated

Aug 20, 2026