Discover datasets

This dataset contains 62-channel EEG recordings from 25 healthy, right-handed subjects performing kinesthetic motor imagery of the right hand versus right elbow, a within-limb discrimination paradigm. Each subject completed 15 sessions over 3 days, yielding 600 trials total, designed to study brain-computer interface (BCI) decoding of different joint movements from the same limb. The dataset supports research into motor rehabilitation and prosthetic control applications using EEG-based BCI systems.

Participants

25

Channels

62 (10-10)

Citations

52

HED

v8.4.0

Size

53.8 GB

Version

v1.0.2

Updated

Aug 18, 2026

Participants

12

Channels

32 (10-10)

Citations

21

HED

v8.4.0

Size

158 MB

Version

v1.0.0

Updated

Jul 10, 2026

This dataset comprises preprocessed electroencephalography (EEG) recordings from 20 healthy participants performing a motor imagery task involving discrete reaching movements in four directions (up, down, left, right) with varying speeds and distances. Participants executed 960 trials across 10 blocks while viewing visual cues, with concurrent eye-tracking and motion capture data. The dataset includes extensive preprocessing with artifact correction, source localization, and classification features, making it suitable for brain-computer interface research and motor imagery decoding studies.

Participants

20

Channels

67 (10-05)

Citations

3

HED

v8.4.0

Size

15.4 GB

Version

v1.0.2

Updated

Aug 18, 2026

This dataset comprises EEG recordings from 20 healthy participants performing motor imagery of handwritten letter production. Participants imagined writing ten different letters (a, d, e, f, j, n, o, s, t, v) using their right index finger in response to visual cues. The study investigates brain-computer interface applications through direct EEG classification and continuous kinematic decoding approaches, achieving classification accuracies of 26.2% for ten-letter and 46.7% for five-letter tasks.

Participants

20

Channels

60 (10-10)

Citations

12

HED

v8.4.0

Size

27.0 GB

Version

v1.0.2

Updated

Aug 18, 2026

This dataset comprises electroencephalography (EEG) recordings from 50 acute stroke patients (1-30 days post-stroke) performing motor imagery tasks of left- and right-handed hand-grip movements. Recorded using a wireless 29-channel EEG system at 500 Hz, the dataset includes raw and preprocessed data, representing the first open resource addressing left- and right-handed motor imagery in the acute stroke population. The dataset supports brain-computer interface (BCI) algorithm development and clinical rehabilitation applications.

Participants

50

Channels

29 (10-10)

Citations

78

HED

v8.4.0

Size

1.15 GB

Version

v1.0.3

Updated

Aug 18, 2026

This dataset comprises EEG recordings from 12 healthy participants performing upper-limb motor imagery tasks centered on elbow movements. Participants executed kinesthetic imagery of nine goal-directed tasks (drawer opening, soup preparation, weight lifting, door opening, plate cleaning, combing, pizza cutting, and pick-and-place operations) plus rest, cued by visual stimuli. The dataset contains 330 trials recorded at 1000 Hz using 17 EEG channels and is designed for brain-computer interface (BCI) research and motor imagery classification studies.

Participants

12

Channels

17 (other)

Citations

13

HED

v8.4.0

Size

4.18 GB

Version

v1.0.3

Updated

Aug 18, 2026

This dataset comprises EEG recordings from 10 patients with spinal cord injury performing five distinct hand and forearm motor imagery tasks: hand open, palmar grasp, lateral grasp, pronation, and supination. The study investigates the decoding of attempted movements from low-frequency EEG signals (movement-related cortical potentials) to support neuroprosthetic applications for upper limb control. Data include 360 trials (72 per class) acquired at 256 Hz with 61 EEG channels and 3 EOG channels, preprocessed with ICA-based artifact rejection and bandpass filtering.

Participants

10

Channels

61 (10-05)

Citations

97

HED

v8.4.0

Size

4.51 GB

Version

v1.0.2

Updated

Aug 18, 2026

The Munich Motor Imagery dataset comprises electroencephalographic recordings from 10 healthy subjects performing two-class motor imagery tasks (left and right hand) cued by visual arrow stimuli. The dataset includes 128-channel EEG data sampled at 500 Hz with preprocessed signals suitable for brain-computer interface research and benchmarking of spatial filtering and classification methods.

Participants

10

Channels

128 (other)

Citations

74

HED

v8.4.0

Size

5.49 GB

Version

v1.0.2

Updated

Aug 18, 2026

The BNCI2003_IVa Motor Imagery dataset comprises EEG recordings from 5 healthy subjects performing motor imagery tasks (right hand and feet movements) in response to visual cues. This preprocessed dataset, originally from BCI Competition III, contains 280 trials per subject recorded with 118 EEG channels at 100 Hz sampling rate. The dataset has been extensively used for benchmarking brain-computer interface classification algorithms, particularly for evaluating common spatial patterns and feature combination methods.

Participants

5

Channels

118 (10-05)

Citations

86

HED

v8.4.0

Size

735 MB

Version

v1.0.3

Updated

Aug 18, 2026

This dataset contains simultaneous scalp and ear-EEG recordings from 6 healthy right-handed adults performing a motor execution task involving left and right hand fist clenching, cued by visual arrow stimuli. Recordings were made at 1000 Hz using a 122-channel Neuroscan SynAmps2 system, yielding 1114 trials, and are intended for benchmarking motor task classification algorithms such as EEGNet. The dataset was converted to BIDS format using MOABB and originates from a study investigating the utility of in-ear EEG sensing for motor task classification.

Participants

6

Channels

122 (10-05)

Citations

12

HED

v8.4.0

Size

18.4 GB

Version

v1.0.3

Updated

Aug 20, 2026