moabb.datasets.BNCI2024_001#

class moabb.datasets.BNCI2024_001[source]#

BNCI 2024-001 Handwritten Character Classification dataset.

Dataset summary

#Subj

20

#Chan

64

#Classes

10

#Trials / class

varies

Trials length

3 s

Freq

512 Hz

#Sessions

1

#Runs

1

Total_trials

varies

Participants

  • Population: healthy

Equipment

  • Amplifier: BrainVision

  • Electrodes: active electrodes

  • Montage: eogl1 eogl2 eogl3 eogr1 af7 af3 afz af4 af8 f7 f5 f3 f1 fz f2 f4 f6 f8 ft7 fc5 fc3 fc1 fcz fc2 fc4 fc6 ft8 t7 c5 c3 c1 cz c2 c4 c6 t8 tp7 cp5 cp3 cp1 cpz cp2 cp4 cp6 tp8 p7 p5 p3 p1 pz p2 p4 p6 p8 ppo1h ppo2h po7 po3 poz po4 po8 o1 oz o2

  • Reference: right mastoid

Preprocessing

  • Data state: raw

  • Bandpass filter: 0.3-70 Hz

  • Steps: resampling, notch filtering, high-pass filtering, bad channel interpolation, EOG derivative computation, low-pass filtering of EOG, epoching, visual artifact rejection

  • Re-reference: car

Data Access

Experimental Protocol

  • Paradigm: imagery

  • Feedback: none during main paradigm; training included visual guidance

  • Stimulus: letter cue

Dataset from [1].

Dataset Description

This dataset contains EEG data from 20 healthy subjects performing handwritten character (letter) writing tasks. Participants wrote 10 different letters (a, d, e, f, j, n, o, s, t, v) while EEG was recorded. The study investigates the classification of handwritten characters from non-invasive EEG through continuous kinematic decoding.

Participants

  • 20 healthy subjects

  • Location: Institute of Neural Engineering, Graz University of Technology, Austria

Recording Details

  • Equipment: BrainVision EEG system with 60 EEG + 4 EOG channels

  • Channels: 60 EEG electrodes + 4 EOG electrodes = 64 total

  • Electrode montage: Extended 10-20 system

  • Sampling rate: 500 Hz

Experimental Procedure

  • 10 letter classes: a, d, e, f, j, n, o, s, t, v

  • Participants wrote letters inside a box while fixating on the screen

  • No visual feedback of the writing was provided during the task

  • 2 experimental rounds per subject, each containing ~32 trials per letter

  • Additional motion capture data was recorded (pen position)

Event Codes

The events correspond to the 10 different letters written by participants:

  • letter_a (1): Letter ‘a’

  • letter_d (2): Letter ‘d’

  • letter_e (3): Letter ‘e’

  • letter_f (4): Letter ‘f’

  • letter_j (5): Letter ‘j’

  • letter_n (6): Letter ‘n’

  • letter_o (7): Letter ‘o’

  • letter_s (8): Letter ‘s’

  • letter_t (9): Letter ‘t’

  • letter_v (10): Letter ‘v’

References

[1]

Crell, M. R., & Muller-Putz, G. R. (2024). Handwritten character classification from EEG through continuous kinematic decoding. Computers in Biology and Medicine, 182, 109132. https://doi.org/10.1016/j.compbiomed.2024.109132

Notes

Added in version 1.3.0.

This dataset is notable for exploring non-invasive EEG-based handwritten character classification, which could enable communication for individuals with limited movement capacity. The study demonstrated that handwritten characters can be classified from non-invasive EEG and that decoding movement kinematics prior to classification improves performance.

On this page