moabb.datasets.BNCI2015_007#

class moabb.datasets.BNCI2015_007(subjects=None, sessions=None)[source]#

BNCI 2015-007 Motion VEP (mVEP) Speller dataset.

Dataset summary

#Subj	16
#Chan	63
#Trials / class	varies NT / T
Trials length	0.7 s
Freq	100 Hz
#Sessions	1

Participants

Population: Healthy
Age: 23.8 (range: 21-30) years
Handedness: normal or corrected-to-normal vision
BCI experience: naive

Equipment

Amplifier: BrainAmp EEG amplifier
Electrodes: active electrode
Montage: 10-10
Reference: linked mastoids

Preprocessing

Data state: filtered
Bandpass filter: 0.016-250 Hz
Steps: downsampling, low-pass filter, baseline correction, artifact rejection
Re-reference: linked mastoids
Notes: For offline analysis: downsampled to 200 Hz, low-pass filtered (42 Hz passband, 49 Hz stopband). For online: downsampled to 100 Hz. Artifact rejection: min-max ≥70 μV. Nontarget epochs filtered to avoid overlap with targets (3 preceding and 4 following stimuli must be nontargets).

Data Access

DOI: 10.1088/1741-2560/9/4/045006
Repository: BNCI Horizon

Experimental Protocol

Paradigm: p300
Task type: visual_speller
Feedback: visual
Stimulus: motion VEP (mVEP)

Found an issue with this dataset?

If you encounter any problems with this dataset (missing files, incorrect metadata, loading errors, etc.), please let us know!

Report an Issue on GitHub

Dataset summary

Name	#Subj	#Chan	#Trials/class	Trials length	Sampling Rate	#Sessions
BNCI2015_007	16	63	~1800 NT / ~360 T	0.7s	100Hz	1

Dataset from [1].

Dataset Description

This dataset implements a motion-onset visual evoked potential (mVEP) based brain-computer interface for gaze-independent spelling. Unlike conventional flash-based P300 spellers that use luminance changes, this paradigm uses motion onset (moving bar stimuli) to elicit visual evoked potentials, specifically the N200 component. This approach has advantages including lower visual fatigue, reduced luminance and contrast requirements, and potential for use in bright environments.

The motion VEP (mVEP) speller operates by presenting moving bar stimuli at different positions in a matrix layout. When the user attends to a target position, the motion onset at that location elicits a characteristic N200 response that can be detected to determine the user’s intended selection.

Participants

16 healthy subjects
Gender: Not specified in metadata
Age: Not specified in metadata
BCI experience: Not specified
Health status: Healthy volunteers
Location: Neurotechnology Group, Technische Universitat Berlin, Germany

Recording Details

Equipment: BrainProducts actiCap active electrode system
Channels: 63 EEG electrodes (standard 10-10 system)
Sampling rate: 100 Hz (downsampled from original recording)
Reference: Nose reference
Montage: standard_1005
Filters: Bandpass filtered during preprocessing
Units: uV (converted to V during loading)

Experimental Procedure

6x6 matrix speller layout (36 possible targets)
Motion onset stimulation (moving bars)
6 stimulus positions per row/column
Overt attention paradigm (gaze-dependent) and covert attention modes
One recording session per subject with multiple runs (typically 2)
Each run contains multiple spelling sequences

Data Organization

Subject codes: fat, gdf, gdg, iac, iba, ibe, ibq, ibs, ibt, ibu, ibv, ibw, ibx, iby, ice, icv
Data URL: http://doc.ml.tu-berlin.de/bbci/BNCIHorizon2020-MVEP/

Event Codes

Target (1): Target stimulus presented (attended)
NonTarget (2): Non-target stimulus presented (not attended)

References

[1]

Treder, M. S., Purwins, H., Miklody, D., Sturm, I., & Blankertz, B. (2012). Decoding auditory attention to instruments in polyphonic music using single-trial EEG classification. Journal of Neural Engineering, 11(2), 026009. https://doi.org/10.1088/1741-2560/11/2/026009

Notes

Added in version 1.2.0.