Related Experiment Video
Updated: Jul 11, 2025
06:09
P300-Based Brain-Computer Interface Speller Performance Estimation with Classifier-Based Latency Estimation
Published on: September 8, 2023
609
Burst c-VEP Based BCI: Optimizing stimulus design for enhanced classification with minimal calibration data and
Kalou Cabrera Castillos1, Simon Ladouce1, Ludovic Darmet1
1Human Factors and Neuroergonomics, Institut Supérieur de l'Aéronautique et de l'Espace, 10 Av. Edouard Belin, Toulouse, 31400, France.
Neuroimage
|November 10, 2023
Summary
This study introduces Burst c-VEP, a new method for brain-computer interfaces that uses slow, flashing visual stimuli. Burst c-VEP improves accuracy and user comfort, reducing calibration time for reactive BCI applications.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Human-Computer Interaction
Background:
- Code-modulated Visual Evoked Potentials (c-VEP) are key for reactive Brain-Computer Interfaces (rBCI).
- Existing c-VEP designs require improvement for user experience, signal quality, and reduced calibration time.
- Aperiodic flickering visual stimuli are utilized in c-VEP for rBCI.
Purpose of the Study:
- To introduce and evaluate an innovative variant of c-VEP, termed
- Burst c-VEP
- to enhance performance and usability.
- To investigate the potential of burst c-VEP to achieve high decoding performance with reduced calibration data.
- To improve user visual comfort by reducing stimulus saliency.
Main Methods:
- Developed and tested
- Burst c-VEP
- stimuli, presenting short bursts of aperiodic flashes at 2-4 Hz.
- Utilized convolutional neural networks (CNNs) for decoding.
- Conducted an offline 4-class c-VEP protocol with 12 participants, manipulating stimulus patterns (burst vs. m-sequence) and amplitude modulations (100% vs. 40%).
Main Results:
- Full amplitude burst c-VEP sequences achieved higher accuracy (90.5%-95.6%) compared to m-sequences (71.4%-85.0%).
- Reduced stimulus intensity to 40% slightly decreased burst c-VEP accuracy (to 94.2%) but significantly improved user experience.
- Mean selection time was 1.5s for both codes, comparable to previous studies.
Conclusions:
- Burst c-VEP demonstrates significant potential for advancing reactive BCI systems in terms of both performance and usability.
- The proposed method offers a promising approach for faster selection and reduced calibration requirements.
- Open-access data and CNN implementation are provided for further research.
