Using a cVEP-Based Brain-Computer Interface to Control a Virtual Agent.

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Noninvasive brain-computer interfaces (BCIs) offer device control via brain activity but suffer from low information transfer rates (ITRs), limiting practical use.
• Codebook visually evoked potentials (cVEP) represent a promising BCI technology with superior ITRs compared to other methods.
• Prior research primarily focused on cVEP applications for spelling, leaving potential for broader use unexplored.

Purpose of the Study:

• To develop and evaluate the first cVEP-based BCI for real-world applications, including navigation and action selection.
• To demonstrate the feasibility of cVEPs in less restricted environments, specifically on dynamic backgrounds.
• To introduce and validate a novel dynamic repetition algorithm for optimizing individual BCI performance.

Main Methods:

• Development of an online cVEP-based BCI controlling a virtual agent in a simulated 3D kitchen environment.
• Investigation of cVEP reliability with stimuli presented on dynamic, changing backgrounds.
• Implementation of a novel dynamic repetition algorithm to individually balance accuracy and speed.

Main Results:

• Successfully triggered cVEPs with stimuli on dynamic backgrounds, demonstrating robustness.
• Achieved average and maximum information transfer rates (ITRs) of 50 bits/min and 68 bits/min, respectively, in a 12-command cVEP-BCI.
• Validated the effectiveness of the dynamic repetition algorithm for personalized BCI optimization.

Conclusions:

• cVEP-BCIs are a fast and robust approach suitable for real-world applications, moving beyond laboratory settings.
• The developed cVEP-BCI system demonstrates potential for enabling everyday tasks through brain control.
• This research advances the practical utility of cVEP technology for broader BCI deployment.