Stimulus characteristics of c-VEP BCI
A brain-computer interface (BCI) can use a multitude of control signals that are decodable from measured EEG. One of the control signals that leads to accurate and fast BCI performance for instance for communication and control, is the code-modulated visual evoked potential (c-VEP). It is the response to a pseudo-random sequence of flashes. Flashes are typically represented as a black background changing to a white one on a screen. The pseudo-random sequences are often borrowed from the telecomunication domain, because these tend to have appropriate properties such as low auto- or cross-correlation between different sequences (assumed to lead to large differences in evoked brain activity).
Still to-date, only little is known about the various stimulus characteristics of the c-VEP BCI, and specifically how these lead to differences in decoding performance as well as user-comfort. Stimulus characteristics may involve (but are not limited to) firstly the presentation rate, for which faster sequences might become less fatiguing to look at and perhaps trials could become shorter (i.e., the BCI faster). Secondly, the used colors, where black/white is full contrast that hypothetically leads to the highest performance with low comfort, while other colors might recruit different visual pathways and be less visual fatiguing. Thirdly, the stimulus does not necessarily need to be just a color, but can be any alternating image, think of studies that were performed with checkerboards, faces or grid stimuli, or with motion. Fourthly, the actual stimulus sequence itself can be selected or optimized (i.e., determining when flashes happen) in such a way that decoding performance and/or user-comfort is maximized. Finally, typically binary sequences are used, but non-binary sequences such as quintary m-sequences denoting gray-scale values might be promising as well.
Image credit: Martínez-Cagigal, V., Thielen, J., Santamaría-Vázquez, E., Pérez-Velasco, S., Desain, P., & Hornero, R. (2021). Brain–computer interfaces based on code-modulated visual evoked potentials (c-VEP): A literature review. Journal of Neural Engineering.
In this project, it shall be investigated how changing certain c-VEP stimulus properties leads to differences in BCI performance as well as user-comfort. Additionally, a posthoc analysis would look into the changes in the neural response that may lead to better or worse BCI performance, which could lead to novel insights for vision neuroscience as well (i.e., how does the brain respond to a visual stimulus as measure by EEG).
Skills / background required
- Very proficient in Python
- Proficient in machine learning
- Knowledge of vision neuroscience