BCIwiki (bciwiki.com) — A closed-loop brain-computer interface (BCI) called MetaIDNet made EEG-based color vision testing 70% faster at the same accuracy in offline tests and 40% more accurate at the same speed in online tests, researchers at the University at Albany, State University of New York and collaborating institutions reported in Sensors on September 30, 2026 (DOI: 10.3390/s26196197).
The team had previously demonstrated objective, quantitative BCI-based color vision testing. Participants look at a stimulus flashing between 2 light sources. Sources that appear to be different colors elicit larger steady-state visual evoked potentials (SSVEPs), while similar-looking colors elicit smaller ones, so the response reflects an individual's color vision.
Accuracy depends on how the color space is sampled and on noise in individual observations. Existing systems sample the color space uniformly and average repeated SSVEPs at each location, trading speed for accuracy. MetaIDNet instead uses Gaussian process regression (GPR) and a neural network to improve that tradeoff.
The researchers ran offline experiments with 10 participants and online experiments with 8, comparing MetaIDNet with 2 existing methods. The authors say the results confirm significant gains in both speed and accuracy. The authors are Hadi Habibzadeh, Daphney-Stavroula Zois and James J. S. Norton.