Brain-computer interface (BCI)-based identification of congenital red-green color vision deficiencies

We demonstrate brain-computer interface (BCI)- based color vision (CV) assessment for the identification of congenital red-green CV deficiencies. Experiments were based on the identification of metamers—light sources with different spectral distributions perceived to be the same color. Metamers elicit steady-state visual evoked potentials (SSVEPs) of minimal size and are different for people with versus without CV deficiencies. Methods: Thirty-one participants (20 control (CTR), 11 color vision deficient (CVdef)) completed behaviorand BCI-based CV assessments. Experiments used a visual stimulus that alternated between a monochromatic light source (yellow; fixed luminance) and a dichromatic light source (red and green; varying luminances) at a fixed frequency. During behavior-based CV assessment, participants identified metamers by manually adjusting the dichromatic source’s settings until its color matched that of the monochromatic source. During BCI-based CV assessment, participants attended a sequence of stimuli (each with a different dichromatic source setting) while electroencephalography was recorded; metamers were defined as the dichromatic source settings that minimized SSVEP size. Results: The behavior- and BCI-identified metamers were identical within each group, but different across groups (i.e., CTR and CVdef and for people with protan- versus deutan-type CV deficiencies). Automatic identification of CVdef individuals (and type of deficiency) was demonstrated using a classification analysis (93% accuracy). Conclusion: Experiments validated BCI-based CV assessment for the identification of congenital red-green CV deficiencies. Significance: BCI-based CV assessment does not require behavioral responses and can be automated, making it suitable for people with cognitive/motor deficits. With further development, BCI-based CV assessment may enable automatic identification of many types of congenital and acquired CV deficiencies.