Abstract
Recent work suggested that it is possible to decode which chromaticity an observer is viewing from the multi-electrode pattern of low frequency EEG activity on the scalp (Bocincova and Johnson, 2018). However, chromatic stimuli may vary also in luminance, and there has been debate about whether differences in the visual evoked potentials (VEPs) from stimuli of different chromaticities are driven by differences in the luminance or the chromaticity of the stimuli (Skiba et al., 2014). Thus, an open question is whether the chromaticity of a stimulus can be decoded even when differences in luminance do not inform the classification. To answer this question, we conducted two experiments in which we systematically varied both the luminance and the chromaticity of a centrally presented disk. In experiment one, we presented two chromaticities (appearing red and green) at three luminance levels (6.5, 10.8, and 17.4 cd/m2, on a 5 cd/m2 background) on separate trials. In experiment 2 we presented 4 chromaticities (appearing red, orange, yellow, and green) at two luminance levels (6.5 and 10.8 cd/m2, on a 5 cd/m2 background). For each observer, we first performed heterochromatic flicker photometry to equate each chromaticity at each luminance level. Next, observers monitored centrally presented chromatic discs (150ms stimulus duration) while EEG was recorded. Using a pattern classifier and the multivariate topography of scalp EEG activity we were able to accurately decode the chromaticity of observed stimuli. We could also decode the luminance level of each stimulus. Critically, we were able to decode the chromaticity of the stimuli when we trained the classifier on the chromaticities presented at one luminance level and tested at a different luminance level. Thus, the multivariate topography of EEG activity can be used to decode which chromaticity is being viewed, even when chromaticity is decoupled from variation in luminance.