Abstract
The standard luminance function, V(λ), is mainly defined by heterochromatic flicker photometry (CIE, 1924). It is, however, often in disagreement with heterochromatic brightness judgments. While luminance is firmly grounded physiologically in the retino-geniculo-cortical M-pathway, no such correlate is known for heterochromatic brightness. Here, we wanted to explore whether steady-state visually evoked potentials (SSVEP) at low temporal frequencies could serve as such a neural correlate in humans (Regan, 2009, Vision Research). We recorded SSVEPs to chromatic stimuli flickering at 3Hz or at 15Hz against a black background. In each trial, the stimulus had one of the “principle” colors on the RGB display (i.e., R, G, B, RG, RB, GB, RGB), and its linearized intensities were swept from 12.5%, 25%, 37.5% to 50%, each step lasting 1 second. At a flicker frequency of 15Hz, which is often used in flicker photometry, the SSVEP amplitudes could be well predicted by stimulus luminance (r2 = .81). However, when the frequency was 3Hz, SSVEP amplitudes were barely related to luminance (r2 = 0.13). They were much better accounted for by stimulus max[r,g,b] values (r2 = 0.65). This max rule (1) weights the red, green, and blue channels equally, (2) combines them non-linearly with a maximum rule, and (3) agrees quite well with recent psychophysical findings on color-weight photometry by Koenderink et al (Vision Research, 2018). Our results provide a consistent, reliable and easy to measure neural correlate of heterochromatic brightness.