Abstract
Purpose:
More than a half-century of psychophysical experiments have indicated blue-yellow and red-green neural channels are responsible for color perception. Physiological recordings have identified single neurons early in the visual pathway with red-green and blue-yellow responses. However, the cone inputs necessary to produce color sensations as measured psychophysically do not align with the characteristics of opponent cells measured physiologically. To reconcile biology with perception, we measured color percepts elicited by stimulation of identified single cones in a human subject.
Methods:
An adaptive optics scanning laser ophthalmoscope was used with retinal densitometry to classify (as L, M or S) ~1000 cones in a subject. Using the same system, single cone stimulation (543 nm) was achieved following published methods [Harmening et al. 2014 J Neuroscience]. The subject was instructed to report the sensation elicited by each stimulus using an electronic tablet. Pilot work indicated that, under our experimental conditions, red, green, blue, yellow and white were sufficient color categories to adequately describe the percepts.
Results:
On a dim white background these single cone conditions did not elicit blue percepts. To encourage blue sensations, we adopted a background dominated by short wavelengths to preferentially hyperpolarize S-cones. Under these conditions, when the flash was detected, the percentage of percepts named blue was 22%, white 60% and red 18%. Stimulation of L-cones rarely resulted in blue percepts (9% of trials). M-cones were significantly more likely to result in blue responses (43% of trials). Finally, when M-cones were targeted, the probability of seeing blue decreased as distance to the nearest S-cone increased.
Conclusions:
These results provide direct evidence that blue percepts can be driven by stimulation of individual M-cones. The increased likelihood of blue responses with M-cone proximity to S-cones suggests this sensation arises when M-cones sum with S-cones in post-receptoral pathways. Small bistratified ganglion cells, assumed to be the retinal pathway responsible for blue color perception, however, difference S- and M-cone signals, exactly the opposite of the trend found here. Our results are consistent with the existence of alternative neural circuitry in the retina responsible for mediating blue sensations in which outputs of S + M cone signals are differenced from L-cones.