Abstract
Purpose :
To examine the relationship between the topography of the three cone types in the human retina and the visual sensitivity of individual receptors measured under cone-isolating conditions.
Methods :
An adaptive optics scanning laser ophthalmoscope was used to assess single-cone increment thresholds across a patch of cones (eccentricity = 1.5 degrees) in two color-normal subjects whose cone mosaics had previously been spectrally classified using cone-resolved retinal densitometry. Test conditions were set to favor detection either by long-wavelength (L-) or middle-wavelength-sensitive (M-) cones. For L-cone isolation, a constant blue background (λ = 470 nm) was used to adapt M-cones selectively, leaving L-cones more sensitive to a red stimulus (λ = 710 nm). M-cone isolation was achieved by presenting a green stimulus (λ = 543 nm) on a red background (λ = 710 nm) which selectively reduced the sensitivity of L-cones. The dim infrared imaging raster (λ = 840 nm) and light leak through the respective stimulus channels also contributed to the appearance of the background, such that for each test configuration, the non-targeted cone class was about twice as adapted as its targeted counterpart. The relationship between a cone’s threshold and the adaptation-induced activity of nearby receptors was examined by evaluating the average threshold for cones in the targeted class (e.g., L-cones in L-cone isolation) as a function of the number of cones of opposite type in the adjacent vicinity.
Results :
We found that the sensitivity of a cone under different cone-isolating conditions was largely predicted by its spectral type. We also observed substantial within-class variability in sensitivity. This variability was accounted for by inhibitory signals originating in adjacent receptors: cones with more highly-adapted neighbors had lower sensitivities. This trend was consistent for both adaptation conditions. Monte Carlo reshuffling of the cone mosaics demonstrated the neighborhood effects were unlikely to arise by chance, and optical modeling ruled out uncorrected stimulus blur as a primary contributor to the neighborhood effect.
Conclusions :
Our results confirm that percepts elicited from single cones can be studied reliably in vivo and suggest a psychophysical platform for unraveling, at the scale of individual cells, the lateral interactions between cone classes in the circuitry subserving color vision.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.