Reversed effect of adapting stimuli on visual sensitivity

Abstract

The sensitivity of the eye to incremental stimuli normally decreases as the intensity of a steady background field increases. We examine here a range of conditions under which an added field anomalously increases sensitivity. If the threshold intensity for violet (A = 423 nm) test flashes is measured on a blue (μ 1 = 473 nm ) field of 10 9.7 quanta s -1 deg -2 and if then yellow (μ 2 = 575 nm) fields of increasing intensity are added to the background the threshold may fall by 0.3—0.4 logarithmic unit. The facilitation occurs whether yes/no or forced-choice procedures are used (experiments 1, 2), is abolished when the blue and yellow fields are presented dichoptically (experiment 5) and varies in magnitude with the duration of exposure to the composite adapting field (experiment 6). The intensity of the yellow field required for maximal facilitation ( a ) increases as the intensity of the blue field increases and ( b ) is typically that intensity required to render the composite field achromatic (experiments 1, 3). The results imply that the sensitivity of the psychophysically defined blue mechanism is not controlled merely by photons absorbed in the shortwavelength receptors. We hypothesize that signals originating in the short-wavelength receptors are confined to opponent-colour channels (assumption 1) and that any opponent-colour channel is most sensitive to input perturbations when at an intermediate value of its response range (assumption 6). We relate the latter principle to line-element analyses of colour discrimination. Variation of μ 2 allows us to estimate the action spectrum of the longwavelength input to the putative opponent channel (experiment 4). This spectrum resembles the photopic luminosity function, V; but for μ 1 > 600 nm the facilitation is less than for μ 1 « 575 nm. We discuss the implications of the present results for the π mechanisms of Stiles. If detection were by opponent channels, might Stiles’s field sensitivities be contaminated by variation in post-receptoral sensitivity when field wavelength was varied or when an auxiliary field was added? Although hindsight does suggest that the large and long targets used by Stiles would favour detection by opponent channels, an unappreciated advantage of the two-colour method may be that the monochromatic field desensitizes opponent channels and ensures that detection is via a nonopponent pathw ay whose (post-receptoral) sensitivity varies little in the course of field-sensitivity m easurem ents. Stiles’s own checks and other evidence suggest that π 4 and π 5 are little distorted by opponent processes and even for the blue mechanism, where the use of long-wavelength auxiliary fields would be likely to modulate the sensitivity of opponent channels, the relative field sensitivity for μ < 500 nm is probably that of the receptors. The present results do suggest that Stiles’s two-colour method can, with care, be extended to the study of chromatically opponent channels in the visual system.