There are at least three psychophysically-defined detection mechanisms in human vision: an additive or luminance (Lum) mechanism, an opponent red vs. green (RG) mechanism, and a yellow vs. blue (YB) mechanism. Some recent studies find evidence for additional, """"""""higher order"""""""" mechanisms, which respond to more specific colors (e.g., orange). Hue cancellation measurements have been extensively used to define two suprathreshold color appearance mechanisms, and in some cases these hue mechanisms have been equated with the detection mechanisms. The first major aim is to test whether RG and YB are the only two detection mechanisms, and whether they are the same as the mechanisms that are responsible for color appearance. The Lum mechanism will also be studied at and above threshold under novel conditions designed to isolate Lum, by comparing thresholds with suprathreshold heterochromatic flicker photometry and minimally-distinct border settings, which are believed to tap the same mechanism. These measurements should provide estimates of the gains of three cone classes entering Lum; these gains are very difficult to determine under normal circumstances.
A second aim i s to measure the """"""""equivalent input noise"""""""" of RG, YB, and Lum. This noise consists of photon noise, uncorrelated neural noise, and neural noise which is correlated between the different cone classes. RG is more sensitive than Lum under many conditions, and only part of this difference is due to RG's greater spatial and temporal summation: some of the sensitivity advantage may be due to lower noise levels in RG, an idea that will be directly tested.
The third aim i s to study interactions which occur between these three channels when the activity of one of them is suprathreshold. A clearly-visible edge seen by one mechanism can reduce another mechanism's threshold. A series of experiments will test a new theory of this interaction, based upon a cross-correlation function. Studying the interactions of these basic mechanisms is a step towards understanding high level visual perception, which often involves determining the hue or lightness of an object that has clearly-visible edges.
|Wang, Quanhong; Richters, David P; Eskew Jr, Rhea T (2014) Noise masking of S-cone increments and decrements. J Vis 14:8|