Psych 129 - Sensory processes

Color

The nature of color

• The perception of color is an entirely subjective experience.
  

• The perceived color of an object depends on the mixture of light wavelengths reflected—i.e., those wavelengths that are not absorbed—by the object.
  
  
• Common psychophysical measures of color space are hue, saturation, and brightness.
  
  
  

• During the same period in which Isaac Newton discovered the binomial theorem, invented the calculus, and devised the theory of gravity, he also bought a prism that he later used to study the nature of light. From these studies, he concluded that white light was actually a superposition of light of many different colors. When white light was passed through the prism, different wavelengths of light were bent by different amounts (although Newton did not know that) and revealed the multi-spectral composition of white light. The notion that white light was composed of many different colors was quite a radical idea to propose at the time, and many philosophers, such as Goethe, refused to believe it.
  
  
• Newton deduced that certain colors are “primary” because they could not be further subdivided. He proposed that there were 7 primary colors, in line with there being 7 notes in an octave in the (primarily western) diatonic scale.
  
  
• Newton further invented the color circle to predict how colors would look when mixed—i.e., using a linear combination rule.
  
  
• Newton got alot right, but went wrong with the 7 primary colors idea. Colors lie along a continuum. The visual system is simply limited in terms of how finely gradations along that continuum can be discriminated, which is why “primary” colors don’t be appear to get further subdivided. An organism with more cone pigments that are more narrowly tuned, such as the Mantis shrimp, would have seen the results of Newton’s experiments quite differently, and might have concluded there are hundreds of primary colors.
  
  
  

• An additive color mixture occurs when two light sources are mixed together—e.g., when red and green light are added they produce yellow.
  
  
• Complimentary colors are those that produce the perception of white when added together—e.g., when you add blue light and yellow light you perceive white light.
  
  
• A metamer, in general, is defined as two things that are physically different but perceived the same. Thus, red and green adding to produce yellow is a color metamer, because yellow could have also been produced by a single light with wavelengths between green and red. Your visual system is fooled into thinking that red+green is the same as yellow, when in fact red+green has a completely different spectral composition than yellow.
  
  
• A subtractive color mixture occurs when two reflecting mediums are mixed - for example when you mix ink or paint. In this case, you get the union of the absorption spectra of the two substances, and thus the intersection of the reflectances of the two substances - e.g., mixing blue and yellow ink produces green ink (not white, as it would with an additive mixture).
  
  
  

• Color constancy refers to our ability to remove the effect of the illuminant in perceiving the color of objects. Without color constancy, objects would appear quite different depending upon the wavelength composition of the light illuminating them. For example, sunlight, tungsten light, and fluorescent light have quite different spectral compositions, and so the spectral composition of light reflecting off of an objects will be quite different depending on which of these light sources is illuminating it. If the visual system did not have the capacity to correct for variations in illumination, then your perception of objects would be radically different—for example, a red shirt may appear orange—depending on the source of illumination.
  
  
• Color constancy is thought to be achieved by the comparison of wavelengths emanating from objects. The visual system assumes that all objects within the same general region of a scene are being illuminated by the same source of light, and so the reflectance of objects is perceived relative to the reflectances of other objects around it.
  
  
• A disadvantageous side-effect of color constancy is that color appearance is not veridical (i.e., truthful). Our visual system does not act as a spectrometer, but rather attempts to correct for illumination. Thus, judging the color appearance of an object depends heavily on its context, or what colors it is surrounded by. An example of this is the “color” brown; really there is no such thing as brown all by itself—an object will look brown only in relation to the other things around it. If you isolate brown all by itself, it looks rather dark orange.
  
  
  

• Trichromacy refers to the fact that the visual system uses three cone types, with three different absorption spectra, to resolve the wavelength composition of light.
  
  
• By contrast, a monochromatic visual system with only one cone type would be unable to differentiate what are to us different colors. The responses of the cones would be confounded by wavelength and intensity. Everything would probably (although who really knows?) appear simply as shades of gray.
  
  
• A dichromatic visual system with two cones types could begin to resolve the ambiguity between wavelength and intensity by comparing responses across cones. But confusions could still be produced by independently varying two or more wavelengths of light.
  
  
• Our trichromatic visual system consists of three cone types - S, M, and L, selective to short (420 nm), medium (530 nm), and long (560 nm) wavelengths respectively. By independently controlling the amount of light at each of these wavelengths, any percept of color can be produced (e.g., this is the principle by which color TV works). But we are still “blind” to variations in spectral composition produced by more than three independent dimensions.
  
  
• Why weren’t humans endowed with four, five or even 300 cones types? It depends on what is necessary for survival. There is a trade-off between resolution in space and resolution in wavelength. Achieving both requires alot more neurons and processing power. Some animals, such as the mantis shrimp, possess as many as 8 cone types, although their acuity is certainly not as good as ours.
  
  
• Of the 8 million cones in the eye, about 1 million are S, and the remaining 7 million L (2/3) and M (1/3). The central most portion of fovea is devoid of S cones. Thus, S cones are overall quite sparse and so we have low spatial resolution in this part of the EM-spectrum. The reason why this is done is most likely because the blue light is most blurred due to chromatic aberration by the lens.
  
  
• Most primates are dichromats, with only two cone types: an S cone and something between an L cone and an M cone. It is thought that the L and M cone split occurred around 40 million years ago, and was perhaps encouraged by the need for monkeys to discern ripe from unripe fruit in the jungle.
  

• Certain colors appear to be mutually exclusive - e.g., blue/yellow or red/green. It doesn’t make sense to speak of something having a blueish-yellow hue, or a redish-green hue. They are like opposites, analogous to north/south or east/west. Also, adapting to blue produces the percept of yellow and adapting to red produces the percept of green (and vice-versa). So there seems to be something going on neurally that makes these colors tug against one another. This apparent tug-of-war between certain colors is termed color opponency.
  
  
• From psychophysical experiments, there appear to be three separate “channels” (i.e., large pools of parallel, neural mechanisms) that subserve the perception of color and which can account for color’s opponent character. One is an achromatic (luminance) channel formed by the addition of the M and L cone responses, [M+L], which alone would give you a percept similar to black and white film. The other two channels, termed chromatic channels, are formed by computing differences among the cones, and hence have an opponent character. One is the blue-yellow channel [M+L - S] and the other is a red-green channel [L+S - M].
  
  
• Color opponency is thought to be subserved by color-opponent cells in the LGN, although the match between the cell types found there and the psychophysics is not exact.
  
  
• Afterimages are likely produced by habituation of cones - e.g., adapting to blue will habituate responses of S cones, so that when you now look at white light the S cones will have less response than normally, so the blue-yellow opponent mechanism, rather than getting balanced excitation and inhibition as it normally would in white light, will instead get more excitation from the M+L cones than inhibition from the S cones, yielding the percept of yellow (or so the theory goes - the real story could be more complicated than this).
  
  
• Color opponency is an efficient coding strategy, as it reduces correlations among neural responses.
  
  
• Color names appear to be well-matched to the LGN opponent-cell response properties.
  
  
  

• Color blindness is most often produced by a missing cone pigment, thus rendering the system dichromatic or monochromatic. There is also something called “anomalous trichromacy” in which a person has all three cone pigments, but in slightly different parts of the spectrum than most normal people.
  
  
• Color blindness or deficiencies can also be acquired. For example, glaucoma and diabetes affect the integrity of S cones. And as people get older, the lens tends to yellow and block out more blue light.
  
  
• Central achromatopsia occurs when all three cone pigments are intact, but there is a neural defect at some central stage of processing (i.e., in the cortex), such as area V4.