Colour Vision 2

Colour vision 2

In the previous lecture the most basic psychological properties of colour vision were described. This lecture looks at the mechanisms which give rise to the properties of our colour vision.

What physical property of light is responsible for variations in colour? Light behaves like a series of waves travelling through space, as with any wave motion we can measure the distance between successive peaks of the wave: this distance is known as the wavelength. For light, colour is the psychological correlate of wavelength. The wavelengths of visible light are rather tiny, about 1/2000th of a millimetre, and are usually measured in nanometres (abbreviation nm; 1 nm = 10^-9 metres). The wavelengths of visible light range from 400 nm at the violet end of the spectrum, to 700 nm at the red end of the spectrum. Remember that colour is a purely subjective phenomenon; to quote Isaac Newton: "For the Rays to speak properly are not coloured".

In earlier lectures the idea of a "tuned" detector was introduced. For example, there is evidence for entities that respond most strongly when a certain, preferred, spatial frequency is present, with decreasing response as the spatial frequency of the stimulus gets more and more different from the preferred value. The photoreceptors of the retina act in this way with respect to wavelength: a given receptor responds most strongly to light of a given wavelength, and more weakly to neighbouring wavelengths. The response of the photoreceptor rises with the intensity of the light as well. Vision in bright light is mediated by the photoreceptors called cones.

Although a photoreceptor’s level of response is a function of two variables (intensity and wavelength) the level of response is itself a single variable (you will see this referred to as the principle of univariance). It follows that we can get an identical cone response from lights of different wavelengths simply by adjusting their relative intensities. For example, if a cone is twice as sensitive to light A as light B, then two units of light B will elicit exactly the same response as one unit of light A. Thus a single cone cannot disentangle wavelength and intensity. The monochromats discussed in the last lecture possess only one type of photoreceptor, and as we saw they are able to get two lights of different composition to look identical simply by adjusting the intensities of the lights.

Now imagine that we have two cones (1 and 2), whose preferred wavelengths are different. We now find that if we adjust our lights A and B to give identical responses in cone 1, then the responses of cone 2 to lights A and B will be different, and vice versa. Thus, by looking at the responses of two different cones, we no longer completely confuse wavelength and intensity. However, if we add a second light B’ to light B, and allow ourselves to adjust the intensities of B and B’ independently, we find that there are intensities of B and B’ for which light A and light (B+B’) produce identical responses in both cones. So two lights are needed to match a third. This is the way a dichromat would behave.

If we have three different cones, we find that we need to mix three different lights in order to be able to match a third. This is how a trichromat would behave. The psychological property of trichromacy is believed to arise because we possess three distinct types of cone in our retinas. These have preferred wavelengths in the blue, green and yellow regions of the spectrum and are referred to as the S (short-wavelength), M(middle-wavelength) and L (long-wavelength) cones.

The three-receptor theory provides simple explanations of colour mixing phenomena. Two lights will appear identical if they produce identical responses from all three cone types. Thus, red and green lights added together appear yellow because they elicit the same set of cone responses as a yellow light would.

It should now be clear that we could construct colour vision systems with four, five or even more different types of cone, yielding the tetrachromats, pentachromats, etc that we discussed as a possibility in the last lecture. The arbitrariness of our colour vision should be obvious here. Our colour vision is largely determined by the properties of our cones: if we had different types of cones, sensitive to different parts of the spectrum, pairs of lights that we now distinguish would look identical, and lights that we currently cannot tell apart would appear different. This even if we restrict ourselves to a system of three cones, there is scope for infinite variation in colour vision.

Postreceptoral encoding of colour

There is good physiological and psychological evidence that the responses of the cones feed into three neural channels. Into the first, achromatic, channel are fed the summed responses of the M and L cones. Because no comparison is made between the cone respones, this channel has no colour discrimination: it signals brightness.

The second channel is the red-green channel, and it transmits the difference between the M and L cone responses.

The third channel is the blue-yellow channel. It takes the sum of the M and L cones (yellow) and subtracts the response of the S cones (blue).

The colour of afterimages is explained naturally by the existence of these channels. Prolonged polarisation of a channel (eg by strong blue light) will cause either fatigue or adaptation of the cells involved. On removal of the stimulation, the channel will return beyond its original neutral point, in this case, to the yellow side of white.

The red-green channel cannot signal red and green at the same time, neither can the blue-yellow channel signal blue and yellow at the same time. However, the state of each channel is independent of the state of the other channel, so that the simultaneous signalling of red+blue, green+blue, red+yellow and green+yellow is possible. Thus the properties of the postreceptoral channels neatly explain the mutual exclusiveness of some colour sensations and the compatibility of others.

Last updated on Wednesday, 10/02/02, at 08:32 PM by