In this video we will cover how color output from the cones gets processed into the colors we recognize. This is part 5 in our series about color vision. Stare at the center of this oddly colored flag for 5 seconds or so.
Then look at this white space. The after image is a properly colored flag. How odd!
Understanding how these colors Leaving an afterimage like this was a clue to early researchers about how color vision works. Here is the plan for what we will cover in this video. The sensation of color begins with the cones in the retina.
The cones sense different wavelengths of light and generate nerve impulses, which undergo a first step of processing by ganglion cells also in the retina. The nerve impulses are then carried to the brain, where color information gets further processed into color opponent channels, which ends up creating the colors you see. In the last video we looked at cones and color in detail, so here we will start with a brief review, but the main subject will be color opponent processing. Light from the outside world is focused by the cornea and lens. Within the eye, the retina is the part that senses light, the specialized cells called rods and cones.
As the photoreceptors receive the light, that triggers a nerve impulse that is sent along the optic nerve back to the brain. If we take a piece of retina and examine it under a microscope, it looks like this. The retina is a layer of nerve tissue that contains three layers of cells. Ganglion cells on the top, bipolar cells in the middle, and located in the bottom cell layer are the photoreceptors, the rods and cones, so named because of their shapes.
the cones function in bright light and give us our color vision in dim light rods give us grayscale vision like this they do not contribute to color vision in daylight estimates of the total number of colors we can distinguish reach past a million how do you sense such a wide range of colors historically in the early eighteen hundreds thomas young and later hermann von helmholtz realized Your eye cannot possibly hold a receptor for each color and have them all spread throughout the retina. They theorized that all colors could be perceived by a combination of just three receptors, one red, one green, and one blue, which much later were identified as the cone sensitivities. They theorized that as an image falls on the retina, each receptor sends a separate color signal to the brain, which assembles that information into an image you recognize.
Here is a typical complement of photoreceptors, with each of the cones shown by the colors they are named for. A more accurate way to look at the cones is by the part of the spectrum that each is most sensitive to. The naming convention works like this.
Blue light has a short wavelength, so these are called either blue cones or S cones. Here is the peak of rod sensitivity, but again they don't contribute to color vision. The green sensing cones operate in the middle wavelengths, so they are called M cones. Toward the red end, these cones sense longer wavelength light, so they are called L cones. Note how the range of each cone overlaps the others.
As showers of photons arrive, they trigger nerve impulses from the cones. If there was only one cone type, shown red in this example, and an equal amount of photons arrived at either 500 nanometers or 620 nanometers, the cone would respond with the same output of nerve impulses. One cone won't get you color vision.
But with more than one type of cone, the visual system can compare the proportion of nerve impulses from each cone type. You can see the same incoming wavelengths trigger different combinations of cone responses, thereby allowing us to discriminate quite finely between colors. How many colors? More than a million.
Amazing. Within the retina, here is how the circuitry is organized. Each cone is connected to other nerve cells in the retina.
Usually, the output of several photoreceptors is gathered by a ganglion cell that sums the response for that area of the retina. The area of cones that feed into a ganglion cell is called a receptive field. In the fovea with the finest detail vision, it is one cone to one ganglion cell.
To picture this, consider an area of the retina. There will be many overlapping receptive fields gathering cone output. Now, a receptive field is organized in a particular way, into center and surround areas like a bullseye. The center is fed by cones of one type, while the surround is fed by cones of another.
If hitting the center excites the ganglion cell, it is an on-center. Hitting the surround area with an opposite input inhibits the ganglion cell, which is called an off-surround. And there can be the opposite setup, an inhibitory off-center and an excitatory on-surround. Color vision researchers have found four arrangements in the retina. One, a center of red cone activation with a surround of green cone inhibition.
Two is a green center with a red surround. Technically, blue also inputs into this system. Here is the other center-surround pair with blue versus yellow. Here are the two opposing pairs, which show that this system is divided into red versus green and blue versus yellow.
One term for this is a cone-opponent system. Let me repeat. We have just seen cones feeding into ganglion cells. in a center-surround organization, creating a cone opponent system.
Note that there is also black versus white, but that doesn't affect our color sense. But we are not through yet. Young and Helmholtz thought they had a sound explanation for color vision using just cones, but there were aspects of color vision that weren't accounted for by this theory. To try and explain those issues, an Austrian named Ewald Hering proposed an alternate system that divided color perception into three channels, blue-yellow, red-green, and black-white. He called this opponent color theory.
He based his opponent concept on two main findings. First, and this is something you may not have noticed, there are certain combinations of colors that don't exist in ordinary perception. For example, we can perceive purple as a combination of blue and red. But we cannot perceive a color that is a combination of blue and yellow.
There is no bluish yellow. Likewise, there is a perception of orange as a combination of yellow and red, but there is no color perceived as greenish red. A mixture gets yellow, but not greenish red. These pairs of colors, blue and yellow, and red and green, are mutually exclusive, so they are called opposing or opponent colors. Supporters of the two theories of color vision, cones versus opponent colors, battled for many years.
Time passed and researchers got a better look at the detailed workings of the retina and color processing. Eventually, it looked like both theories are correct. So let's take a look at how the opponent color process works.
Color information that started in the retina is divided into three separate channels. One channel is for luminance. black versus white.
Another channel is for blue versus yellow and the third channel is for green versus red. Here are all three color axes. Any color can be specified by a location in this opponent color space.
Here is how each of the channels works. The channel for luminance receives its input from red and green cones. That determines the black versus white range. In the blue-yellow color channel, the blue input is from the blue cone, of course, but where is the yellow cone? Yellow, you may remember, comes from the combination of red plus green.
So, by the relative input of the cones into the blue-yellow channel, the incoming light is judged to be either bluer or yellower. The red-green color channel also gets input from all three cones. In this channel, the incoming light is judged to be either redder or greener.
So here is the diagram of the whole circuit, which yields the opponent color channels. Now, let's give you a different view, then we will return to this circuit. Here is another way of looking at the result of opponent color response.
This is the blue versus yellow channel. Below 500 nanometers, the blue side dominates. Above 500 nanometers, yellow dominates.
The red versus green channel looks like this. with red dominating at each end and green in the middle. When you put the opponent channel functions together, here is the result.
Let's take a look at a specific example. Take blue at 450 nanometers. The red channel has input, as does the blue channel.
The sum of the two inputs is this specific blue. Every color is perceived through a balance between the two channels. Now let's see if we can put all this together. Starting with the cones, their input into the ganglion cells makes the first separation into blue-yellow and red-green orientation. These are the cone opponent channels.
When this was first discovered, it seemed this would be the mechanism for color opponents. Unfortunately, it is not that easy. Experiments show that this output does not match our final color experience.
There is another step in color processing that occurs in the cortex that yields color results that do match our perceptions. What does that mean, match our perceptions? If color processing was only...
by the three cones, this is approximately what the spectrum would look like. After cone-opponent processing, this is what arrives at the lateral geniculate nucleus before reaching the brain. Most of the colors are there, but they are not in quite the right places.
After the final stage of processing, the colors end up in the places we are accustomed to seeing them. We can't leave opponent colors without specifically recognizing the Unique hues. For a long time it has been recognized there are four psychologically elementary colors. Blue, green, yellow, and red. They are unique in that they are pure, not seen as combinations of other colors.
In theory, these colors occur where one channel is at its neutral point, leaving the other channel unopposed. Number one. There is a spot at 580 nanometers where red and green cancel. leaving yellow unopposed.
This is the site of unique yellow at 580 nanometers. Red and green also cancel at a second spot. This leaves blue unopposed. In other words, this is the site of unique blue at about 475 nanometers.
Likewise, there is a spot where blue and yellow cancel. This is the site of unique green at about 500 nanometers. Unique red is tricky.
It falls outside the spectrum. I've taken these numbers from the original Jamison and Hurwitz paper. Later research has shown quite a bit of variation between people in the location of the unique hues.
For your amusement, you might notice these colors turn up in lots of places. Let's now go back to our flag demonstration. The complementary nature of the colors in the afterimage was the other main clue that there had to be more to color vision than the three cone colors. Staring at the yellow fatigues that part of the blue-yellow channel, leaving blue temporarily dominant. Likewise, staring at the green leaves red temporarily dominant.
And that is how the afterimage occurs. Let us finish by making an estimate of the amount of colors we can see. one way to think about this is to start with the spectrum we can distinguish about a hundred and eighty pure spectral hues then there is all the variation in saturation of luminance to be accounted for here is one way of calculating a number provided by professor knights of the university of washington first with one cone type there is no color color space is one-dimensional in which we can distinguish about two hundred levels of gray Adding a second cone creates a two-dimensional color plane. On the blue-yellow horizontal color axis, we can distinguish about 50 color steps. So 50 times 200 equals 10,000 colors.
Adding a third cone adds a third dimension. The red-green axis adds 100 steps of discrimination. If the result was a rectangular solid, then 10,000 times 100 equals 1 million colors. This is Neitz's diagram, including the full color map, from which he projects over 2 million possible colors.
In summary, perception of color starts with the cones. Cones input into retinal ganglion cells by a center-surround organization, creating a blue-yellow and red-green orientation. These are the cone opponent channels. When this was first discovered, it seemed this would be the mechanism for color opponents.
but experiments show this output does not match our final color experience. The next step occurs in the cortex, where the color opponent process yields color results that do match our perceptions. Here is the estimated color result from each of the stages of color processing.
This is a good place to point out that researchers are still working on each of the above steps. In the next video, we cover what vision is like with different numbers of cones. That gives us part of the story of what vision is like in other animals. And in another video we focus on the very interesting story of color vision in primates.
Here are selected references if you want to read further.