Most of you have probably seen this before, but if you haven’t, look at the flag for 30 seconds (if it doesn’t work with 30, try 60), and then look at the white space underneath it.
You should see a red, white, and blue flag when you look at the white space. That is a color afterimage.
Again, most of you probably know how this works, but just in case you don’t, I’ll briefly explain it. As I’m sure you know, when light comes through the iris, it is projected onto the retina by the lens. The retina is covered with photoreceptors, which come in two types, rods and cones. Photoreceptors are pretty much always active, and when there are no photons striking them, they release a certain amount of the neurotransmitter glutamate. When exposed to light, they become hyperpolarized, and release less glutamate. Cones, the photoreceptors that respond to light, come in three types: those that respond (by releasing less glutamate) to red, those that respond to green, and those that respond to blue.
Attached to photoreceptors are cells called bipolar cells. Bipolar cells come in two types: inhibitory and excitatory. When a photoreceptor becomes hyperpolarized and stops releasing glutamate, the attached inhibitory bipolar cells decrease the activity of a third type of cell, ganglion cells. The excitatory bipolar cells in turn increase the activity of ganglion cells. As a result of the way photoreceptors and bipolar cells are organized, each ganglion cell responds to two colors. One of the colors will excite the cell, and the other will inhibit it. For example, some ganglion cells are excited by cones that respond to red, but inhibited by cones that respond to green.
This is how we get the afterimage. When cones are exposed to a particular color for an extended period of time, they undergo what’s called neural adaptation. This means that they become less sensitive to that color. So, for example, if we stare at a green splotch for an extended period of time, the cones that respond to green will stop responding to green. In fact, they’ll actually become more depolarized than they are in the dark, meaning they’ll release more glutamate than they would in the dark. Meanwhile, neighboring cells that respond to other colors (e.g., red) will still be releasing glutamate at the normal rate. This will mean that a ganglion cell in that area that are inhibited by green but excited by red will begin to act like they’re being exposed to red. So, we see an a red afterimage.
In addition to color afterimages, there are other well-known types of aftereffects. The most striking is probably the motion aftereffect. Unfortunately, I don’t know how to stick movies into moveable type, so I can’t show you one here, but if you’re curious, you can see one version, the “waterfall illusion,” here. Motion aftereffects are also a result of neural adaptation. In this case, the adaptation occurs in cells in the primary visual cortex that respond to motion in particular directions. In the waterfall illusion, cells that respond to downward motion adapt, and neighboring cells that respond to upward motion become relatively more active, and thus we get the illusion of upward motion.
As the explanation of color afterimages and motion aftereffects indicates, one of the interesting implications of aftereffects for a particular dimension (e.g., color or motion) is that there are neurons (or groups of neurons) in the visual system that respond to values on that dimension. Tomorrow, I’ll post on recent research that uses the demonstration of aftereffects as evidence for the existence of neurons that respond to information in a more surprising domain.
UPDATE: In comments, “Pdf23ds” provided a link to this striking color afterimage: The Castle.