Adams noted that you can get an interesting effect by staring for a minute or so at a waterfall (keeping the eyes steady). For a short time afterwards, a stationary scene (the rocks adjoining the falls) will appear to move in the opposite direction. Versions of the spiral seen here were first used in 1849 by the Belgian physicist Joseph Plateau. 

This spiral, with three concentric and conflicting zones, was created by 
American magician Jerry Andrus, who specializes in creating marvelous
three-dimensional optical illusions. A wonderful toy version of this  spiral can be purchased from Binary Arts.
 

So What's Going On?
 

    Like all aftereffects, this one is due to the fact that the nerve cells that signal motion in the direction the stimulus is moving in fatigue after several seconds or more of continuous firing activity. In humans, cells that signal the direction of a moving stimulus are not to be found in the retina, but first appear in the cortex. 

     Your brain represents a sensory quality, such as motion, brightness, color or depth, not in terms of the firing of one group of neurons, but in terms of the electrical activity of one group of neurons relative to the electrical activity of another group of neurons. For example, the activity of neurons coding for clockwise motion relative to the activity of those coding for counterclockwise motion.

     Clockwise motion is signaled by the fact that the neurons coding for clockwise movement fire more strongly than neurons coding for counterclockwise motion. If the clockwise neurons now fire less strongly because their electrochemical batteries run down after several seconds or more of continuous activity, the balance between these two groups is disrupted. 

     If you now look at a stationary target (such as a wall or the back of your hand), the neurons coding for clockwise motion remain inhibited relative to the neurons coding for counterclockwise  motion and the brain concludes, quite sensibly, that a target that moved in the opposite direction to the original stimulus is present. This aftereffect is quite intense for a few seconds, and can last for up to 20 seconds or so. 
 

     What is so strange about this aftereffect illusion is its paradoxical nature.  Although your hand or friend's face is being seriously distorted it does not change. This is true for all motion aftereffects: the apparently moving features nevertheless seem to stay still! A sensation of expansion or contraction (depending on the direction of the spinning disk) does exist, but the contours do not appear to be going anywhere. This perceptual paradox suggests that the visual system detects and represents velocity (motion) and position using different neurons and neuronal subsystems. Thus, one system can adapt (or even be destroyed such as in a bullet wound through the brain)  while the other one remains intact and non-adapted.  Because the neuronal system signaling the position and the shape of an object is not adapted, you see motion without any change in the aftereffect.
 

      An interesting question is why you do not see a strong motion aftereffect after we have been driving for minutes or hours on the freeway. After all, during all this time the movement of the entire world outside our car must have totally fatigued our motion systems. Yet even if we stop abruptly, we do not perceive that the world around us suddenly moves backwards. (Some people, however, have noticed slight effects). 

      Two reasons appear likely. First, if the motion covers the entire field of view of the eye, little or no motion aftereffect of any kind occurs. The brain is concerned with signaling motion relative to some stationary point or relative to some other part of the image moving in a different direction. Humans are not very good at detecting motion of an object when it is seen with no background or when the entire field of view moves roughly in the same direction. 

     Recent experiments have shown that you can also achieve a motion aftereffect from a non-moving stimulus -- induced motion. In addition, studies have shown that after many successive viewings learning may occur so that the stimulus pattern induces a motion aftereffect even when it is not moving.

     This indicates that the motion aftereffect is still not completely understood, nor can it be completely explained by the fatiguing of neurons.

Second Experiment

     Fixate on the spinning disk again for thirty seconds or more, but this time with only one eye. When the time is up close that eye and immediately look at the back of your hand with your other eye (the one that did not see the stimulus). You should still see a motion aftereffect, although this time the effect will be weaker.

So What's Going On?
 

     This demonstration tells us where in the visual pathway your brain is doing the processing. It shows that the effect you are seeing is partly due to neurons that only receive input from one eye (so-called monocular neurons) and partly by neurons that receive input from both eyes (binocular neurons). For the latter ones, it does not matter whether the original adapting stimulus came from the right or from the left eye. These neurons explain why the aftereffect transfers -- Interoccular Transfer. The evidence of ttransfer seen here helps determine at what stage in the visual system this effect arises. Stationary black and white afterimages originate with nerve cells in the retina. So, none of these afterimages transfer from one eye to the other. However, research carried out on monkeys, our closest animal relatives, has revealed that neurons responsive to the direction of a moving stimulus do not occur until the primary visual cortex. Therefore, these motion aftereffects are due to cortical -- not retinal -- processing.