Keeping the world straight in our heads

When you were a child, did you ever bend over and look between your legs to see what the world looked like upside down? If you were like me, you were disappointed: for me, anyways, the world didn't look as different as I had hoped. Though turning things upside down does make it more difficult to get around, we're actually quite good at adapting to changing our head position. You can do an experiment to see just how good you are at it. Fix your eyes on an object ten feet or so away. Now tilt your head to the left and to the right. Does it appear as if the image is tilting back and forth as you tilt your head? Perhaps a bit, but the effect is nothing like what it would be if you took a movie camera and filmed a scene while similarly tilting the camera.

Now try another experiment. Look at a wall or some other large object ten feet or so away. Now rapidly scan the object with your eyeballs so as to get a good view of the whole thing. Again, the wall doesn't appear to move as you do this, even though clearly the image projected on the back of your eyeball (the retina, in case you've forgotten that eye-dissection lab from high school) is constantly changing. Again, imagine filming a movie scene the same way: the film would be a jumbled mess.

Clearly our mind is doing something to counteract the constant movements we make with our heads, bodies, and eyes. David Whitney, David A. Westwood, and Melvyn Goodale (of CIHR Group, the University of Western Ontario, and Dalhousie University, respectively) devised a set of experiments to determine what inputs we use to straighten out those crooked images on our retinas ("The Influence of Visual Motion on Fast Reaching Movements to a Stationary Object," Nature, 2002).

There really are two possibilities: either our minds detect when our bodies are moving and compensate accordingly, or we look for visual cues that motion is occurring and use that information to compensate.

In their first experiment, Whitney et al. showed participants a computer screen with the following image:

i-157d86ee1cc140e51e892f7275134ac6-whitney.gif

The two striped rectangles displayed animated motion: the black bars on the right slowly moved up, and those on the left slowly moved down. Then a small rectangle flashed to the right of the rectangles. Viewers were instructed to move their hand as rapidly as possible and poke the computer monitor where the small rectangle had appeared (don't ever do this to one of Greta's computers). Viewers consistently poked the screen above where the rectangle flashed, as if it was moving along with the larger rectangles.

Periodically, the movement of the large rectangles changed directions. When the rectangle on the right showed downward movement, viewers poked below where the small rectangle appeared. If the motion changed direction near the time the small rectangle appeared, then viewers started to move their hand in one direction, and shifted their motion to match the shift in the large rectangles.

Whitney and his colleagues had been using a chin rest to make sure participants didn't move their heads, so these results suggested that people do use visual cues to see if the "world" is moving.

Next they strapped volunteers in a chair and rotated it back and forth in a dark room. They flashed an LED in front of viewers and asked them to reach for it. This was essentially the same experiment they had done before, but the sensation of motion was used instead of a visual indicator of motion. Unlike in the first experiment, participants were unable to compensate for the motion of their bodies. In another condition, a set of background LEDs was also included, so that participants had a visual cue that showed them how far they were moving. In this condition, the results were similar to the first experiment.

So it appears that we actually use visual cues to help us see objects as stable, despite the fact that the image on our retina moves around all the time, both due to the movement of our bodies, and our eyeballs. It seems that we constantly examine the data coming into our eyes and adjust it so that we experience a stable world. Perhaps most amazing of all is the fact that we do all this unconsciously, from the time we are small children.

More like this

There are two powerful physiological processes that the article fails to mention. The first is nystagmus, the compensatory motion of the eyes in the skull which is controlled by signals from the organs of balance. The eyes, during nystagmus, move slowly through 10-20 degrees in such a direction as to "freeze" the image on the retina. When they have moved such an amount, they flick back to their starting position while at the same time "dimming" the image of what is being viewed. After the flick is complete, they once again resume their slow traverse undimmed. The result is the appearance of a relatively stable viewing target being looked at by means of successive jumps in direction of gaze.

The power of this mechanism can be appreciated by spinning oneself, at constant angular velocity, for a minute or so. During this time, the nystagmus will subside due to adaptation, and the moving surrounds will become quite blurred. A second demonstration can be had by holding one's open hand about 2 feet in front of one's face. If the hand is rapidly moved from side to side through about 30 degrees arc the image of one's fingers will be quite blurred. In contrast, if the head is shaken through about the same angle and speed, the fingers will remain clearly visible. The difference is the contribution made by the semicircular canals and their control of eye movement.

The second effect that was not mentioned is due to the adaptation of motion receptors. It has long been known that their are specific cells in the visual system that respond only to movement (speed and/or direction) of a visual target or pattern. This specific information, when added to other components such as edges, color, etc. are composited into the image we perceive. Those receptors, like all neurological processes, are subject to adaptation. That means that over a period of time, usually 10's of seconds, those receptors will adapt and no longer contribute significantly to the perception. When the constant motion ceases or reverses direction, the adapted receptors contribute signals with large errors to the overall perception.

It appears, then, that not only do we use the visual cues you describe in your blog, but we rely on inertial signals from the organs of balance. In addition, those organs put out signals that correspond to our head's velocity and these are compared with the signals from the visual system's velocity receptors. A complex picture indeed. (It might also be of interest to learn that motion sickness is almost invariably the result of a significant mismatch between these to types of signals.)

By Richard Malcolm (not verified) on 17 Mar 2005 #permalink