Getting a new view of motion: Iman Brivanlou (left) and Michael Berry do experiments to see how animals and humans anticipate the position of moving objects. Photo by Kris Snibbe.

Pretend you are Mark McGwire and you've just come to bat in the last inning of a key game. The pitcher winds up, then throws a fastball to you at 90 miles per hour. Because you're human, you have a handicap. It takes about a tenth of a second for your eye to tell your brain where the ball is. But in that time the ball travels about 13 feet. Therefore, you would swing at the ball long after it passes you by.

Those who play softball, tennis, ping-pong, and other hit-the-ball games face the same problem. While your eyes and brain work to pinpoint the ball in space, the ball can move six or more feet ahead.

Most people, especially McGwire, do a good job of making contact with the ball despite the delay. How do they do it?

A group of researchers at Harvard University asked themselves this very question.

"We guessed that the retina of the eye might begin processing visual information before it sends it to the brain," explains Michael Berry II, a postdoctoral fellow in molecular and cellular biology. "Then we did experiments proving that the eye is capable of projecting ahead the trajectory of a moving object so it's seen at its actual location. In fact, activity in the eye can even lead, or anticipate, the position of a steadily moving object."

Shifting the Peak

Berry and his colleagues didn't start out trying to solve this problem. They were working for Markus Meister, a professor of molecular and cellular biology, on ways to accurately measure responses to moving objects by the retina.

The retina is a light-sensitive screen at the back of the eye that converts visual scenes into nerve pulses that travel over the optic nerve to the brain. To make their measurements, Berry, Meister, and research assistants Iman Brivanlou and Thomas Jordan place sheets of retinal cells from rabbits and salamanders over a grid of electrodes that measure which cells respond when different patterns of light are flashed at them.

During this research, Meister noticed that a continuously moving bar of light caused a wave of excitation to move across the retina. He further noticed that the retinal cells fired at, or slightly ahead of, the front edge of the bar. The Harvard team was also aware that other researchers had found a similar visual illusion: when people look at a flashing light and a moving light, both at the same position, they perceive the moving light to be ahead.

"We decided to look into this phenomenon more closely," Berry notes. In doing so, the researchers discovered a clever mechanism at work.

As an object enters the visual field of a retinal nerve cell, the cell starts to fire rapid pulses. As the edge of the object sweeps across the field, a negative feedback mechanism quickly turns down the cell's light sensitivity so it fires less rapidly. That has the effect of quickly shifting the peak firing rate to the front edge of the moving object. When the message goes to the brain, the object is perceived where it is and not where it was.

It makes sense for such a system to evolve. A hungry predator looking for its next meal needs to know the actual position of its prey, not where it was a tenth of a second before. Likewise, a rabbit about to dodge the charge of a coyote requires up-to-date position information.

Matching Sounds

Hearing doesn't involve the same time delay. Sounds vibrate relatively large membranes ("ear drums"), and nerve cells attached to these membranes respond more quickly than light- gathering cells in the eye.

"The auditory delay is only a few thousandths of a second compared to a tenth of a second," Berry points out. "Inputs of sight and sound come together in the brain's thalamus, so it's the mark of an elegant design to correct the visual delay in the retina before reaching that point." The thalamus consists of two walnut-size bundles of nerve tissue deep in the brain.

Berry speculates that higher areas of the brain possess the same capability to anticipate movements as the retina. With the eyes following a tennis ball, the brain might realize that, when an opponent hits a certain type of shot, he or she tends to move toward the net. In other words, the primitive mechanism that foresees the trajectory of a ball or of prey, might also be used by a more abstract system at a higher level to predict behavior.

Since many animals hunt at night, retinal anticipation must also work in the dark. The eye contains structures called rods, which are sensitive to light levels millions of times lower than daylight.

"Rods can change sensitivity over a wide range, allowing eyes to adapt to many levels of light," Berry explains. "But the price you pay for this increased sensitivity is a decrease in response time. A leopard hunting at night may be twice as slow in its visual responses as in daytime. As for humans, their ability to anticipate poops out at night. We're adapted to sleep at night, not hunt."

For retinal anticipation to work best, a target must move at more or less constant speed and direction. Over short periods of time, it's reasonable to assume that baseballs, tennis balls, and traveling meals will do just that. But a batter or a hunter can't be sure that's the way things will always go.

Berry has begun doing experiments to determine how the retina handles sudden shifts in speed and direction. In one experiment, a light suddenly appears, then starts to move. Compared with a steadily moving light, it takes twice as long, about one-fifth of a second, for the correct anticipation response to set itself up.

That's one reason why McGwire misses some of those breaking balls, and a lion can fail to catch a rapidly zigzagging antelope.