Narrator: Vision is one of our most important senses. In order to understand how this sense works, let us go to the beginning of the visual pathway, the eyes. The eyes don't really see. They are mere sensors, beaming information to be deciphered and reassembled by at least three million brain cells.

In most vertebrates, each waking second, the eyes send a billion items of information to the brain. But what do all these neurons do? And how do they allow us to see?

First of all, the lenses of our eyes turn everything upside down. Then, the visual information is beamed along the heavy fibers of the optic nerve. The first bend in the pathway is the optic chiasm, where some of the information from the left and right eyes crisscross. Then, the pathways reach a major staging stop at the back of the head, a thin veneer of brain tissue. Here, at the primary visual cortex, also called the striate cortex, visual information begins its processing.

At the University of California at Berkeley, a team of scientists led by Russell De Valois and Roger Tootell set out to find just what information this primary visual cortex receives. This image was flashed before the eyes of a monkey. How would the nerve cells at the back of the brain encode the picture? Are visual images exposed onto the back of the brain the way images are exposed in a darkroom onto photographic paper? Dr. De Valois and colleagues came up with a surprising answer.

Russell De Valois: The techniques we used was to give the monkey minute quantities of radioactive sugar, which is taken up by the most active cells in the cortex. Then when we looked at the x-ray picture of the cortex, we could see which cells were most responsive to this pattern. And the amazing thing was the precision of this pattern that we saw in the cortex, where you can see all of the stripes and rings laid out beautifully on the cortical surface. We were really amazed at how precise the mapping was onto the cortex.

Narrator: Photographs on the brain? Not quite. Without doubt, Russell De Valois and his colleagues have shown that there is precision in the way individual nerve cells fire in response to what the eyes see. But this autoradiograph doesn't yet explain how those nerve cells actually decoded the image.

[music playing]

[applause]

To solve that mystery, Dr. David Hubel of Harvard has invested nearly three decades of research in pursuit of the answer. He and his colleague Dr. Torsten Wiesel made discoveries which won them the Nobel Prize. The brain surprised them. The way in which nerve cells of the primary visual cortex responded to an image flew in the face of their own intuition.

David Hubel: To start with, the cortex is a plate of cells a couple of millimeters thick and a foot or so in surface area. The part of the cortex that we're dealing with, the primary visual cortex, is a chunk of that about the size of maybe a credit card. And here's the real thing. This is an actual brain of a macaque monkey.

This is the primary visual cortex, spread out smoothly and with a tucked under part that we can't see. Here, we've made a cut in the cortex. And to see what the actual cortex looks like under a microscope, we can walk into this region and look to the left.

And you see on the next picture what that looks like. Here is the smooth surface of the primary visual cortex. This is the part that's tucked underneath. And now, you begin to see that it's a layered structure.

Some places, the cells are packed tightly. Other places, they're looser. And underneath every square millimeter of cortex, which is about that much, you have something like 100,000 cells.

Narrator: The researchers actually listened in to individual nerve cells firing in the anesthetized cat, as they presented it with different visual images.

David Hubel: When we started working, Torsten and I, in the late '50s, we set up our first experiments. And they didn't go well. Because at the beginning, we couldn't make the cells fire at all. We'd shine lights all over the screen. And nothing seemed to work.

And rather by accident one day, we were shining small spots, either white spots or black spots, onto the screen. And we found that the black dot seemed to be working in a way that at first we couldn't understand until we found that it was the process of slipping the piece of glass into the projector, which swept a line, a very faint, precise, narrow line, across the retina. And every time we did that, we'd get a response.

And even more than that, the line produced responses that swept across the screen in one direction but not in the reverse direction. Of course, that could have simply been an oddball cell. We didn't know whether we'd ever find another such cell.

But after some weeks or months, it became pretty clear that most of the cells that we encountered in the visual cortex demanded just that kind of stimulus, although from one cell to the next the orientation varied. A number of other things differed. The line was the important thing.

Narrator: Russell De Valois's findings do not necessarily contradict but extend the findings of Hubel and Wiesel. They imply that some nerve cells are sensitive to the spatial frequencies of light, as in the subtle effects of sunshine on water, shadows spangling up onto other shadows, a prismatic swirl of texture where edges do not seem to be the key factors.

Russell De Valois: It's very nonintuitive. But if you take a grading pattern and add it to another pattern, it still looks like just sort of a vague, random pattern. But as you continue to add patterns of different periodicity, different variations across space, different orientations, pretty soon it builds up into a more and more complex pattern— in fact, one that you can recognize as an identifiable human face.

Narrator: In the 19th century, impressionist painters sensed that such spatial frequencies might help create and enrich imagery. Whole canvases emerged out of formless points. These seemingly random dots show how spatial frequencies can operate, how clear images are sustained out of a vibration of color and light. But David Hubel is not convinced by this theory of vision.

David Hubel: It might seem as a real surprise that the cells of the striate cortex have this preference for lines. Because after all, it's obvious that we don't see lines all the time. We see objects. But if you think of it for a minute, you realize that objects themselves consist of light dark contours.

And what the visual cortex is doing is asking at any point in the visual world, is there a contour? And if so, what's the direction of that contour? So it really isn't lines that are being detected. It's contours and their directions. And that is probably sufficient information to tell you all you need to know when you look at a scene.

Narrator: The primary visual cortex, this brain area the size of a credit card, has been studied extensively. But laws about the way it decodes vision are still not easy to lay down. While we may surmise that orientation, edge, and frequency detection are indeed roles performed by some of these cells, there are many other cells whose function is still unknown.