How the Brain "Fills In" the Blanks
April 25, 1991
Hold for Nature embargo: April 25, 1991
Media Contact: Warren R. Froelich, 534-8564
VISION: HOW THE BRAIN "FILLS IN" THE BLANKS
How do we see the world without confronting "blind spots" at every turn?
From a purely biological sense, our vision should be continually interrupted by blind spots since the place where the optic nerve is inserted into the eyeball is devoid of light receptors -- it is completely blind.
Yet, we do not see a "hole" in the world corresponding to this spot. Why?
According to a series of experiments led by Vilayanur Ramachandran, a physician and psychology professor at the University of California, San Diego, the human brain performs some neurological sleight-of-hand and fills in the blanks.
"The brain apparently does a statistical estimate of whatever is in the surround and uses that information to fill in the gaps," said Ramachandran.
In his experiments, published in the April 25 issue of the British science journal Nature, Ramachandran demonstrates how the brain actively creates a neural representation of the surrounding visual field to fill in the blind spot.
"There are two theories," Ramachandran said. "One is that you just ignore what is going on there. The reason you don't see a hole or a gap is that the brain simply doesn't notice it.
"The other view is that the hole is there, but it is in some loose sense filled in with what's around it in some sense."
Ramachandran's studies, which involve the use of computer- generated images, suggest that the latter hypothesis is correct.
Other scientists participating in the study include Will Aiken, an undergraduate psychology student at UCSD; and R.L. Gregory, an internationally renowned psychologist at the University of Bristol, United Kingdom.
Ramachandran said his studies could help psychologists and others interested in neurological function better understand how the brain processes visual information to create images of the world.
The experiments also could offer clinical insights into certain eye diseases, including glaucoma where the majority of such patients suffer some form of scotomas -- regions in the visual field where things are invisible.
Other researchers, including those interested in robotics, might use the studies to help create so-called intelligent machines capable of seeing visual objects.
"You don't have a mechanism just to fill in the blind spots," explained Ramachandran. "It's a mechanism that has evolved to assign a surface in between borders. That is, once you discover the borders of an object like a chair, you don't have to bother with all the fine details. You just say if the stuff is here, it's going to be the same stuff all over the object.
"It's this process you are tapping into when you are studying the blind spot," he added, "so this may give us clues into how the brain does surface interpolation, and then we can think of applying the rules to robotics."
In one example from the study, subjects were shown a television screen filled with snowy interference, except for a small empty square-shaped image. The subject was instructed to stare at the square, while the snow danced around the screen. After several seconds, the square seemed to disappear and the place occupied by it was filled in with the visual background noise, or "twinkling."
In another experiment, several people were shown a snowy television transmission with a red background and a small grey square. After staring at a tiny dark spot on the image, the square appeared to turn red and then became filled with the snow.
Ramachandran said the first example suggested that the brain cells responsible for seeing the square "became fatigued." When this occurs, the cells of the brain responsible for "filling-in" take over. These cells may have large "receptive fields" that overlap or straddle the borders of the square.
The results of the second experiment imply there are two sections of the brain's visual cortex involved in this uncanny process, he added. The first is responsible for color and form; the second is concerned with motion.
Ramachandran's studies further suggest that the brain's ability to fill-in-the-blanks may be clever, but also is limited.
In one visual demonstration, for example, subjects were shown a piece of paper containing a straight line with a circle in the middle and a small black dot a few inches to the side. With the right eye shut, the subject was asked to fixate on the small black dot while moving the picture to and fro. In this case, the circle seemed to disappear, leaving behind the straight line -- the brain seemed to complete the line across the blind spot.
In a second demonstration, the corner of a square was positioned inside the blind spot. However, no matter how the subject moved the paper, the corner remained incomplete.
"The brain is not clever enough to say look, this has three corners and the fourth corner must be there," said Ramachandran, "but if you have a line, then it can form an estimate that there is a line there.
"So there are clear limits to what the visual part of the brain can do."
(April 25, 1991)