Neuroscientists locate area of brain responsible for 3-D vision
PASADENA--Researchers have found the brain circuitry that allows us to see the world in three dimensions even when one eye is closed.
In the current issue of the journal Nature, a team of neuroscientists at the California Institute of Technology reports that the middle temporal area (MT) of the brain renders certain visual motion cues into 3-D perceptions. Area MT is a small cortical area in each cerebral hemisphere located approximately an inch or two above the ears in both humans and non-human primates.
"We see the world in three dimensions even though the image in our retina is flat and in two dimensions," says Richard Andersen, who is the James G. Boswell Professor of Neuroscience at Caltech and principal investigator of the study, which was also conducted with postdoctoral fellow David Bradley and graduate student Grace Chang.
"So to derive the third dimension of depth, our nervous system has to process certain visual motion cues."
Andersen says that many people may assume that 3-D vision is explained solely by our having two eyes that provide a stereoscopic view of the world. But stereoscopic vision is fairly new in evolution, he says, while the depth-from-motion process is much more fundamental and primitive.
"In certain contrived situations, you can actually be better off by closing one eye," he says. For example, in a video animation his team created for the research project, an image of a cylinder is constructed entirely with points of light on a black field. When the cylinder is frozen, the viewer normally sees only a rectangular flat plane of light dots. But when the image is rotated, the viewer perceives a three-dimensional object.
"In this case, your stereoscopic vision may tell you the image is flat," he explains. "But area MT still overrides what you see with two eyes and gives you depth."
What's actually happening in the brain at such a time is a processing of the motions the eye perceives on the screen. While the spinning cylinder appears to have three dimensions, it actually comprises a series of dots that move horizontally across the screen at varying speeds. The dots near the edge of the cylinder image move more slowly across the screen, while the dots at the center move more quickly.
The brain picks up these varying speeds as natural motions in the world. The right and left edges of a cylinder naturally seem to move more slowly than the portion of the cylinder directly in front because the edges are moving forward and backward in that reference frame. And while there are no stereoscopic views in this display, the brain can still reconstruct the perception of depth using only the motions of the dots.
An especially important aspect of the study is the fact that viewers have a bias as to which direction they perceive the image to rotate, which changes spontaneously every few seconds. Because the cylinder is merely a group of dots moving at varying speeds, the image can appear to be rotating either clockwise or counterclockwise (see QuickTime movies below). Both human viewers and rhesus monkeys tend to perceive the cylinder as moving left to right (or counterclockwise), and then with time see it reverse.
"The beauty of this illusion is that the stimulus is always the same, but at different instances in time the perception is completely different," Andersen says.
"If the MT neurons were only coding the direction of motion of the dots in two dimensions, the cells would not change, since the physical stimulus never actually changes," he adds.
"So what we're actually watching is brain activity changing when an interpretation of the three-dimensional world changes," Andersen says.
Andersen says the research is aimed primarily at a fundamental scientific understanding of the biology of perception. However, the research could also eventually impact the treatment of human patients with vision deficits. In the very far future, the research could also perhaps be exploited for a technology leading to artificial vision.