Wallpaper for the mind

Discover, January 31, 1995

In 1994, America became addicted to autostereograms–those swatches of psychedelic wallpaper that dissolve into three-dimensional images when you stare at them cross-eyed long enough. What the slack-jawed millions may not have realized, though, as they stared at books and posters, is that they were experiencing an enduring mystery of neurology: When the brain perceives a 3-D object, which comes first, the object or the 3-D?

The mystery has its roots in a previous 3-D craze, back in the nineteenth century. Victorian researchers discovered that they could create a 3-D illusion if they took photographs of an object with two cameras a few inches apart and had a person look at the images through a stereoscope, which allows each eye to see just one of the photos. The first stereograms were both a commercial smash and a neurological breakthrough: scientists realized that depth perception arises from the way the brain compares signals from the two eyes, which see an object from slightly different angles.

For over a century researchers assumed that the brain needed to recognize the signals as an object before it could compute the object’s 3-D shape. In 1960, however, Bela Julesz, a psychologist now at Rutgers University, challenged that idea with a new kind of stereogram made of two identical fields of randomly scattered dots. In each field he drew an imaginary square around some of the dots and shifted them slightly to one side, filling in the blank gaps with more dots. If you looked at either field alone, you couldn’t see the square. But when Julesz put both of them into a stereoscope, people saw a dot-covered square floating in front of a similar background. He and others concluded that depth perception is one of the first things the brain extracts from the visual signal, by comparing the left-eye and right-eye images dot by dot. Object recognition must come later.

Autostereograms, which can be viewed without a stereoscope, were invented in 1979 by psychologist Christopher Tyler. They consist of repeating vertical strips, like wallpaper, in which the pattern elements–random dots or something more complicated–have been shifted to one side, a la Julesz. They’ve been shifted in such a way that when you look at the pattern cross-eyed (or in some cases look “through” it), the neighboring strips overlap. The pattern elements then fuse into left-eye and right-eye images of a single hidden object, which appears to be floating in space.

Although Bela Julesz can be considered the grandfather of the autostereogram craze, it turns out he wasn’t entirely right about how the brain perceives 3-D objects. Vilayanur Ramachandran of the University of California at San Diego has shown as much by making a stereogram out of an optical illusion. The illusion consisted of three circles, each of which had a wedge cut out of it, arranged so the wedges formed the corners of an imaginary triangle. Even though there are gaps in the sides of the triangle, you see it whole because your brain fills in the gaps. Ramachandran created left-eye and right-eye versions of this illusion and had people look at them through a stereoscope. The people saw the illusory triangle floating in 3-D–even though the gaps prevented them from making the point-by-point comparison of left and right images that Julesz thought was essential to depth perception.

Ramachandran thinks the object-recognition and depth-perception regions of the brain may work in tandem, bouncing signals back and forth. That might explain the sensation some people have when looking at autostereograms: when they start to see the outline of a hidden object, the 3-D illusion suddenly kicks in. “The regions may be like two drunks,” says Ramachandran. “Neither one can make it down the street alone, but if they lean up against each other, they stay upright.”