Perception is mathematically impossible.

This might seem like a bold statement—after all, you are perceiving these letters right now—but it’s nonetheless true. Imagine a black-and-white line drawing of a cube on a sheet of paper. Although this drawing looks to us like a picture of a cube, there are actually an infinity of other three-dimensional figures that could have produced the same set of lines when collapsed on the page. But we don't notice any of these alternatives. Happily for all of us, our visual systems have more to go on than just bare perceptual input. They use heuristics and short cuts, based on the physics and statistics of the natural world, to make the “best guesses” about the nature of reality. Just as we interpret a two-dimensional drawing as representing a three-dimensional object, we interpret the two-dimensional visual input of a real scene as indicating a three-dimensional world. Our perceptual system makes this inference automatically, using educated guesses to fill in the gaps and make perception possible.

It turns out that our brains use the same intelligent guessing process to reconstruct the past, in addition to using it help perceive the world. Memory itself is not like a video-recording, with a moment-by-moment sensory image. In fact, it’s more like a puzzle: we piece together our memories, based on both what we actually remember and what seems most likely given our knowledge of the world. Just as we make educated guesses in perception, our minds’ best educated guesses help “fill in the gaps” of memory, reconstructing the most plausible picture of what happened in our past.

The most striking demonstration of the minds’ guessing game occurs when we find ways to fool the system into guessing wrong. When we trick the visual system, we see a “visual illusion”—a static image might appear as if it’s moving, or a concave surface will look convex.  When we fool the memory system, we form a false memory—a phenomenon made famous by researcher Elizabeth Loftus, who showed that it is relatively easy to make people remember events that never occurred. As long as the falsely remembered event could plausibly have occurred, all it takes is a bit of suggestion or even exposure to a related idea to create a false memory.

In the Blink of an Eye
In past literature, visual illusions and false memories have been studied separately. After all, they seem qualitatively different: visual illusions are immediate, whereas false memories seemed to develop over an extended period of time. A surprising new study blurs the line between these two phenomena, however. The study, conducted by Helene Intraub and Christopher A. Dickinson, both of the University of Delaware, reveals an example of false memory occurring within 42 milliseconds—about half the amount of time it takes to blink your eye.

Intraub and Dickinson’s study relied upon a phenomenon known as “boundary extension”, an example of false memory found when recalling pictures. When we see a picture of a location—say, a yard with a garbage can in front of a fence—we tend to remember the scene as though more of the fence were visible surrounding the garbage can. In other words, we extend the boundaries of the image, believing that we saw more fence than was actually present. This phenomenon is usually interpreted as a constructive memory error—our memory system extrapolates the view of the scene to a wider angle than was actually present. 

The new study, published in the November issue of the journal Psychological Science, asked how quickly this boundary extension happens. The researchers showed subjects a picture, erased it for a very short period of time by overlaying a new image, and then showed a new picture that was either the same as the first image or a slightly zoomed-out view of the same place. They found that when people saw the exact same picture again, they thought the second picture was more zoomed-in than the first one they had seen. When they saw a slightly zoomed-out version of the picture they had seen before, however, they thought this picture matched the first one. This experience is the classic boundary extension effect. So what was the shocking part? The gap between the first and second picture was less than 1/20th of a second. In less than the blink of an eye, people remembered a systematically modified version of pictures they had seen. This modification is, by far, the fastest false memory ever found.

Although it is still possible that boundary extension is purely a result of our memory system, the incredible speed of this phenomenon suggests a more parsimonious explanation: that boundary extension may in part be caused by the guesses of our visual system itself. The new dataset thus blurs the boundaries between the initial representation of a picture (via the visual system) and the storage of that picture in memory.

So is boundary extension a visual illusion or a false memory? Perhaps these two phenomena are not as different as previously thought. False memories and visual illusions both occur quickly and easily, and both seem to rely on the same cognitive mechanism: the fundamental property of perception and memory to fill in gaps with educated guesses, information that seems most plausible given the context. The bottom line? The work of Intraub and colleagues adds to a growing movement that suggests that memory and perception may be simply two sides of the same coin. 

Are you a scientist? Have you recently read a peer-reviewed paper that you want to write about? Then contact Mind Matters editor Jonah Lehrer, the science writer behind the blog The Frontal Cortex and the book Proust Was a Neuroscientist. His next book, How We Decide, will be available in February 2009.