Brain warps scenes to cancel out perceptual delays
You’re living in the past. Not because of nostalgia, but because of delays in your visual system. It takes at least one-tenth of a second for signals from your eyes to travel to the visual areas of your brain and rise to perception.
As a result, the image of the world you see is already outdated. When you’re moving through the world, you’re not seeing things where they actually are.
How the brain deals with these delays has been a subject of fascination and research in the laboratory of Dr. David Eagleman, assistant professor of neuroscience at Baylor College of Medicine.
This week, in Frontiers in Psychology, Eagleman and his student Don Vaughn report a new discovery which may unmask the mechanisms by which the brain strives to perceive a more up-to-the-second view.
It had been previously proposed that the brain might be able to deal with delays by warping a moving scene, making it look more like it should if there were no delay. This “perceiving the present” hypothesis, first postulated by evolutionary biologist Dr. Mark Changizi, suggests an explanation for various visual illusions. For example, in the Hering illusion, two straight, parallel bars look bent when presented against a background of radial lines. Changizi proposed that the background has the effect of fooling the brain into thinking it is moving forward. As a result, the visual system “warps” the bars so they appear as they will in the next moment.
To understand why the bars look bent, Eagleman suggests imagining driving on the Golden Gate Bridge.
“As you approach a pair of pillars, they move rapidly apart at eye level. But their distant tops don’t move apart as rapidly, because they are farther away from you.”
In other words, when you move forward, points closest to the horizontal plane change their relationship to you the most quickly. So if the brain wanted to extrapolate how approaching parallel bars would look in the next moment, it would have the middle of the bars bend apart the most.
Perceiving the present
Eagleman and Vaughn found the “perceiving the present” hypothesis intriguing, but there was little direct data to support it, and no model for how the brain might accomplish such a task, biologically speaking.
To put the hypothesis to the test, they asked participants to watch a flowing “starfield”—hundreds of white dots moving from the center of a screen to the edges. (Think of the way stars look in Star Trek when the Enterprise moves into warp drive.) The starfield gives the brain the impression that it is moving forward through a scene.
The researchers then flashed two bars on top of the starfield, and discovered that the bars appeared bent, just as in the classical version of the illusion. To quantify the illusion, they flashed the bars repeatedly, and had participants use two buttons to adjust the curvature of the bars until they appeared straight. When the bars looked straight to the viewers, they were not actually straight—and this is what allowed the researchers to know how much curvature in the opposite direction was needed to cancel out the powerful illusion of bending.
The important discovery came when Vaughn and Eagleman had the dots flow the other direction—inward—as though the viewer were receding instead of moving forward. In this case, straight bars flashed on the screen looked bent outward, exactly the same size and direction as they had in the first experiment.
“In other words,” said Vaughn, “the illusion was exactly the same no matter which direction the dots flowed. This was an unexpected surprise.”
This was the clue the researchers needed to understand how the brain was accomplishing the warping.
“By showing that the illusion was the same whether the dots were flowing forward or backward, we could conclude that the spatial warping is not an active, sophisticated computation,” Eagleman said. “Instead, the brain uses a simple mechanism of motion detection—irrespective of direction—and warps the scene under the assumption of forward motion. Except in the laboratory or the back of an old station wagon, this assumption works well most of the time.”
This detail allowed Vaughn and Eagleman to narrow down the mechanism of spatial warping to a part of the brain called the primary visual cortex, or V1, a region which can detect motion but is agnostic about which direction something moves along a line.
“As a result, perceiving the present is something the brain gets right much of the time but not all of the time,” Vaughn said.
Eagleman added: “Rather than run expensive computations on the fly, the brain appears to bake some assumptions into the circuitry—such as forward motion—allowing a simple mechanism of warping the scene to do most of the heavy lifting.”
As a result, the brain can mitigate neural delays under normal circumstances. “Our findings put a new twist on a 152-year-old illusion, and narrow down the brain mechanisms behind the feat,” said Eagleman.
For a demonstration of the “dynamic Hering illusion” with the flowing starfield, see www.eaglemanlab.net/hering.
Vaughn is now at the University of California Los Angeles.
Funding for this work came from the National Institutes of Health grant to David Eagleman (R01NS053960).