The University of Iowa
The University of Iowa News Services Home News Releases UI in the News Subscribe to UI News Contact Us

 

CONTACT: MARY GERAGHTY
100 Old Public Library
Iowa City IA 52242
(319) 384-0011; fax (319) 384-0024
e-mail: mary-geraghty@uiowa.edu

Release: Aug. 25, 1999

(NOTE: Embargoed until 1 p.m. CDT Aug. 25 in accordance with Nature press guidelines.)

UI researchers show rapid shifts of attention in human brain

IOWA CITY, Iowa — Finding a friend in a crowd. Searching for a nickel in a pile of coins. Finding a jar of mustard in the refrigerator. These are examples of visual search tasks, and two University of Iowa researchers have made an important step forward in understanding how these tasks are performed by the human visual system.

Geoffrey F. Woodman, a UI graduate student, and Steven J. Luck, a UI associate professor of psychology, have found that when humans look at collections of objects, attention is focused on a single object at a time and moves from one object to the next in about a tenth of a second. Their research is published in the August 26 issue of the journal Nature.

Research has shown that the more complex a visual image is, the longer it takes for the human brain to process the whole scene. Scientists have long debated the reason for this, with some arguing that the longer processing time indicates serial processing and others believing that it indicates parallel processing.

As an analogy, Luck said that serial processing is like a construction company with 100 employees who work together to build one house at a time, whereas parallel processing is like a company that works on 10 houses simultaneously with 10 employees working on each house. Similarly, the human brain might devote all of its capacity to one object at a time and shift rapidly from one object to the next, or it might process all of the objects in a visual scene simultaneously, with only a fraction of the brain’s capacity devoted to each object. In the construction example, it would take twice as long for either company to build 20 houses as to build 10 houses. In a visual search, it would take twice as long for either a serial process or a parallel process to scan 20 objects as to scan 10.

Woodman and Luck are the first to show definitively that the human brain processes complex images serially -- attending to only one object at a time but shifting rapidly from object to object.

"This may seem counterintuitive because it doesn't feel like we perceive only one object at a time," Luck said. Although it may seem that when you look at a scene you are seeing the whole picture, each object in the scene commands your attention in rapid succession, Luck said.

Woodman and Luck demonstrated serial processing by having subjects search for a specific object in a display containing many other objects. Subjects were shown a group of squares and told to search for the one with a gap in the left side. Most of the squares were black, but the picture also included one red and one green square on opposite sides of the image. The subjects were told that the target square probably would be red but might be green. Luck explained that this was done to bias the subjects to search in a specific order.

In this experiment, Woodman and Luck used EEG recordings, which monitor brain waves in the same way as an EKG monitors heart activity. To track attention with the EEG, they measured a brain wave that indicates where attention is focused. They found that attention, as indicated by this brain wave, shifted rapidly from the red square on one side of the image to the green square on the other side, which demonstrates serial processing.

Woodman compared this type of visual activity to searching for a piece of a jigsaw puzzle among a pile of pieces that will make up the completed picture. You can ignore pieces that are the wrong color, just as the subjects in his experiment ignored the black squares. But once you find several pieces that are the right color, you must focus your attention on them individually to see if they also have the right shape.

Luck said that this new research will contribute to the long-standing quest to understand how human vision works, which may someday lead to computerized visual systems. But, he cautions, there is still a long way to go before we will be able to create robots or other forms of artificial intelligence that can match human vision.

"A huge proportion of the brain is devoted to vision," Luck said. "We have so much brain power devoted to vision that perceiving scenes seems easy, but it is really an incredibly complex process."

For more information, Luck is available by phone at (319) 335-2422 or by email at steven-luck@uiowa.edu and Woodman is available by email at geoff-woodman@uiowa.edu