The brain scans simultaneously measured neural activity via blood flow in thousands of locations across the brain. Researchers used regularized linear regression analysis, which finds correlations in data, to build models showing how each of the roughly 50,000 locations near the cortex responded to each of the 935 categories of objects and actions seen in the movie clips. Next, they compared how much of the cortex was devoted to detecting humans or vehicles depending on whether or not each of those categories was the search target.
They found that when participants searched for humans, relatively more of the cortex was devoted to humans, and when they searched for vehicles, more of the cortex was devoted to vehicles. For example, areas that were normally involved in recognizing specific visual categories such as plants or buildings switched to become tuned to humans or vehicles, vastly expanding the area of the brain engaged in the search.
"These changes occur across many brain regions, not only those devoted to vision. In fact, the largest changes are seen in the prefrontal cortex, which is usually thought to be involved in abstract thought, long-term planning and other complex mental tasks," Cukur said.
The findings build on an earlier UC Berkeley brain imaging study that showed how the brain organizes thousands of animate and inanimate objects into what researchers call a "continuous semantic space." Those findings challenged previous assumptions that every visual category is represented in a separate region of visual cortex. Instead, researchers found that categories are actually represented in highly organized, continuous maps.
The latest study goes further to show how the brain's semantic space is warped during visual search, depending on the search target. Researchers have posted their results in an interactive, online brain viewer.
|Contact: Yasmin Anwar|
University of California - Berkeley