IOWA CITY, IowaSince the emergence of machine vision in the 1960s, debate has raged over whether a parallel or serial architecture is best. Researchers modeling visual processes in the brain observed parallelism in neural structures, but didn't know enough about how visual information was being represented to resolve the issue. Now University of Iowa researchers say they've solved this vision research question: Does the brain operate in parallel or serially?
"We are the first research group to show definitively that the human brain processes images serially-paying attention to only one object at a time and shifting rapidly from object to object," said University of Iowa professor Steven Luck. According to the new insights, the brain does perform many tasks in parallel, such as muscle coordination for walking in the park while simultaneously listening to birds chirping. These are cognitive operations that involve separate processing on different types of data.
For such diverse tasks it is clear that the brain does operate in parallel. But when it comes to tasks involving similar data items, the brain appears to time-division multiplex, that is, focus its attention on one object at a time so quickly that the conscious mind is not aware of it. "It's counterintuitive because it seems to our conscious mind that we are comparing objects simultaneously, but we now think that the brain's parallelism is similar to a computer's-that is, a computer has millions upon millions of simultaneously acting transistors, but at the functional level it is operating serially-one instruction at a time," Luck said. The new theory says that the brain operates the same way at the functional level; it processes information serially, even though the underlying neural hardware is operating in parallel.
Luck was able to determine whether the brain's processing was parallel or serial through an experiment he performed in 1994. This experiment identified a pattern in brain waves known as N2PC, which stands for the second negative peak (N2) of the posterior contralateral (PC). The N2PC identifies the location of brain waves as emerging from either the right or left side of the brain. By arranging the experimental situation, Luck was able to use N2PC to identify whether a person was processing visual signals one at a time or simultaneously.
He enlisted the help of graduate student Geoffrey Woodman to perform the experiment and study the collected data. The experimental setup presented to subjects a landscape-shaped display of different colored blocks, most of which were black except for a red block on the left side and a green block on the right. The subjects were instructed to find the block with a nick in it and told that it was probably red but could be green. Those instructions allowed subjects to either process all blocks in parallel, focus their attention on just the red and green blocks simultaneously or search for the correct block in the same order each trial-that is, red, then green, then black.
"It was important that we knew the order in which they paid attention to the colored objects, because the N2PC works by correlating the brain waves coming from each side of the brain over many statistical trials, so we had to always have them search in the same order," Luck said. By observing the brain activity of the subjects performing the search and recognizing tasks using N2PC, Luck and Woodman discovered that the brain turned its attention from one block to the next at intervals of about 1/10th second. "There wasn't a single subject that did the task in parallel," Luck said.