In the laboratory, rats traverse a device known as the radial arm maze, which simulates the pathways rats naturally create. One might assume that under these conditions the rats would need some encouragement, such as a food reward, to scour this environment as exhaustively as they would their natural habitats. But William Timberlake, a professor of psychology on the Bloomington campus, has found that rewards are unnecessary.
Timberlake and postdoctoral researcher Wesley White, who at the time was an IU doctoral student, tested the performance of three groups of rats on the radial arm maze with eight arms, or pathways, radiating from its central platform. One group of rats was kept hungry before the experiment, and each of these rats received a food reward the first time it reached the end of each of the eight arms. A second group of rats was also kept hungry before the experiment, but the rats from this group did not receive a food reward for reaching the ends of arms. A third group consisted of rats that were fed before the experiment. These rats did not receive a food reward.
The third group turned out to be the least efficient at exploring the maze, making more "mistakes"--that is, backtracking to a previously visited arm before entering all unexplored arms. Between the other two groups, however, there was virtually no difference in performance. "Whether food-rewarded or not, deprived rats traversed the radial arm maze efficiently," Timberlake and White reported. And even the third group chose previously unexplored arms at levels above chance.
"Rats are exceptionally good at this," Timberlake says. "Averaged over days, they can enter 7.5 arms out of 8 novel choices." The evolutionary advantages for rats in remembering where they have been is obvious, as Timberlake and White observed in the report that accompanied this study: "Considerable unrewarded locomotion is often required to find food. Rats that persistently explored their environments efficiently in spite of momentary food scarcity were probably more likely in the long run to ingest an adequate diet, survive, and reproduce."
But how do they do it? An early guess was that the rats rely on odor detection, Timberlake says. "If they smelled the results of putting their paws on the maze arm, and came back to making a choice again, they might sniff the front of the arm and say 'I've already been down that way.'" But field observers know that rats in the wild exhibit efficient foraging strategies under conditions that preclude odor detection--for example, after a heavy ice storm and snowfall have obliterated the rats' previously created paths.
Instead, Timberlake says, rats seem to direct their movements in relation to "large vertical stimuli, such as a fence post or tree." Even in the laboratory, he says, "they're orienting toward spatial cues," such as a doorway or window. Knowing that rats use visual cues to keep track of where they have been lends insight into the nature of the animal's capacity for short-term, or working, memory. Experiments with the radial arm maze, says Timberlake, further reveal that the rat's working memory is quite tenacious.
When removed from the radial arm maze in the middle of their search and replaced up to two hours later, rats can remember which arms they have previously visited. "This is a robust working memory, more robust than remembering phone numbers on a short-term basis in humans," Timberlake says. Considering the interruptions in foraging that a rat might experience in the wild--a rainstorm perhaps, or an invasion by predators--the rat's robust short-term memory is a valuable tool. As Timberlake observes, "Usually things that work extremely well with an animal in a laboratory setting have a real connection with something about the animal's ecology."