How do our minds distinguish between what is important in the world around us and what is not? Research has long shown that when two cues both signal a reward, an animal will learn to associate the reward with only one of them. The cue that is most salient or that is learned first overshadows or blocks learning about the other; the less salient cue is redundant. The phenomenon, known as cue competition, is so well established in learning theory that it approaches law.
Now, however, a study in rats has revealed a striking exception to that law. In a series of experiments published in the October Journal of Experimental Psychology: Animal Behavior Processes (Vol. 27, No. 4), researchers have found that animals' learning about the shape of their environment--in this case, a pool from which they must escape--is not overshadowed or blocked by the presence of a vertical rod, or "beacon," that marks the escape location.
"We have, in learning theory, this extremely successful model of cue competition, which seems to work for everything from drug addictions to human causal reasoning," comments psychologist Sara J. Shettleworth, PhD, of the University of Toronto. "But now, it seems that processing of a geometric space doesn't work that way."
"That really did come as a surprise," says the study's lead author, experimental psychologist John M. Pearce, PhD, of Cardiff University in the United Kingdom, who in previous research found that a beacon overshadowed rats' learning about other, more distant landmarks.
Calling the new findings a "refreshing contrast," psychologist Ken Cheng, PhD, who studies spatial learning at Macquarie University in Australia, argues that the results suggest that "there's something special about how animals learn about geometry."
No sign of overshadowing
In a series of experiments to uncover whether learning about the shape of an environment is overshadowed or blocked by the presence of a beacon, Pearce and colleagues trained rats to escape from a triangular pool with one curved side by swimming to a submerged platform. After each of 36 training trials, the pool was rotated 90 degrees so that no extraneous cues, such as the noise of a room fan, could influence rats' ability to find the platform.
For one group of rats, a rod--or beacon--was attached to the platform, providing a landmark that the animals could use to find their escape. For a second group, there was no beacon, so that the only cue that the rats had to help them find the platform was its location relative to the geometry of the pool. For a third group of rats, dubbed the "random" group, the platform and a beacon were randomly moved from one location to another across the experimental trials, so that the animals had no opportunity to learn over time where the platform would be relative to the shape of the pool.
After the training sessions, the investigators tested the rats' learning of where the platform had been. The platform was removed from the pool, as was the beacon for animals that had been given one. The animals were put back in the pool, and the researchers measured how long they spent searching for the platform in the region of the pool in which it had been for the training trials.
In four experiments, the researchers found that the presence of a beacon did not block or overshadow the rats' learning about the shape of their environment: Rats that had been shown a beacon during training spent just as much time searching in the correct location as did rats that had not. Further, rats in both the beacon and no-beacon groups showed a greater preference for the correct location than did rats in the "random" group.
These results indicate that animals that had been trained with a beacon nonetheless learned about the shape of the pool and were able to put that knowledge to use during testing.
"The tried-and-true principle of cue competition, which applies so well to standard associative learning tasks, and has also been shown to apply to some spatial learning tasks, doesn't apply when it comes to learning about shape," concludes Pearce. "This suggests that there's something rather special about spatial learning based on the shape of the environment. The exciting question now is, what is special?"
Compass and ruler
The new findings are consistent with the notion, proposed by Cheng, that animals' brains possess a specialized "geometric module" for processing information about the shape of an environment. He and others have published research bolstering that idea, but have not shown conclusively that geometric learning is not overshadowed or blocked by other cues.
Pearce notes that the results also support a different proposition: that the shape of their environment provides a frame of reference by which animals can judge the spatial relations--distances and angles--among objects within the environment.
This geometric frame of reference, Pearce suggests, might be "a compass and a ruler," by which animals can judge distances and angles between objects. In this model, then, the geometry of one's environment operates not independently from landmarks and other spatial cues, as Cheng has proposed, but as a tool for aiding navigation from one object to another.
Pearce favors the frame of reference notion, in part because in some of his experiments, the beacon not only did not overshadow but actually facilitated learning based on geometry.
While the generality of the new findings must still be tested, the implications for brain function could be significant, says study co-author Mark A. Good, PhD, a behavioral neuroscientist at Cardiff University. In preliminary studies, he and his colleagues have found that rats with damage to the hippocampus, a brain area known to be involved in spatial learning, are impaired in their ability to navigate using the shape of their environment. Further research, they hope, will help elucidate how the brain--and in particular, the hippocampus--processes shape information.This article is part of the Monitor's "Science Watch" series, which reports news from APA's journals.