Place learning the default option

The most important and new concept developed by O'Keefe and Nadel (1978) was that animals possess the capacity to identify a given position in space, on the basis of its relations with other positions. This abstract capacity was exemplified by adult rats in the well known Morris navigation task (1984), in which rats must reach an invisible submerged platform at a fixed position in a circular pool of opaque water. To meet the requirement for place learning in this situation, they must reach the escape position with help of distant information only, with little training, following a direct path from any starting position in the pool. In addition, the normal rats spend a significant amount of time searching on its exact position when the escape platform is absent. They can also memorize a new escape position in the same environment following one demonstration trial only. Rats with hippocampal lesions are unable to satisfy the third criterion and require a special training procedure to meet the two first conditions.

Interestingly, the very concept of place learning by rats appears to be counterintuitive, probably because anyone expects animals to be attached to local objects and not to behave on the basis of an abstract representation. However, it appears to be the default option of rats in any spatial task. If briefly trained to find a visible platform at a fixed position in space, they will mainly remember its spatial position for further escape trials, ignoring the visible cue if the latter was placed at another position in the pool (see McDonald et al., 2004, for a recent review).

This hippocampus dependent strategy appears to be the primary one in normal subjects, when they are not submitted to intense stress. During an exploratory phase in a stable environment, selecting simple features for later stimulus response associations could be regarded as an economic and simple strategy. However, such process would take time and require repeated exposures in the same environment. Thus, an early spatial representation is more likely to emerge from a global encoding based on distributed movement. The association of simple responses in a cumulative manner is more likely to occur during overtraining and in subjects with a dysfunctional hippocampus. Indeed, rats with hippocampal lesions which cannot process stimulus-stimulus relations from general "panoramas", are also more rapid than normal subjects in learning discrete stimulus response associations (Packard et al., 1989; McDonald and White, 1993; Packard, 1999). One can thus hypothesize that the two systems are antagonist in a new situation, where the dominant hippocampus promotes "spatial optimal foraging" strategies, disregarding the food reinforcement rule for which stimulus response strategies might be more efficient.

This suggests a hierarchy of orientation strategies depending on the stability of the environment and on the regularity with which an animal has been allowed to follow a given trajectory. For the squirrels retrieving nuts, the place where they had buried a nut is of higher relevance than the search for nut odor or other stereotyped behavior (Jacobs, 1995). A similar bias toward a spatial position was described in food caching birds trained to find nuts in salient feeders in an aviary (Brodbeck, 1994). In a more general perspective, this suggests that the organization of spatial orientation aims at overcoming local variability in an environment in which local cues are likely to change over time.

0 0

Post a comment