Constraints on water maze spatial learning in rats: implications for behavioral studies of brain damage and recovery of function.

In an effort to develop spatial learning tasks not requiring food or water deprivation for use in studies of recovery of function after brain damage, T-maze spatial alternation learning was examined in intact rats using water maze swim-escape procedures. Consistent with previous studies, rewarded spatial alternation involving food or water deprivation was readily learned by intact rats. However, none of the groups of rats trained in the swim-escape tasks learned to alternate goal arm choices in the water maze at reliable rates. This was true regardless of whether non-correction or correction procedures were used, and regardless of intertrial delay intervals. Although average alternation rates over sessions did increase from chance levels, the majority of the rats did not reach criterion levels, even with as many as 38 consecutive days of testing. In contrast, a conditional spatial alternation task in the water maze, using a win-shift procedure, was readily learned. Surprisingly, a win-stay version of this conditional spatial task was not learned over 21 days of testing. These unexpected constraints on spatial learning and memory processes in rats cannot be attributed simply to failure of spatial information processing, nor to strict limitations on working memory in swim-escape tasks, since excellent spatial navigation abilities have been documented, and mastery of at least some working-memory tasks have now been demonstrated in swim-escape tasks.