| |
Abstract:
The Morris water task (MWT) has been used to study spatial
learning in rat from systems to molecular levels. MWT performance
depends upon the integrity of hippocampus and initial acquisition
depends upon hippocampal synaptic plasticity. In a computerized
version (vMWT), humans learn to locate a hidden goal using distal
cues in a manner similar to rats. Hippocampal dependency of vMWT
learning has been demonstrated recently with lesion and functional
neuroimaging approaches (Astur, et al., 1998). We contrast basic
associative and cognitive mapping views of this form of spatial
learning. According to the latter, such learning occurs in an
automatic, all-or-none manner and involves creating and updating a
unified spatial representation in response to environmental change.
The former holds that place learning should accrue incrementally
and exhibit associative phenomena such as blocking. To test these
views, we measured blocking in vMWT learning. Three sets (A, B, C)
of four distal visual cues were used. Each set unambiguously
predicted the goal location. The experiment was conducted in three
phases: I. Pretraining with set A, C or no pretraining; II.
Training with sets A and B; III. Probe with set A or B. Blocking
was evident in poorer performance by the set A pretraining group on
standard learning measures when probed with set B. The results show
that vMWT performance relies on associative learning rules.
|
