An analysis of within-compound associations in spatial learning

Abstract

The nature of spatial learning has been argued to be qualitatively different from that of associative learning. Compelling evidence for this argument is provided by experiments showing a lack of typical associative cue competition between spatial and non-spatial cues. However, this lack of cue-competition is also evident in wholly nonspatial experiments and has been explained by the presence of within-compound associations: an associative phenomenon. This thesis aims to determine whether such
associations can explain similar cue-competition failures in spatial learning.
In a series of experiments it is shown that these within-compound associations exist between spatial and non-spatial cues in the rat, and that they can account for the frequent failure to observe typical cue-competition between these cues. In addition, it is demonstrated that the extent to which this failure occurs depends upon the relative
salience of the cues in question.
In related experiments, it is also shown that these within-compound associations between spatial and non-spatial cues exist in humans. However, manifestation of these associations appears to depend on the gender of the participant, with associations forming in males but not in females. Further experiments suggest that this difference is
likely due to the fact that the females are much less able to learn about the spatial cues in question.
It is argued that spatial learning need not be qualitatively different from associative learning if such associative phenomena as within-compound associations are accounted for.