No abstract

No abstract

Differential responding to changes in the stimulus situation, long central to the concept of stimulus control, also provides the implicit conceptual basis for assessing the nature of a variety of associative relationships. However, there is substantial evidence that the perception of stimulus similarity is not a static property. Generalization gradients to contextual as well as discriminative stimuli flatten over time, and this increase in perceived similarity presumably reflects forgetting of the detailed characteristics or attributes of stimuli. Methodologically, the flattening of the gradient imposes an important constraint: The effect of a stimulus shift will be highly sensitive to the length of the delay interval between training and testing. Conceptually, the loss of memory for stimulus attributes also implies that the sources of interference in retention can increase over time.

Old memory, when reactivated by cue exposure, was disrupted by mild or deep hypothermia treatments. New memory was impaired only by deep cooling. Moreover, old but not new learning showed spontaneous recovery. Old reactivated memory may be qualitatively different from newly acquired memory.

This review briefly traces some of the history of the phenomenon of what has come to be called “reconsolidation.” The early findings of retrograde amnesia for an old but reactivated memory led to several interesting but largely behaviorally oriented studies. With only a few sporadic exceptions, research in the area languished until about 2000, when several articles caught the attention of the neuroscience community and led to a number of studies examining the phenomenon at several different levels of analysis. We consider several of the current issues generated by those studies, present a retrieval based model that may account for some findings, and indicate some possible new directions on this topic.