Dissociating Implicit and Explicit Memory Retrieval Processes: Fmri Evidence That Conceptual Priming Does Not Support the Contribution of Familiarity to Recognition Memory.

Abstract:
Prior exposure to stimuli modulates different brain networks depending on the memory test employed. Lateral and medial parietal, dorsal middle frontal gyrus and anterior prefrontal cortex respond more for old than new items, reflecting 'explicit retrieval success'. By contrast, left ventral and dorsal inferior frontal gyrus and left temporal cortex respond less for old than new items, reflecting 'conceptual priming'. We demonstrate a functional dissociation between these networks. At study subjects made abstract/concrete judgements to words. At test, old/new recognition (episodic task) and abstract/concrete judgements (semantic task) were performed in separate blocks. Only task instructions differed across blocks; test items were equivalent. Event-related procedures were employed within the blocks to detect differences between old and new item. Results revealed that priming effects were only present during the semantic memory task, whereas retrieval success effects were present during both tasks. This dissociation provides several important functional-neuroanatomical constraints on cognitive accounts of memory retrieval. First, processing within 'retrieval success' regions is relatively obligatory rather than being driven by task-demands. Second, priming effects appear to reflect a form of 'transfer appropriate processing'. Third, priming effects were not present during, and presumably did not contribute to, successful retrieval during the episodic memory task. These findings strongly suggest that conceptual priming does not support the contribution of familiarity to recognition memory judgements.