A Representational Account of Memory and Perception Across the Lifespan

The representational hierarchical model suggests that regions of the medial temporal lobes (MTL) differ in terms of the representational content they support (i.e., the perirhinal cortex represents objects, while the hippocampus represents scenes). The model posits that these representations are useful for all cognitive functions – be it memory or perception – and predicts that stimulus-specific processing deficits will emerge based on the neuroanatomical locus of dysfunction. Thus, the goal of the present dissertation was to examine (1) stimulus-specific processing deficits in older adults at-risk for mild cognitive impairment (MCI) and in healthy older adults, and (2) whether manipulations that impact memory also impact perception in young adults. In Chapter 2, we hypothesized that cognitive status related to MCI would result in object-specific memory decline, based on past work demonstrating that MCI pathology originates in the perirhinal cortex. We assessed recognition of objects and scenes in older adults and correlated recognition performance to cognitive assessment scores. As predicted, we found that object recognition showed a stronger relationship to cognitive status than did scene recognition. In Chapter 3, we hypothesized that healthy older adults would show scene-specific associative memory deficits based on putative hippocampal dysfunction. We examined associative memory for object-and-scene pairs and found that older adults were not able to use scene contexts as effectively as young adults to recall associated objects. In Chapter 4, we examined whether context reinstatement would affect both object memory and object perception, as predicted by the representational hierarchical model. We found that when object memory was tested within the same scene as study, object recognition was boosted. Critically, we also found that subsequent object perception was improved, even in the absence of a scene. The results of these experiments support representational accounts of memory by demonstrating stimulus-specific decline resulting from putative neural dysfunction. Moreover, they add novel support to the idea that stimulus representations are impacted in similar ways across different cognitive domains. Thus, I show that examining cognition by appealing to representational content instead of cognitive taxonomy is a powerful way to elucidate a broad range of novel empirical phenomena.Ph.D