Rethinking the neuroanatomical organization of cognition: Recognition memory in visual cortex

项目总结

PROJECT SUMMARY By 2035, the number of US residents over 65 years will surpass the number under the age of 35 years, totaling 20% of the US population. Meanwhile, by 2050 the world population is projected to increase by 17%, to include 1.6 billion people aged 65 or older. Progressive deficits in memory are a common fact of aging, in addition to being a primary symptom of Alzheimer's disease. Yet, treatments for memory disorders such as seen in aging and Alzheimer's are limited. One obstacle is that we still have not attained a thorough understanding of how brain systems are functionally organized to produce memory in the healthy brain. Based on case studies like patient H.M., traditional accounts of cognition explain memory and perception in terms of processes computed by separable brain regions (or distinct brain networks). In contrast, the Representational-Hierarchical (R-H) account posits that a brain region can be engaged in any cognitive task, if it contains optimal representations for solving the task. The specific objective of this proposal is to test my overarching hypothesis that brain regions contribute to a cognitive task – be it perceptual or mnemonic – according to the representations they contain. The R-H Theory postulates that the organization of cognition (i.e., memory and perception) in the brain is better understood in terms of a hierarchical continuum of representations, as opposed to anatomical modules (or circumscribed "networks") for distinct cognitive processes like "episodic memory". The central hypothesis of this proposal is that declarative memory may be computed in sensory cortex without involving the medial temporal lobe (MTL). Aim 1 will determine whether recognition memory can be supported by structures outside of canonical memory regions (i.e., MTL), in brain areas known to mediate perception (e.g., visual cortex). In this fMRI study, I will test participants' (long-term, declarative) recognition memory for stimuli they have studied. I hypothesize that if to-be-remembered items are abstract visual objects comprising conjunctions of visual features, and if the task requires recognition of familiar versus novel conjunctions of features, then the brain regions supporting memory will fall outside of MTL, in visual cortical sites such as lateral occipital cortex. Aim 2 will test whether aging (associated with incipient MTL deterioration) leads to greater deficits in recognition memory for high-level associative stimuli than for low-level visual stimuli. I will use the visual stimuli from Aim 1, along with analogously constructed word-pair stimuli. We assume that MTL networks deteriorate in aging. My working hypothesis, derived from the R-H Theory, is that older adults will perform worse for word pairs than visual stimuli, relative to younger adults, because word-pairs (but not visual stimuli) require high-level representations, which are MTL-dependent. This proposal has translational and theoretical significance. The results will determine whether the R-H Theory provides a better explanation of how human brains give rise to cognition. Both aims may lay the groundwork for improved diagnosis and treatment of memory disorders in aging and Alzheimer's, potentially alleviating patient suffering and familial burdens.