The Hippocampus: Where a Cognitive Model meets Cognitive Neuroscience

摘要: The Hippocampus: Where a Cognitive Model meets Neuroscience Bradley C. Love (love@psy.utexas.edu) and Todd M. Gureckis (gureckis@love.psy.utexas.edu) Department of Psychology, University Texas at Austin; 1 Station A8000; Austin, TX 78712 USA Models goal the present work is explore possible mappings between an existing model from cognitive psychology functional brain regions. There are numerous these somewhat different levels analysis. For Supervised Unsupervised STratified Incremental Network (SUSTAIN; Love, Medin, & Gureckis, 2004; Sakamoto in press) model, mapping straightforward: aspects appear to map onto structures brain. SUSTAIN holds that humans represent category information terms natural bundles information, referred as clusters. example, knowledge mammals might be represented by several clusters (e.g., primates, four-legged mammals, whales, bats, etc.). posits learners form new response surprising events, such when child first told whale mammal not fish. In this poster, we will focus on SUSTAIN’s cluster formation process. Our hypothesis healthy intact hippocampus necessary for forming support cortical learning temporal lobe (cf., Gluck Myers, 1993). Forming can seen constructing conjunctive codes. A wide variety tasks rely codes episodic memory (a conjunction item context), sequence (item position), list discrimination list), relations item). All heavily (see Brown Aggleton, 2001, review). Assuming reduced ability clusters, has been able developmental trends infant (hippocampus fully developed) performance amnesiacs with hippocampal lesions (Gureckis 2003). Rules Exceptions: An Aging Study account function predicts normal aging disproportionately affect exception items rule-plus-exception classification studies. To master exception, must recruited encode it, despite fact similar or likely already exist memory. As age, accumulation cortisol released stressful events differentially leads atrophy reduceas functioning (Lupien et al., 1998). study design, three B followed simple rule if large, then A. small, B.). ran counter rules. older adults one B, leading increasing application insensitivity old vs. novel rule-following age. contrast, younger recruit each storing them apart rule- following items, which predictions those population. Human Results Thirty-seven undergraduates thirty- seven (51-84 years-old, mean=67.9) Austin VA outpatient clinic participated study. SUSTAIN’S held. Only subset results reported here. phase, type (rule exception) population interacted exhibited smaller difference accuracy (.27 .61) than did population, F(1, 72) = 35.39, MSe 1.09, p < .001. test phase negatively correlated (r=-.38 -.72, respectively), whereas correlations were positive (r=.52 .49, respectively). transfer, subjects made consistent responses studied all about same rate, .70 .69, t 1, applied more frequently (.88 .77) examples, t(36) 4.99, References Brown, P., J. P. (2001). Recognition memory: What roles perirhinal cortex hippocampus? Nature Neuroscience, 2, 51-61. Gluck, A., (1993). Hippocampal mediation stimulus representation: computational theory. Hippocampus, 3, 491-516. T.M., B.C. (2004). Common Mechanisms Infant Adult Category Learning. Infancy, 5, 173-198. B.C., D.L, T.M SUSTAIN: Psychological Review, 111, 309-332. Lupien, S.J., DeLeon, M, DeSanti S, Convit A, Tarshish, C., Nair, NPV, McEwen, B.S., Hauger, R.L., Meaney, (1998). Longitudinal increase during human deficits. 69-73. Sakamoto, Y., B. (in Schematic Influences Learning Memory. Journal Experimental Psychology: General.