IBN 98-16612 BEST (COLLABORATIVE with MINAI) The hippocampus is known to be involved in the performance of crucial cognitive tasks such as memory and spatial representation. A very large body of experimental data is now available for the hippocampus, but the interpretation of these data requires the formulation of theoretical models for hippocampal function. This project applies a computational modeling approach to a very important but poorly understood aspect of hippocampal function. Experimental data suggest that representations of place (where an animal is located with respect to its surroundings) in the hippocampus are not based solely on sensory features (e.g., visual cues), but also reflect a cognitive assessment of location. Thus, two visually identical environments can be identified as different locations based on contextual information. The computational mechanism by which this context-dependence arises is not clear. The present project addresses this issue by building on the hypothesis that two sub-regions of the hippocampus --- the dentate gyrus and the hilus --- are crucial to the emergence of context- dependent representations in the hippocampus. Over its duration, the project will comprise four tasks: (1) Implementation, simulation, and analysis of an abstract computational model for context processing in a network based on the gross architecture of the dentate gyrus and hilus; (2) Implementation and simulation of physiologically and anatomically more detailed models within a phenomenological model of the hippocampal system; (3) Experimental verification of the assumptions underlying the models, and the predictions derived from them; and (4) Modification of the models in light of the new experimental data. The aim is to set up a continuous interaction between theory and experiment that leads to a better understanding of the hippocampal system. This project continues a collaboration that Dr. Minai and Dr. Best have developed over the last two years under NSF's Collaborative Research Initiation Program and represents a strong long-term commitment to theoretical, experimental, and computational research on the hippocampal system. It is expected that this collaboration will elucidate critical information-processing principles in the brain, and also lead to a fruitful application of these principles in areas such as robotics and intelligent systems.
