Encoding Spatial Context: A Hypothesis on the Function of the

There is now a very extensive body of data on the spatial correlates of neural activity in the hippocampal system of rodents. Place-specific activity is found throughout the system and lesions of various hippocampal regions cause severe deficits on spatial tasks. This has led to the hypothesis that the representation of spatial environments is a primary function of the rodent hippocampus. In this paper, we focus on a set of very interesting experimental observations which suggest that spatial representations in the hippocampus proper are highly context-specific and robust to sensory changes. This suggests the existence of a context-representation system which biases the pocampal representations. Based on experimental data from the literature, we argue that the dentate gyrushilus system is the locus of context processing, and present a simplified neural network model to support this hypothesis. 1