Integration of spatial and non-spatial information by heterogeneous dentate gyrus granule cells

Abstract

The hippocampus is the key site for learning and memory and for processing of spatial information in the brain. It is divided into three main subregions: the dentate gyrus (DG), the CA3 area, and the CA1 region, which are linearly interconnected to form a so-called trisynaptic circuit. Thus, the DG sits in a strategic position to gate the flow of information from the neocortex into the hippocampal network. The granule cells (GCs), the main cell type in the DG, receive 'where' and 'what' information from the medial and lateral entorhinal cortex, respectively. How they process this mixed information remains enigmatic. By characterizing the spatial information encoded by the excitatory postsynaptic potentials (EPSPs) in GCs, we demonstrated that the majority of GCs received spatially tuned synaptic input. However, only a minority of GCs successfully converted spatially tuned input to spatially tuned output. Furthermore, we found that mature GCs were highly heterogeneous in terms of their dendritic morphology and intrinsic excitability, which contributes to the sparse and heterogeneous firing of GCs. Finally, we discuss the possible origin of this neural heterogeneity and its potential role in enlarging the computational power of the DG, facilitating pattern separation in this network.