Effect Of Semi-lunar Granule Cells In Hippocampal Dentate Gyrus On Pattern Separation

The hippocampus is one of the most important brain regions for pattern separation,the process that converts similar inputs into highly differentiated,non-overlapping representations of information.Exploring how the hippocampus processes incoming information through pattern separation could help general artificial intelligence learn how to store information more efficiently and in large quantities.Sparse firing granule cells(GCs)in the dentate gyrus(DG)have been thought to perform this computational task.However,it is not clear which factors in the network affect the characteristics of sparse discharge of granule cells.Long-duration plateau potential and bursting firing properties of semi-lunar granule cells(SGCs)may contribute to spare firing of GCs which could facilitates pattern separation.In order to investigate whether and how SGCs contribute to pattern separation.In this paper,a simple and biologically relevant dentate gyrus computational model is established.Model shows that:1.Confirmed that sparse firing of GCs is the key factor for the pattern separation,and inhibitory neurons in the networks are one of the reasons for the sparse firing of GCs.2.A computational model includes SGCs is established.Based on the effect of pattern separation,the parameters related to the SGCs are estimated.Among them,the strength of the connection between the SGCs-BCs and SGCs-MCs are about 6%and 10%respectively,and the number of the SGCs is about 30,000-40,000.3.Based on result,excitability of granule cells increases and pattern separation weakens when the SGCs are absent.The effect of the SGCs in the network can be compensated by enhancing the inhibitory factors in the computational model.SGCs in DG network like a signal amplifier which magnify input signal by their high strength firing to activate inhibitory cells in DG,therefore promoting pattern separation.