How connectivity rules and synaptic properties shape the efficacy of pattern separation in the entorhinal cortex–dentate gyrus–CA3 network

Pattern separation is a fundamental brain computation that converts small differences in input patterns into large output patterns. Several synaptic mechanisms of pattern have been proposed, including code expansion, inhibition and plasticity; however, which these play role the entorhinal cortex (EC)–dentate gyrus (DG)–CA3 circuit, classical remains unclear. Here we show biologically realistic, full-scale EC–DG–CA3 circuit model, granule cells (GCs) parvalbumin-positive inhibitory interneurons (PV+-INs) DG, an efficient separator. Both external gamma-modulated internal lateral mediated by PV+-INs substantially contributed to separation. local connectivity fast signaling at GC–PV+-IN synapses were important for maximum effectiveness. Similarly, mossy fiber with conditional detonator properties By contrast, perforant path Hebbian plasticity direct EC–CA3 connection shifted network towards completion. Our results demonstrate specific optimize higher-order computations biological networks might be useful improve deep learning capabilities technical networks. The authors present model cortex–dentate gyrus–CA3 based on experimental data plays key