An associative neural network to model the developing mammalian hippocampus

Abstract

Electrophysiological recording of pyramidal hippocampal cells along early postnatal development shows a pattern of maturation consisting of a progressive reduction of the accommodation and increasing excitability. Electrophysiological, pharmacological, behavioural and lesion techniques permit to manipulate cellular, synaptic and connectivity properties in order to explain how cellular and synaptic mechanisms interact with the pattern of connectivity to give rise to a behaviorally important output pattern. These techniques, although powerful, have their limitations in that only some of the potentially important cellular or synaptic properties are amenable to experimentation. We propose a complementary approach using an associative network model based Hebbian laws, able to simulate the biological system, whose sequential output depends on the interference between a slow and a fast components.