A note on bistability in a simple synapseless ‘point neuron’ model

Abstract

Neural network models include at least one synaptic interaction between a pair of neurons so that plasticity can be conferred on the system. The single neuron is usually treated unrealistically in these models, functioning as a simple machine that transforms weighted synaptic input to firing frequency in a sigmoidal fashion. Knowledge gained in the past 15 years about the ways voltage-gated ion channels (the molecules of excitability) work, calls for re-evaluation of the role played by intrinsic neural excitability properties in learning and memory. We show that complex memory processes, ranging from seconds to ‘lifetime’, can be induced in a point neuron by including a realistic gating machinery of a potassium selective voltage-gated ion channel that is known to express in hippocampal cells.