Activity-Dependent Modification of Intrinsic Neuronal Properties

The processes that develop and maintain the intrinsic electrical properties of neurons are modeled by allowing the maximal strength of membrane conductances to be slowly varying functions of the intracellular calcium concentration. The resulting dynamic regulation of conductances allows model neurons to self-assemble their membrane currents and to react to and recover from external perturbations. This dramatically increases the stability of the model neuron and makes its intrinsic characteristics activity-dependent. For example, model neurons in a two-cell network spontaneously differentiate in response to each other’s activity. In a spatially extended model neuron, the dynamic regulatory mechanism causes a non-uniform distribution of currents to develop in response to both the morphology and the activity of the neuron. Introduction Work on connectionist neural networks has focused largely on the effect of activitydependent synaptic plasticity on network function [1]. At the same time, enormous experimental effort has gone into studying long-term potentiation and depression of synapses [2]. While the role of synaptic changes is of great importance to learning and memory, it should not be forgotten that the intrinsic characteristics of individual neurons can also be modified by activity [3]. Connectionist networks tend to treat individual neurons as relatively trivial dynamic systems. Real neurons are, of course, far from simple and they can exhibit a wide range of endogenous behaviors including tonic spiking, periodic firing in bursts and bistability between silent and firing states, even in the absence of synaptic input [4]. The intrinsic electrical characteristics of a neuron depend on the strength and distribution of its membrane currents. Modeling studies indicate that even small changes in current strengths or distributions can have a dramatic effect on neuronal behavior. In mathematical modeling terms, realistic neuron models have poor structural stability properties. In light of this fact, the robust stability of biological neurons is surprising and even more remarkable when we realize that the ion channels that conduct membrane currents are continually degrading and being replaced. Furthermore, neurons must adjust to their own growth and to possible changes in the extracellular environment.