Burst generation by electrically coupled network in the snail helisoma: analysis using computer simulation.

Abstract

The effectiveness of electrical coupling between neurons as a mechanism for mediating single and repetitive bursts is investigated here using computer simulation. The cyberchron network in the snail Helisoma generates repetitive bursts controlling the animal's feeding behavior and served as the basic model for the simulation studies described in this paper. The action potential properties of individual neurons were modeled by the Rall equations describing generalized action potentials. Several properties of electrical coupling and its role in burst generation were demonstrated, including: (1) A neuron in an electrically coupled network can generate action potentials at a higher frequency than an isolated neuron with similar membrane properties due to the loading through the electrical junctions. However, the ability of electrically coupled neurons to generate high-frequency bursts of action potentials requires a concomitantly greater amount of driving current to overcome the junctional loading. (2) Temporal and spatial summation of synaptic input onto a neuron is maintained at its most effective level because the postsynaptic current is integrated across the long postsynaptic membrane time constant. (3) Initial simulations concentrated on a pair of electrically coupled neurons which were below threshold. Stimulation of one of the two neurons with a short pulse resulted in a reverberation or regenerative excitation between the two neurons. The reverberation terminated after a number of action potentials dependent on the specific model parameters. Similar results were obtained with a network containing a greater number (20) of model neurons if approximately one-half of the neurons were stimulated simultaneously. However, none of the cases studied produced more than a single discrete burst. (4) Simulations were also conducted on 20-neuron networks containing two subpopulations of model neurons differing in their values of coupling resistance and excitability. Some networks of this type required stimulation of only one cell to make the two subpopulations of model neurons reverberate with one another. Such simulations suggest the possibility that 'preferred' input pathways involving a small number of neurons would be capable of 'turning on' the activity of the entire network.