Complex periodic behaviour in a neural network model with activity-dependent neurite outgrowth

Empirical studies have demonstrated that electrical activity of the neuron can directly affect neurite outgrowth. High levels of activity cause neurites to retract, whereas low levels allow further outgrowth. Previously we studied networks in which all the cells reacted in the same way on electrical activity. Since experiments have shown that neurons may in fact react differentially, we study in this paper networks in which the range of activity where outgrowth takes place varies among cells. We show that this can lead to complex periodic behaviour in electrical activity and connectivity of individual cells. The precise behaviour depends on the spatial distribution of the cells and the distribution of the outgrowth properties over the cells. Any other cellular property that adapts slowly to electrical activity such that neuronal activity is attempted to be maintained at a given level, can lead to similar results. (C) 1996 Academic Press Limited.