Complex Patterns of Oscillations in a Neural Network Model with Activity-Dependent Outgrowth

Abstract

Many processes that play a role in shaping the structure of the nervous system are modulated by electrical activity. For example, electrical activity can affect neurite outgrowth: high levels of activity, resulting in high intracellular calcium concentrations, cause neurites to retract, whereas low levels of activity, and consequently low calcium concentrations, allow further outgrowth [1]. As a result of this and other activity-dependent processes, a reciprocal influence exits between the formation of connectivity (”slow dynamics”) and activity (”fast dynamics”). We have made a start at unravelling the implications of activitydependent neurite outgrowth [2, 3], and have been able to show that several interesting properties arise as the result of interactions among outgrowth, excitation and inhibition: (i) a transient overproduction (’overshoot’) during development with respect to connectivity; (ii) the neuritic fields of inhibitory cells tend to become smaller than those of excitatory cells; (iii) the spatial distribution of inhibitory cells becomes important in determining the level of inhibition; (iv) pruning of connections can no longer take place if the network has grown without activity for longer than a certain time (’critical period’). The results show many similarities with findings in cultures of dissociated cells.