Electrical coupling provides an important pathway for signal transmission between neurons. In several regions of the mammalian brain electrical synapses have been detected, and their role in the synchronization of neural networks and the generation of oscillations has been studied theoretically. Recently, it has been found that the amplitude of the postsynaptic potential is maximized for a specific diameter of the postsynaptic fiber.

In this thesis, the impact of the fiber's diameter on the success or failure of the action potential initiation and propagation is studied theoretically. Systems of two coupled neurons, as well as small networks, are investigated. The passive and voltage-dependent properties of the neurons are implemented using compartment modeling. The results of the simulations show that for neurons with non-branching dendrites an action potential is initiated only for a specific, optimal diameter. In contrast, for neurons with branching structures the signal transmission improves monotonically with increasing diameter. By studying a model network with a ring architecture it is demonstrated that network activity crucially depends on the diameter of the coupled fibers.