Results

The parameters of neuron and synaptic models were initially selected to allow for generation of bursts of action potentials in a small network with a sufficient number of connections. For this set of parameters, multiple random connected generated networks were used for simulation. The results for two networks are shown. Fig. 2 shows traces of membrane potentials for all neurons in small networks of 81 excitatory (top) and 9 inhibitory neurons(bottom). Networks were activated by supplying background Poisson activity to the inputs of 4 selected neurons. Panel A shows activity of the network with 2 excitatory and 3 inhibitory inputs per neuron. There is no synchronized bursting activity and the measure of synchronization is low, $\Delta =0.027$. After increasing the number of excitatory connections to 3 and inhibitory to 5, without changing other parameters, some bursts appear in the output (Fig. 2B) and the synchrony measure increases to 0.077. Increases in the number of connections to 4 excitatory and 6 inhibitory per neuron further increases synchrony of bursting activity. All neurons generate bursts and they are more synchronized (Fig. 2C). The measure of synchrony is significantly higher ($\Delta = 0.25$). Similar results were obtained for the second network illustrated in Fig. 3. This network was activated by applying background activity to 9 selected neurons. Panel A shows traces of membrane potential for network with 2 excitatory and 3 inhibitory synapses on each neuron. A few neurons generate bursts but the activity is not highly synchronized; the measure of synchrony $\Delta = 0.036$. In panel B there are traces from simulation of network with 4 excitatory and 9 inhibitory connections per neuron. The activity is more synchronized ($\Delta=0.098$). Fig. 3C shows that increasing of number of excitatory and inhibitory connections to 6 and 9 respectively, increases synchrony of bursting activity significantly to $\Delta=0.18$. Similar results were obtained for other networks. These suggests that increased connectivity during sprouting can indeed contribute to epileptogenesis, even if sprouting involves inhibitory connections.