Abstracts

Rationale: Mutations in genes encoding sodium channels SCN1A and SCN1B are responsible for childhood epilepsies, and several anticonvulsants that target sodium channels worsen some epileptic syndromes. These findings lead to the paradoxical idea that a reduction in sodium channel function may contribute to seizure activity. A leading hypothesis is that the locus of this effect resides in sodium channels on interneurons. However, this hypothesis does not exclude a contribution from reduced sodium currents in principal neurons. In the course of studying spatial patterns of propagation of epileptic spikes and seizures using a computer model of the CA3 region of hippocampus, we found that reduction in the probability of axonal conduction was an important determinant of whether the population discharged as a single spike or a seizure. Methods: To simulate network activity we used a two-dimensional computer model of the CA3 neural network (described in Swiercz et al. J Clin Neurophysiol 2007 Apr;24(2):165-74). The majority of neurons in the network are pyramidal cells with a smaller number of interneurons. The majority of connections are excitatory recurrent collateral synapses. Connectivity between cells is random with uniform connection probability that decreases with distance. In order to facilitate the occurrence of spontaneous synchronous bursts of activity we increased excitatory synaptic strength or decreased inhibitory synaptic strength. Then we modeled sodium channels mutations as an increasing probability of action potential initiation and propagation. Results: With high probabilities of action potential initiation and propagation, network activity occurred as short synchronous bursts separated by relatively long periods with low levels of spontaneous activity. Increasing the probability of action potential failure initially increased the interval between synchronous population events. Further increases in failure rate surprisingly resulted in sustained reentrant activity. Increasing the failure rate beyond this point stopped all network activity, as expected. Our computational results are being validated in vitro in organotypic slices of region CA3 of rats hippocampus. Action potentials are gradually compromised by low concentrations of TTX, with or without GABA receptor blockade, and network activity is recorded electrophysiologically. Conclusions: Our findings demonstrate the possibility of developing reentrant synchronous activity in neural networks by progressively increasing the number of impaired sodium channels. The mechanism responsible for this phenomenon is partial desynchronization of neuronal activity. This reduces the degree of activity-dependent synaptic depression, so that when waves of excitation return to their origin, the origin has already recovered from activity-dependent depression. Under these circumstances, repeated activation of the same pathway can take place, and sustained reentrant activity is possible.