Abstracts

Rationale: It is commonly assumed that large-scale network remodeling is required to create the conditions for spontaneous recurrent spreading seizures. Thalamic excitability has been previously reported in a chronic neocortical epilepsy model, and it has been theorized that development of a "hyperexcitable intra-thalamic network" is necessary for the thalamus to generate epileptiform activity. We report here a study of subcortical seizure activity in an acute rat penicillin neocortical seizure model. Methods: We investigated seizures in 8 wild type adult Sprague Dawley rats. Penicillin solution was microinjected into left area V1. Tungsten microelectrodes were used to record from ipsilateral and contralateral neocortex (all rats), simultaneously with ventroposterior medial nucleus (VPM) of thalamus (5 rats), hippocampus (2 rats), and locus coeruleus (1 rat). Neural signals were sampled at 40 kHz using a Plexon system. Because seizure spread can be overestimated when examining local field potentials (LFP) only, we required a multiunit activity (MUA, 500 ?" 3000 Hz) correlate to define the presence of seizure activity. Fluorescent dye was used to track the diffusion of injected penicillin. Results: The dye revealed that penicillin remained restricted to within 1 mm of the injection site. In the rats with VPM electrodes, we recorded an average of 21.8 seizures (range 5-30). Ictal activity was also seen in the thalamus in an average of 18.4 seizures per rat (as indicated by MUA). As predicted by previous acute animal and human work, we found that low frequency ictal discharges can appear distant to the injection site without accompanying MUA activity. In the two rats with hippocampal depth electrodes, we recorded 17 and 24 seizures each, with all seizures involving the hippocampus. In the rat with a locus coeruleus depth electrode, 10 seizures were recorded, with five appearing in the LC. Conclusions: Subcortical structures were frequently involved in seizures despite no direct effect from the penicillin injected into a distant neocortical site. Thus, we have demonstrated that neither a chronic development process nor a pathological epileptogenic network is necessary for large-scale seizure spread, and moreover that seizure spread may routinely involve subcortical structures. This finding also provides important implications on the utility and benefit of acute seizure models to produce meaningful information about seizure propagation. Funding: NIH/NINDS R01 NS084142 Raymond and Beverly Sackler Convergence Award