Abstracts

Rationale: Mechanisms of seizure initiation are unclear. One way to address the question is to evaluate neuronal activity in brain regions where seizures initiate. In patients with temporal lobe epilepsy and in epileptic pilocarpine-treated rats seizures frequently begin in the hippocampal formation. Methods: Tetrodes were used to obtain local field potentials and unit recordings from 272 principal cells and 107 interneurons in the dentate gyrus, CA3, CA1, and subiculum of 16 epileptic pilocarpine-treated rats as they experienced over 1000 spontaneous seizures. Results: Average action potential firing rates of subicular principal cells and interneurons increased progressively beginning 4 min before a spontaneous seizure. More than half of the recorded subicular neurons displayed increased preictal activity with significant consistency across multiple seizures. Substantial but smaller proportions of neurons in CA1 and the dentate gyrus displayed similar patterns of increased preictal activity, but CA3 neurons only rarely did. Theta was more likely to occur before seizure onset, and theta-related increased firing frequency contributed predominantly to preictal activity of subicular neurons. Preictal activation of other hippocampal neurons was largely independent of theta. Preictal inactivation occurred in over 25% of CA1 interneurons. Near seizure onset many hippocampal interneurons displayed a burst of activity followed by a brief pause in firing. The end of the pause coincided with the onset of rapid increases in seizure amplitude. A longer lasting reduction in interneuron firing rate occurred later during the seizure and was associated with reduced action potential amplitude, suggesting depolarization block. Initial attempts to identify CA1 interneuron subtypes based on theta-phase preference and activity during sharp waves suggest heterogeneity in seizure-related activity patterns. Conclusions: These findings reveal: (1) increased firing of many hippocampal neurons before seizure onset, which might be useful for seizure prediction; (2) heterogeneity in preictal activity patterns between hippocampal subfields, suggesting different roles in ictogenesis; (3) significant consistency in preictal activity patterns of individual neurons, suggesting specific seizure-generating networks exist that might be targeted for treatment; and (4) a substantial proportion of CA1 interneurons whose activity decreases before seizure onset, which might be ictogenic. Finally, our previous local field potential seizure analysis of various brain regions in epileptic pilocarpine-treated rats suggests only 9% of seizures in the present study were likely to have begun in the right, dorsal hippocampus where recordings were obtained. Thus, the significantly consistent preictal activation of substantial proportions of neurons in the recorded region suggests spontaneous seizures are preceded by progressive activation of a network that extends beyond the site of electrographic seizure onset.