Abstracts

Rationale: The behavior of populations of hippocampal neurons and their interactions are important for understanding the process of seizure generation in temporal lobe epilepsy. Hyperexcitability in hippocampal networks has long been thought to be a mechanism for the initiation of spontaneous seizures. However, more recently the role of inhibition in this process has been investigated, especially for seizures with low-voltage fast (LVF) activity near the onset. Methods: To investigate this, a rat model of chronic epilepsy was employed, with pilocarpine as the agent used for induction of status epilepticus. In the months following status epilepticus, spontaneous seizures were observed to occur with an approximate frequency of 4 per week. At this time, rats were implanted with 7 drivable tetrodes which were slowly advanced into the CA3 field of the hippocampus. Broadband signals (0.8 Hz - 6.5 kHz) were amplified using a 31 channel wireless headstage, sampled at 40 kHz and stored along with video recordings. Features extracted from the action potential waveforms were automatically clustered and final single neuron discrimination was done manually. Pyramidal cells were differentiated from interneurons on the basis of firing rate, waveform shape, and autocorrelation. To gain insight into the process of seizure generation, we examined the average firing rate and spike field coherence of these two populations of neurons during the 2 hours leading up to each seizure onset. The baseline mean and standard deviation of these measures were obtained from the time period starting 2 hours prior to seizure onset and ending 1 minute before onset. Results: The majority of seizures (> 50%) were preceded by LVF activity, which began with an LFP spike and slow wave 5-10 seconds prior to the first appearance of large amplitude ictal spiking. These seizures (n = 18) were used for further analysis. The period of time between the initial slow wave and the start of ictal spiking was accompanied by a significant (p < 0.05) increase in interneuron firing rate for all seizures in the analysis. This increase was, on average, greater than 2 times the baseline interneuron firing rate. Firing rate remained elevated until the appearance of the large amplitude ictal spikes, then decreased below baseline mean for the duration of seizure. Pyramidal cell firing rates decreased significantly while interneuron firing rates were high, but rebounded as ictal spiking began. Interneuron action potential times also showed an increased coherence with LFP oscillations in the frequency bands of either 10-30 Hz, 60-150 Hz, or sometimes both, suggesting increased synchrony during the period of elevated activity. Conclusions: These data are consistent with recent seizure generation hypotheses that synchronized interneuron networks play a role in the recruitment of synchronously firing pyramidal cells in some types of temporal lobe epilepsy seizures.