Abstracts

Rationale: There are many acute cerebral insults that lead to the development of chronic epilepsy in some, but not all, people who suffer from such insults. There is considerable interest in the development of antiepileptogenic drugs, and some progress has been made at identifying such agents. If a reliable predictor of epileptogenesis could be developed that would identify people susceptible to develop chronic epilepsy after brain insult, this would allow for early prognostic treatment by use of prophylactic drugs in just these patients. As a first step towards this end, in this study, we have used nonlinear synchronization techniques to evaluate continuous (over months) electroencephalographic (EEG) recordings in rats with chronic epilepsy induced by the lithium/pilocarpine model of status epilepticus and compared the results of the analysis with the ones obtained from rats that do not develop epilepsy (control rats). The hypothesis we tested was that progressive changes in network synchronization of EEG are predictive of epileptogenesis. Methods: Adult male Sprague-Dawley rats were used in this pilot study. All rats were implanted with an array of 10 Tungsten microwires targeting the cortical and the deeper structures of the limbic region. SE was induced by 3 mmol/kg intraperitoneal lithium chloride followed 24 hrs later by 30 mg/kg pilocarpine subcutaneously. EEG was continuously recorded and the rats were divided into two groups based on the EEG stage when their SE was stopped: Group 1-SE EEG stage III, which is a marker for a less severe brain insult; rats treated at this stage have a lower probability of developing epilepsy, and Group 2-SE EEG stage V, which is a marker for a very severe brain insult; rats treated at this stage have a higher probability of developing epilepsy. Continuous EEG recording was then carried out over 10 weeks in all rats that survived the initial insult of SE to verify the time of onset, frequency, and duration of spontaneous seizures. The analysis of EEG using nonlinear dynamics was performed retrospectively. The average measure of synchrony across all EEG channels was monitored over time until the rats developed chronic epilepsy. Results: We observed a long-term monotonic trend towards synchronization over a 5 week period (See Figure 1 and caption) following the initial SE episode and prior to development of chronic epilepsy in all the rats in Group 2. On the contrary, rats from Group 1 did not show a monotonic increase in synchrony during the same time period. We believe that this trend in network synchrony can be followed in real-time and be used as a marker for epileptogenesis. Conclusions: In summary, we have preliminary results that suggest a way to use computerized analysis of EEG recordings to predict which animals will develop epilepsy after SE. This approach constitutes a first step in identifying EEG-based biomarkers of epileptogenesis and could then be used to predict epileptogenesis after traumatic brain injury (TBI) in human patients.