Abstracts

Rationale: High Frequency Oscillations (HFO) (80-500 Hz) have been implicated in epilepsy but also appear in a variety of other conditions. During clinical recordings, HFO rates appear to increase prior to seizure onset. These observations have been duplicated in animal models. But seizure rates are also known to be state-of-vigilance (SOV) dependent. We hypothesize that the relationship between HFOs and seizure might be complicated by HFO rate changes with SOV. Methods: EEG and head acceleration were recorded continuously from five adult male Long-Evans rats 300 450 g in weight using a custom-made electronics and acquisition system. Seizures were induced by microinjection at the time of implant of 5-10 ng of tetanus toxin into the left ventral hippocampus during surgery for electrode placement. Bipolar pairs of 50 um diameter depth electrodes are placed in the dorsal hippocampus and recorded, along with field potentials from cortical screws placed bilaterally, sampled at 2 kHz. The animals were monitored continuously for up to 30 days following a 12 h on off light dark artificial lighting cycle. Seizures and sleep scores were automatically detected using algorithms previously published by our lab. Ten second intervals were used for sleep scores (3 stage NREM, REM, and Wake) and one second intervals were used for seizures. High frequency Oscillations were detected if they were at least (6 SD) above the hourly mean when band passed filtered between 80-500 Hz and they had at least 6 distinct peaks.Results: High Frequency Oscillations, SOV, and seizures were identified using quantitative automatic classification method in both seizure (N=5) and non-seizure animals. Seizure event rates peaked in latter half of the light cycle. SOV bout durations also had a strong diurnal variation, with, as expected, highest period spent in Wake during early part of the dark cycle. Overall, HFO rates from depth electrodes peaked in the middle of the light cycle. This dependence was not observed for HFOs from screw electrodes. The day-averaged rate of HFO is state dependent with REM having the lowest rate, Awake having the second lowest rate but a fairly high variance for time of day, and NREM sleep having the highest rate of HFO per second. This dependence was observed for HFOs from both hippocampal depth recordings and cortical screw recordings. Significant (p<0.05) correlation was found between HFO rates observed hourly seizure rate from 7/11 depth electrodes (from 5 rats). No significant correlation was observed between screw-observed HFO rates and seizure rates.Conclusions: Diurnal variations are observed in HFO, SOV, and Seizure rates. HFO rates are also state dependent. HFO rate correlates with seizure rates. But this correlation cannot obviously be decoupled from its dependence on SOV. Therefore it is not clear that HFO rate is a direct indicator of seizure susceptibility or predictor of seizure onset. Further work on the dependence of HFOs on SOV may enhance our understanding of the underlying mechanisms of seizure generation. Work funded through NIH-R01NS065096.