Abstracts

Rationale: In the last few years, several laboratories have investigated the use of different bands of spectral content to identify novel biomarkers for seizures. High-frequency oscillations (HFOs, 80-500 Hz) have been considered as indicators of abnormal neuronal network activity in recordings from epileptic patients and animal models of epilepsy. At the other end of the frequency spectrum, infraslow activity (ISA, 0.01-0.1 Hz) has also been implicated in seizure initiation. However, the relationship between these oscillations and their possible correlation with different seizure onset patterns are poorly understood and controversial. Although few animal studies suggest a correlation between HFO characteristics and different seizure onset patterns, this effect is not often studied in humans and, when studied, tends to be inconclusive. In this study, we sought to determine whether certain temporal sequences of spectral power changes are indicative of specific seizure onset patterns. Such sequences could be used to identify possible mechanisms underlying the onset of different seizures. Methods: Seizures recorded from patients with medical refractory epilepsy who underwent presurgical evaluation with intracranial electrodes were analyzed. Seizures were classified based on their electrographic patterns identified by at least two EEG expert reviewers, and only the seizures whose pattern was deduced by both reviewers were analyzed. Seizure onset times and seizure onset regions were identified by clinical reviewers blind to this study. Low and high frequency oscillations were identified and analyzed to elucidate the features of different frequency bands at the initiation of seizures as well as the peri-seizure period. Temporal changes in all frequency bands were calculated and were averaged for each time point over all recorded regions, for each seizure. Finally, these changes for each pattern were plotted to identify the possible correlation between seizures sharing the same pattern. Results: Seizures (n=56) recorded from 9 patients diagnosed with medical refractory epilepsy were analyzed. Changes in the ISA and HFO characteristics were analysed for each seizure. Clustering analysis on the temporal changes of different frequency bands revealed that seizures sharing the same electrographical pattern express related changes in ISA and HFO characteristics that are consistent within the same pattern group. These findings suggest that activity in these widely different frequency bands and their correlations may provide additional information about how seizures might differ in initiation, propagation and termination. Conclusions: This analysis demonstrates that different underlying mechanisms might underlie the generation and the sustain of seizures with different onset patterns. This difference can be revealed by studying the changes in the pattern of different frequency oscillations occurring at seizure onset and peri-seizure period. With the help of these analyses we eventually may be able to pinpoint the involvement of different neuronal elements in the generation and termination of different seizures. Funding: NIH