Abstracts

Rationale: Interictal high-frequency oscillations (HFOs) are emerging as promising biomarkers of epileptogenicity. Little is known, however, on ictal HFOs, whose immediate relationship to the seizure per se makes them potentially the best indicator of the epileptogenic zone. We investigated the time course of changes in HFOs preceding and following seizure onset and quantified the relationships between ictal HFOs, lesional tissue and the seizure-onset zone (SOZ).Methods: Visual inspection and wavelet analysis were used to identify HFOs (ripples, 80-250Hz; fast ripples=FRs, 250-500Hz) in 14 seizures from 10 consecutive patients undergoing intracranial EEG (500Hz filter and 2000Hz sampling) for drug-refractory lesional focal epilepsy. A total of 704 channels were assessed (mean of 50.3 channels/seizure). For each seizure, four EEG sections were selected for HFO assessment: 1) baseline (8s), ending 30s prior to the first EEG change suggestive of seizure activity; 2) immediate preictal (8s), ending at the first EEG change suggestive of seizure activity; 3) ictal onset (4-8s), starting at the first unequivocal ictal EEG change; and 4) ictal evolution (4-8s), starting at the first EEG sign of seizure propagation. We evaluated changes in HFO rates across the four sections, and investigated whether HFO rates (no HFOs vs low vs medium-high rates; rates were defined as low or medium-high according to whether these were ? or > than the mean rates in a given section +1 SD) correlate to lesional tissue and the SOZ. Statistical analysis included univariate and multivariate analyses.Results: HFO rates did not differ significantly between baseline and immediate preictal sections, but increased substantially in both ictal sections (all p<0.01, ripples and FRs; Figure). In all sections, there was a higher proportion of channels with medium-high HFO rates inside the lesional tissue compared to channels without HFOs or with low HFO rates (all p<0.001, ripples and FRs). In each ictal section, there was also a higher proportion of channels with medium-high HFOs rates inside the SOZ compared to those without HFOs or with low HFO rates (p<0.001, ripples and FRs). No such differences were found in baseline or immediate preictal sections. The number of channels with medium-high HFO rates increased progressively over time, particularly in the two ictal sections (p<0.05, ripples and FRs).Conclusions: There was no change in HFO rates in the immediate preictal period compared to baseline. This is in contradiction with earlier findings and may be related to the definition of the seizure onset, in our case the first sign suggestive of seizure activity . HFO rates clearly increase upon seizure onset. Moreover, higher rates of ictal HFOs correlate with lesional tissue and the SOZ, which supports their intimate association with underlying epileptogenicity. Given the similarly promising data on interictal HFOs, future studies should assess the comparative value of interictal and ictal HFOs in identifying the epileptogenic zone. Supported by AES, AANF, Epilepsy Foundation and CIHR.