Abstracts

Rationale: Determine the diagnostic accuracy of high-frequency oscillation (HFO) rates for localizing epileptogenic regions in the temporal lobe.  Methods: Intra-operative iEEG recordings, 2-10 minutes in duration, from 45 patients with temporal lobe epilepsy (TLE) obtained at the time of impantation, and 24 patients at the time of resection were analyzed using a custom HFO detector. During the recording some subjects received boluses of remifentanil to increase epileptogenic activity. The identified HFO events were classified and characterized as true or false ripple/fast ripple on spikes, or ripples/fast ripples on oscillations, and confirmed by visual inspection. The accuracy of HFO rates for localizing the seizure onset zone was performed using a ROC approach.  The positive and negative predictive value of resected and residual HFO rates on seizure outcome were calculated using a cutoff value for the threshold HFO rate used to calculate the predictive values will be selected so as to maximize Youden’s J. Results: Preliminary analysis demonstrates that in a cohort of seven patients anesthetized using remifentanil with multiple subdural strips, and a hippocampal depth electrode, undergoing a tailored temporal lobe resection, the PPV of resected and residual ripple rates on seizure free outcome was = 61.5% (56.5-67.3 95%CI), the NPV of resected and residual ripple rates on seizure free outcome was = 63.0% (52.8-74.5 95%CI). In a smaller cohort of TLE patients (n=3) undergoing depth electrode implantation the area under the ROC curve for localizing the seizure onset zone using HFO rates was 82%. Conclusions: HFO rates measured in brief intra-operative recordings can be helpful in localizing epileptogenic brain regions in the temporal lobe. Funding: This work was supported by NIH/NINDS K23NS094633 (SAW).