Fast Automated Detection Method Of Seizure Onset Zone By High Frequency Oscillations

The latest study found that high frequency oscillations(HFOs) is a new biomarker for epileptogenic tissue. It mainly appears in the core region of seizure onset zone(SOZ), and has higher frequency, amplitude and smaller propagation range. It is sometimes difficult to localize SOZ by epileptiform discharges conventionally, because of the wide propagation range; This localizing last 72 hours from craniotomy to resection; the relationship between epileptiform discharge and seizure is still unknown. All these uncertainties are easy to increase the risk of infection in patients with prolonged surgery and misdiagnosis. In recent year, HFOs becomes the golden standard of SOZ localizing. Visual marking HFOs by experienced doctors is very time consuming and highly subjective, therefore, numerous’ automatic HFOs detection method have been developed and become hotspot.According to the current study, we summed up the propagation characteristic in clinical practice:1) HFOs occurres mainly in the core region of SOZ.2) HFOs more appears in seizures or interictal period, and has a higher frequency and intensity.3) Rs and FRs are more specific and precise compare to epileptiform discharge, FRs is the most. Some form of the HFOs can be recorded in 3 or more than 3 adjacent channels.Therefore,thought that all these channels are in the core of the SOZ.In this study we designed a SOZ localizing method by marking HFOs. We applied the power spectral density technique to find the suspected channels as curve template, and compare cross correlation between curve template and two adjacent channels; Meanwhile, we analyzed signal in time-frequency domain by using complex Morlet wavelet transform. The use of these two methods achieve fast and accurate localizing. At last, we test 6 cases, the results show that this method has good sensitivity and specificity, helpful to localizing SOZ before surgery, and significant to the success of epilepsy surgery.