Micelle Reed, University of Utah
Health
For patients with medically intractable epilepsy, a neurological disorder characterized by seizures that are unable to be controlled with medication, surgical resection of the seizure generating zone is necessary to obtain seizure freedom. Intracranial electroencephalography (iEEG) is used for determining areas for resection when noninvasive techniques fail to pinpoint a specific area. High frequency oscillations (HFOs), observed through iEEG, are successful biomarkers for the seizure generating zones and are more localized to the source of seizures than areas of propagation. The most common method for determining HFO occurrence lies in expert epileptologist interpretation of the iEEG data, although this method is limited to small data sets and the expertise of the doctor. This study uses the signal processing techniques of spectrogram analysis and continuous waveform transforms to find high frequency content in sampled patient data. Through the use of Friedman’s Tests, statistical difference between channels is determined and subsequent Wilcoxon signed-rank tests are performed to find the channels with statistically greater high frequency content. This allows for an unbiased, automated determination of seizure generating channels. Localization of the seizure generating area can be decided because of the 1-1 correspondence between the channel signal and macroelectrode placement on the brain. If functional mapping reveals the cost of resection of that area of the brain to be less than the benefit of reduction in seizure activity, surgery will be performed. Through the determination of the true seizure generating zone, surgical resection will lead to the best patient outcome of potential seizure freedom and improved quality of life.