Date Approved


Graduate Degree Type


Degree Name

Engineering (M.S.E.)

Degree Program

School of Engineering

First Advisor

Dr. Rob Bossemeyer

Second Advisor

Dr. Paul Fishback

Third Advisor

Dr. Samhita Rhodes

Fourth Advisor

Dr. Konstantin Elisevich


One third of the patients diagnosed with focal epilepsy do not respond to antiepileptic drugs. For these patients the possible diagnosis options to give seizure freedom or at least reduce seizure frequencies significantly would be surgical resection or seizure interrupting implantable devices. The success of these procedures depends on accurate detection of the region causing seizure also known as epileptic zone. This requires detail pre-surgical evaluation including Invasive Video Electroencephalographic Monitoring (IVEM). The resulting great volume of intracranial Electroencephalography (iEEG) signal is visually examined by an expert epileptologist which can be time consuming, extremely complex, and not always effective. We have introduced an automated method to help the epileptologist analyze the iEEG signals.

Literature suggest that signals recorded from brain regions subject to seizure activity produce a short durational high gamma ripple activity in the iEEG called High Frequency Oscillations (HFOs). The algorithm presented in this thesis uses an automated time-frequency space analysis method to detect HFOs and distinguish them from high frequency artifacts. As HFOs are short-lived high frequency oscillations, the time-frequency space analysis method chosen should have good time and frequency resolution capabilities. The Stockwell transform was used for this purpose which is a variable window version of the Short Time Fourier Transform (STFT). We have modified the detection algorithm to analyze the multi-channel iEEG data obtained from patients monitored at the Spectrum Health Epilepsy Monitoring Unit (EMU) and found that the electrode site recordings exhibiting higher HFO rate are within the Seizure Onset Zone (SOZ) determined by visual examination of the iEEG recordings by the epileptologist. These electrodes also continue to show higher HFO rate throughout the entire study. The HFO analysis presented in this thesis suggests that HFO detection and identification may be used to reduce IVEM monitoring time by aiding the neurosurgeon delineating the epileptic zone in relatively shorter time. This will lead to better surgical outcome or succesful implantation of the seizure intervention devices.