Author ORCID Identifier
Date Available
5-17-2023
Year of Publication
2020
Degree Name
Doctor of Philosophy (PhD)
Document Type
Doctoral Dissertation
College
Engineering
Department/School/Program
Biomedical Engineering
First Advisor
Dr. Sridhar Sunderam
Abstract
About a third of all epilepsy patients may experience seizures that are resistant to medication. Surgical resection may be an option for some of these patients, but localization of the epileptogenic zone is difficult and expensive. For this process, the seizures must be observed and recorded in a clinical setting over several days. High-frequency oscillations (HFOs) have emerged as a possible biomarker for predicting the epileptogenic zone from a brief interictal recording of brain electrical activity without having to wait for seizures. The broad objective of this dissertation is to devise methods and criteria for accurate detection, sampling and localization of HFO activity for use in surgical diagnostics.
Spikes, which are also associated with epilepsy and can occur with or without HFOs, can give the appearance of HFOs when they pass through filters commonly used in HFO detection algorithms. Here, we develop and validate an unsupervised algorithm for HFO detection from the electrocorticogram (ECoG) of epilepsy patients that minimizes false HFO detections produced by such filtering artifacts. This novel HFO detector was found to perform with high specificity and moderate but acceptable sensitivity.
Secondly, HFO activity is known to be modulated by interictal vigilance state, but a comprehensive analysis of the effect is lacking. Here, we use a machine learning approach to score vigilance state in continuous overnight interictal ECoG data and examine the correlation with HFO activity. Through this approach we were able to partition each recording into four states corresponding roughly to wakefulness, rapid eye movement (REM) sleep, light non-REM sleep, and slow wave sleep in order of increasing arousal threshold. HFO activity was found to increase monotonically with this ordering of vigilance states and was greatest in slow wave sleep. We suggest how this approach could help identify segments of ECoG for HFO sampling with high yield without requiring conventional polysomnography.
Finally, we wish to know when to record and how many HFOs are needed in a sample to accurately localize the cortical region of HFO activity—the HFO zone—as a surrogate for the epileptogenic zone desired by the physician. Based on a detailed analysis, we concluded that there were no fundamental structural differences between spatial HFO profiles occurring in different vigilance states. Based on this analysis, we were able to make recommendations for the minimum number of HFOs to be observed at any recording location for it to be considered HFO-active regardless of the length of the recording.
Digital Object Identifier (DOI)
https://doi.org/10.13023/etd.2020.183
Funding Information
1) The Higher Committee for Education Development (HCED), Ministry of Higher Education and University of Babylon in Iraq. (2012-2018).
2) National Science Foundation Grant No. 1539068. (2018-2020).
Recommended Citation
Al-Bakri, Amir Fared Partu, "HIGH FREQUENCY OSCILLATIONS IN THE EPILEPTIC BRAIN: ACCURATE DETECTION, EFFECT OF VIGILANCE STATE, AND SAMPLE SIZE CONSIDERATIONS" (2020). Theses and Dissertations--Biomedical Engineering. 64.
https://uknowledge.uky.edu/cbme_etds/64
