Computational Models Of Resection In Drug-Resistant Epilepsy

Degree type
Doctor of Philosophy (PhD)
Graduate group
Bioengineering
Discipline
Subject
Epilepsy
Machine Learning
Medical Imaging
Networks
Surgery
Computer Sciences
Medicine and Health Sciences
Funder
Grant number
License
Copyright date
2018-09-28T20:17:00-07:00
Distributor
Related resources
Contributor
Abstract

Epilepsy is the second most common neurological disorder worldwide, and severely limits the quality of life of those who suffer from it. Patients who do not respond to standard anti-epileptic medications often require surgery and resection of large volumes of brain tissue. This standard surgical therapy, which has remained largely unchanged in the past 30 years, is based on epileptologists identifying regions of seizure activity visually on intracranial electrophysiology recordings. Despite extensive attempts and research to improve surgical efficacy, a significant number of patients still fail to achieve complete post-surgical seizure freedom and suffer significant neuropsychological deficits. With the advent of newer neurostimulation technologies and focal laser-based ablation, minimally invasive therapies are possible, but the challenge is where to apply them. There is great promise, however, in new techniques to identify optimal resection targets and merge them with quantitative multi-modal neuroimaging and network dynamics derived from electrophysiology data. In this thesis, I present a novel approach to combining multimodal neuroimaging, expertly annotated clinical neurophysiology, and quantitative imaging and network measures to inform clinical decision making in an integrated fashion. I find that network-based dynamics hold promise for identifying focal and disparate targets in the epileptic network that respond well to localized therapies. I also find that translating these findings from bench to bedside requires large-sample clinical trials that span multiple institutions, multiple imaging modalities, and clinical expertise. Open-source data sharing platforms, such as the tools I have created as part of the following work, will enable these big data studies to take place. My hope is that the work presented in this thesis will herald a new approach to epilepsy interventions, such as highly targeted devices or ablation to minimize side effects, and help patients who suffer from this destructive illness.

Advisor
Brian Litt
Andrew Tsourkas
Date of degree
2017-01-01
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation