Classification of Pre-clinical Seizure State Transitions in Scalp EEG using Global Cross-frequency Coupling Index
Year: 2017
Type: dissertation
Abstract: In 30% of individuals that suffer from epilepsy, seizures remain uncontrolled by drug therapy. The unpredictability of clinical seizures reduces patient quality of life, ability to function in society, and increases their risk of SUDEP. This work aims to improve the state of the art of pre-clinical seizure detection systems so that they can be implemented to monitor patients, particularly, in the home environments outside the clinic. Simultaneous intracranial and scalp EEG recordings from patients with drug resistant epilepsy were used to examine cross-frequency coupling (CFC) features as a marker for the detection of a pre-clinical seizure state transition. Machine learning based classifiers were trained and combined in an expert system to identify pre-clinical seizure state transitions in a multi-patient, multi-seizure data set. This work shows that CFC features could be implemented in a home monitoring system to give advanced warning to patients and caregivers of an upcoming clinical seizure.
Source: TSpace (University of Toronto)
Author: Trevor Hilton
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2783297345