Quantitative analysis of epileptic and meditation EEG signals

• Main Author: Jing, Jin
• Other Authors: Justin Dauwels
• Format: Theses and Dissertations
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing; DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Medical electronics
• Online Access: https://hdl.handle.net/10356/68880
• Institution: Nanyang Technological University

Electroencephalography (EEG) refers to the recording of the brain's spontaneous electrical activity over a period of time. EEG remains the primary test to diagnose brain disorders such as epilepsy, sleep disorders, coma, encephalopathies, and brain death, which cause abnormalities in EEG readings. In this thesis, novel signal processing methods are developed for quantitative EEG analysis. We demonstrate that brief interictal intracranial EEG can be used to delineate the seizure onset zone (SOZ) for patients with epilepsy. Unlike traditional methods based on single EEG features, multiple features are applied to localize the SOZ in this thesis. An automated system is also developed for detecting high frequency oscillations (ripples and fast ripples) in epileptiform EEG. Moreover, in order to assist EEG interpretation and improve the efficiency of the neurologists, we develop a novel system that dramatically accelerates the process of acquiring expert-annotated EEG records based on template matching under Dynamic Time Warping. Last but not least, we also develop an automated system for paroxysmal gamma wave (PGW) extraction in scalp EEG recordings of meditation practitioners. We have better insights about the brain activity during meditation through PGW-related analysis. In this thesis, we consider three applications of EEG analysis: • Application 1: To localize the SOZ in the brain of epilepsy patients • Application 2: To accelerate the annotation of epileptiform spikes • Application 3: To study the brain activity during meditation Both Applications 1 and 2 are related to EEG recordings of patients with epilepsy, but in different EEG data types, intracranial and scalp EEG, respectively. On the other hand, in Application 3 we analyze the scalp EEG recorded from the meditation practitioners. The common EEG pattern that joins three applications together is a special waveform named \spike". It is also known as “interictal discharges” or “paroxysmal gamma waves (PGWs)” under different applications. Interictal spikes are the key diagnostic biomarker for epilepsy, predicting seizure recurrence, and allowing a physician to make a confident diagnosis of epilepsy and to prescribe appropriate treatment. On the other hand, the PGWs found in meditation EEG can be characterized by distinct high-frequency biphasic patterns, with a sharp peak closed to an interictal spike. Strong EEG activity has been observed in Bhramari Pranayama meditators, exhibiting frequent PGWs in active brain regions. However, visual inspection and manual extraction of spikes (or PGWs) is tedious, time-consuming and subjective, relying on experts who are in short supply. As a result, there is a great need for automated systems for spike detection. To this end, we have developed and utilized spike detection algorithms for each application, emphasizing on different aspects to cater different needs.