Signal Analysis of Gamma Brain Oscillations Recorded with Low-Tc and High-Tc SQUID-Based Magnetoencephalography - A Matching Pursuit Approach

Examensarbete för masterexamen

Please use this identifier to cite or link to this item:
Bibliographical item details
Type: Examensarbete för masterexamen
Title: Signal Analysis of Gamma Brain Oscillations Recorded with Low-Tc and High-Tc SQUID-Based Magnetoencephalography - A Matching Pursuit Approach
Authors: Goudian, Tamara
Abstract: Brain diseases present a growing social and economic burden worldwide. However, understanding the pathogenesis leading to many such diseases remains a great challenge. Magnetoencephalography (MEG) is a promising safe and noninvasive imaging technique with demonstrated utility in several research and clinical neuroscience applications. Most of the common MEG devices detect magnetic signals emitted from the brain using highly sensitive sensors - superconducting quantum interference devices (SQUIDs) - that operate below a specific temperature known as the critical temperature (Tc). To support the usefulness of MEG, suitable analysis techniques are needed to extract meaningful information from MEG recordings. Several standard time-frequency analysis techniques have been used for this purpose but impose a trade-off between frequency and time resolutions making it difficult to capture transient and rhythmic components of signals simultaneously. One technique that offers improvement with this respect is Matching Pursuit (MP). This project explores the MP algorithm as an option for the analysis of brain oscillations in MEG recordings obtained from human subjects following. This is motivated by its success when applied to electroencephalography (EEG) recordings from monkeys in recent research efforts. Here, MP is used to study gamma oscillations, which are associated with cortical functions involving high-level cognitive and sensory processes. Two implementations were investigated: one using a dyadic dictionary, and the second using a stochastic dictionary. The aim is to test MP’s capacity to extract different or new information from MEG recordings: first differences in data obtained from low-Tc SQUID-based devices and high-Tc SQUID-based devices, and then differences between human subjects with and without autism spectrum disorders (ASD). The results of both tests reveal some insightful differences that may form a basis for further investigation. Nonetheless, further efforts are needed to examine other implementations of MP to provide better accuracy and performance for use in clinical settings. Overall, there is a need to provide enhanced analysis methods and more solid interpretations of their outputs.
Keywords: Magnetoencephalography, Time-Frequency Analysis, Matching Pursuit, Dyadic Dictionary, Stochastic Dictionary, Low-Tc SQUIDs, High-Tc SQUIDs, On-Scalp MEG, Autism Spectrum Disorders.
Issue Date: 2019
Publisher: Chalmers tekniska högskola / Institutionen för mikroteknologi och nanovetenskap (MC2)
Collection:Examensarbeten för masterexamen // Master Theses

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.