Improved Classification of Optic Neuritis Patients Using Brain Visual Network Transfer Functions
Date
2017-12-22
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
One method to investigate how information propagates throughout brain networks is the transfer function (TF), which determines the amplification or attenuation of frequency components of signals from one brain region to another. Previous functional magnetic resonance imaging (fMRI) studies have demonstrated a disrupted cortical visual network (CVN) in the presence of optic neuritis (ON), which is often associated with the development of multiple sclerosis (MS). In this thesis, new approaches were developed to optimize TF metrics for resting state fMRI data for the purpose of distinguishing between the CVNs of healthy volunteers and ON patients. TF metrics were validated using receiver operating characteristics. Further development permitted the ability to distinguish CVNs between patients experiencing ON as a clinically isolated syndrome and ON patients with relapsing-remitting multiple sclerosis. Such a distinction has implications for the understanding of MS development and progression. Artificial neural networks were also explored as a potential tool to combine several TF metrics to further increase accuracy.
Description
Keywords
Image processing, Transfer function, Optic Neuritis, Multiple Sclerosis, Resting-state fMRI, Machine Learning, ROC analysis
Citation
Shahrabi Farahani, E. (2017). Improved Classification of Optic Neuritis Patients Using Brain Visual Network Transfer Functions (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.