Novel Transfer Function Based Approaches for Analysis of Resting-State Connectivity in Patients with Optic Neuritis

Date
2014-01-28
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Generating reliable imaging markers to study connectivity changes of brain regions associated with optic neuritis (ON) allows ON to be used as a system model for advanced treatment and pathology of multiple sclerosis (MS). Markers that can differentiate ON subject groups are obtained using new transfer function based approaches to characterize the connectivity paths of visual signal propagation. We suggest three thresholding methods to alleviate the effect of possible noisy peaks in the transfer function spectrum. Artificial neural networks (ANN) are used to provide an improved classifier via the integration of the proposed metrics. We propose a new simulated annealing approach to improve the sensitivity of ANN trained with small datasets such as obtained from ON data. We evaluate the ability of the metrics to differentiate between normal subjects and ON patients, with and without MS, using two-way and three-way receiver operator characteristics.
Description
Keywords
Engineering--Electronics and Electrical
Citation
Choudhury, S. H. (2014). Novel Transfer Function Based Approaches for Analysis of Resting-State Connectivity in Patients with Optic Neuritis (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/27136