Electrical and Computer Engineering ETDs

Publication Date

9-9-2010

Abstract

Medical image analysis techniques are becoming ever useful in allowing us to better understand the complexities and constructs of the brain and its functions. These analysis methods have proven to be integral in revealing trends in brain activity within individuals with mental disorders that are distinguishable from what would be considered normal' activity within healthy populations. This has led us to gaining a better understanding of functional connectivity within the brain, especially within populations suffering from mental disorders. Functional magnetic resonance imaging (fMRI) is one of the leading techniques currently being implemented to explore cognitive function and aberrant brain activity resulting from mental illness. Functional connectivity has investigated the associations of spatially-remote neuronal activations in the brain. Independent component analysis (ICA) is the leading analytical method in functional connectivity research and has been extensively implemented in the analysis of fMRI data, allowing us to draw group inferences from that data. However, there is mounting interest in the functional network connectivity (FNC) among components estimated through group ICA. This type of analysis allows us to delve further into the temporal dependencies among components or 'regions' within the brain. In this thesis, we investigate the implementation of group ICA and FNC analysis on two large-scale psychopathology studies -— the first from a multi-site study involving the comparison of schizophrenia patients with healthy controls using a sensorimotor task paradigm, and the second from an investigation in psychopathy in prisoners performing an auditory oddball task. In both studies, we analyzed the fMRI data with group ICA and implemented FNC analysis on the resulting ICA output. The purpose of these studies was to investigate differences in modulation of task-related and default mode networks, identifying any potential temporal dependencies among selected components. Our ultimate objective was to demonstrate the effective application of group ICA and FNC analysis together on two large-scale studies of two, distinct psychopathologies. The results of this combined method of analysis establish the practicality and general applicability of group ICA-FNC analysis in the growing fields of functional connectivity and functional network connectivity.'

Keywords

Brain--Magnetic resonance imaging--Data processing., Brain--Localization of functions--Multilevel models (Statistics), Psychology, Pathological--Data processing.

Sponsors

Department of Energy, National Institute of Health

Document Type

Thesis

Language

English

Degree Name

Computer Engineering

Level of Degree

Masters

Department Name

Electrical and Computer Engineering

First Committee Member (Chair)

Pattichis, Marios

Second Committee Member

Clark, Vince

Third Committee Member

Kiehl, Kent

Share

COinS