Date of Award

Spring 2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Biomedical Engineering

First Advisor

Beardsley, Scott

Second Advisor

Liebenthal, Einat

Third Advisor

Ropella, Kristina

Abstract

The purpose of this dissertation was to apply joint independent component analysis (jICA) to electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) to characterize the neuroimage coupling between the two modalities. EEG and fMRI are complimentary imaging techniques which have been used in conjunction to investigate neural activity. Understanding how these two imaging modalities relate to each other not only enables better multimodal analysis, but also has clinical implications as well. In particular, Alzheimer’s, Parkinson’s, hypertension, and ischemic stroke are all known to impact the cerebral blood flow, and by extension alter the relationship between EEG and fMRI. By characterizing the relationship between EEG and fMRI within healthy subjects, it allows for comparison with a diseased population, and may offer ways to detect some of these conditions earlier. The correspondence between fMRI and EEG was first examined, and a methodological approach which was capable of informing to what degree the fMRI and EEG sources corresponded to each other was developed. Once it was certain that the EEG activity observed corresponded to the fMRI activity collected a methodological approach was developed to characterize the coupling between fMRI and EEG. Finally, this dissertation addresses the question of whether the use of jICA to perform this analysis increases the sensitivity to subcortical sources to determine to what degree subcortical sources should be taken into consideration for future studies. This dissertation was the first to propose a way to characterize the relationship between fMRI and EEG signals using blind source separation. Additionally, it was the first to show that jICA significantly improves the detection of subcortical activity, particularly in the case when both physiological noise and a cortical source are present. This new knowledge can be used to design studies to investigate subcortical signals, as well as to begin characterizing the relationship between fMRI and EEG across various task conditions.

Available for download on Wednesday, November 04, 2020

Share

COinS