Graduation Semester and Year

2016

Language

English

Document Type

Thesis

Degree Name

Master of Science in Biomedical Engineering

Department

Bioengineering

First Advisor

Georgios Alexandrakis

Second Advisor

Hanli Liu

Third Advisor

Jean Gao

Abstract

It is necessary for distributed neural assemblies to work in collaboration to achieve the most elementary brain functions. Activities within segregated neural assemblies are integrated to process an incredibly diverse collection of cognitive tasks. Measures such as correlation, cross-correlation, coherence and phase delay make limited use of the information that is available with the advanced multichannel and multimodal recordings of the brain’s activity. They can also give spurious and false results. Most importantly, these bivariate measures are not capable of describing the interaction among multiple neural assemblies and the effective connectivity among them. Through causality analysis, we can identify these interactions and describe how information is transmitted across multiple parts, and make evident the presence of feedbacks in the network. In this thesis, using Granger’s causality and direct transfer function, I studied the brain's connectivity map at resting state. The direct transfer function is a normalized measure that uses the notion of Granger’s causality to determine how much each channel can be considered a source of information for another channel at a particular frequency. I found out that at resting state with the eyes open, sources of information at both the theta and the alpha frequency band are localized to the frontal region of the head. I also used a modified version of Granger’s causality, applicable to mixed frequency data, to investigate if the brain’s electrical activity measured through EEG is causal to its hemodynamic activity measured using fNIRs. Through this algorithm, I found no evidence of brain’s electrical activity being causal to its hemodynamic activity. I also investigated the relation between the spontaneous occipital alpha rhythm and the brain’s hemodynamic fluctuations. Although my results show that there is a significant correlation between the occipital alpha rhythm and the hemodynamic fluctuations at the electrode sites F8 and F7, causality analysis did not reveal any causal influence from the alpha rhythm to the hemodynamic activity.

Keywords

Mixed frequency data, Causality, EEG, fNIRS

Disciplines

Biomedical Engineering and Bioengineering | Engineering

Comments

Degree granted by The University of Texas at Arlington

Download

Share

COinS