<%BANNER%>

Reliability and Validity of Diffusion Imaging Methods Assessing Caudate to Frontal Lobe Function

Permanent Link: http://ufdc.ufl.edu/UFE0043030/00001

Material Information

Title: Reliability and Validity of Diffusion Imaging Methods Assessing Caudate to Frontal Lobe Function
Physical Description: 1 online resource (51 p.)
Language: english
Creator: NGUYEN,PETER T
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: CAUDATE -- COGNITION -- DIFFUSION -- DTI -- DWI -- FRONTAL -- MOW -- MRI -- RELIABILITY -- VALIDITY -- WHITE
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Diffusion magnetic resonance imaging (MRI) is a novel method of studying the human brain, in vivo. There has been a growing interest in utilizing diffusion imaging methods to examine white matter in the human brain. However, there is a relative gap in the literature examining the reliability of these methods. Given the rapid proliferation of diffusion imaging studies, it is vital to establish whether these methods are able to provide stable and consistent results in a single human participant over multiple separate trials. Diffusion tensor imaging (DTI) is a diffusion imaging post-processing method that quantifies white matter integrity and predicts directionality. Although it is effective at tracking highly directional regions such as the corpus callosum, it is limited in regions with crossing fibers, which may result in problems when examining cognitive constructs. The Mixture of Wishart (MOW) method is designed to provide better estimations of white matter integrity. The current study was divided into two aims: For the first aim, the test-retest reliability for white matter integrity and gray matter quantification were examined. Methods involved re-measuring an n-tridecane phantom and a single healthy participant over 10 separate MRI sessions. Reliability of diffusion data acquisition and post-processing using traditional DTI measurements and the novel MOW method were evaluated. Reliability of gray matter structure volumes (frontal lobe, caudate) was also examined. It was hypothesized that all white matter and gray matter variables would be acquired reliably across the separate trials. The second aim investigated the construct validity of a novel method of quantifying white matter integrity (edge weights), based on either the traditional DTI or MOW methods, for dissociating verbal fluency from visuospatial abilities as a function of white matter integrity between the left frontal lobe and left caudate. Methods involved examining white matter integrity between the left caudate and left frontal lobe for a sample of 39 older adults (16 non-demented Parkinson?s, 17 matched-controls) who had completed a set of neuropsychological tests. It was hypothesized that the MOW method, relative to the traditional DTI method, would provide more robust edge weights and stronger associations with verb fal fluency due to its increased ability to navigate and track crossing fibers. For aim one, diffusion and gray matter structural volumes were reliable across trials for both methods, although DTI and MOW edge weight values had two time points with noteworthy amounts of variance. For aim 2, left hemisphere frontal lobe to caudate connections revealed a dissociation between verbal fluency tasks (COWA) and visuospatial tasks (JLO) for both DTI and MOW methods. This is one of the first studies comparing the traditional DTI to the novel MOW approach. Regarding reliability, gray matter volumes and fractional anisotropy values were consistent across all ten trials although DTI and MOW edge weight values displayed significant variability in two time points. In terms of construct validity, results suggest a relationship between left frontal lobe to left caudate connectivity and verbal fluency, regardless of which diffusion post-processing method is used. No significant differences were found between the two imaging methods. Further studies are needed to fully assess the reliability and full utility of the DTI and MOW methods.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by PETER T NGUYEN.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Price, Catherine.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043030:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043030/00001

Material Information

Title: Reliability and Validity of Diffusion Imaging Methods Assessing Caudate to Frontal Lobe Function
Physical Description: 1 online resource (51 p.)
Language: english
Creator: NGUYEN,PETER T
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: CAUDATE -- COGNITION -- DIFFUSION -- DTI -- DWI -- FRONTAL -- MOW -- MRI -- RELIABILITY -- VALIDITY -- WHITE
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Diffusion magnetic resonance imaging (MRI) is a novel method of studying the human brain, in vivo. There has been a growing interest in utilizing diffusion imaging methods to examine white matter in the human brain. However, there is a relative gap in the literature examining the reliability of these methods. Given the rapid proliferation of diffusion imaging studies, it is vital to establish whether these methods are able to provide stable and consistent results in a single human participant over multiple separate trials. Diffusion tensor imaging (DTI) is a diffusion imaging post-processing method that quantifies white matter integrity and predicts directionality. Although it is effective at tracking highly directional regions such as the corpus callosum, it is limited in regions with crossing fibers, which may result in problems when examining cognitive constructs. The Mixture of Wishart (MOW) method is designed to provide better estimations of white matter integrity. The current study was divided into two aims: For the first aim, the test-retest reliability for white matter integrity and gray matter quantification were examined. Methods involved re-measuring an n-tridecane phantom and a single healthy participant over 10 separate MRI sessions. Reliability of diffusion data acquisition and post-processing using traditional DTI measurements and the novel MOW method were evaluated. Reliability of gray matter structure volumes (frontal lobe, caudate) was also examined. It was hypothesized that all white matter and gray matter variables would be acquired reliably across the separate trials. The second aim investigated the construct validity of a novel method of quantifying white matter integrity (edge weights), based on either the traditional DTI or MOW methods, for dissociating verbal fluency from visuospatial abilities as a function of white matter integrity between the left frontal lobe and left caudate. Methods involved examining white matter integrity between the left caudate and left frontal lobe for a sample of 39 older adults (16 non-demented Parkinson?s, 17 matched-controls) who had completed a set of neuropsychological tests. It was hypothesized that the MOW method, relative to the traditional DTI method, would provide more robust edge weights and stronger associations with verb fal fluency due to its increased ability to navigate and track crossing fibers. For aim one, diffusion and gray matter structural volumes were reliable across trials for both methods, although DTI and MOW edge weight values had two time points with noteworthy amounts of variance. For aim 2, left hemisphere frontal lobe to caudate connections revealed a dissociation between verbal fluency tasks (COWA) and visuospatial tasks (JLO) for both DTI and MOW methods. This is one of the first studies comparing the traditional DTI to the novel MOW approach. Regarding reliability, gray matter volumes and fractional anisotropy values were consistent across all ten trials although DTI and MOW edge weight values displayed significant variability in two time points. In terms of construct validity, results suggest a relationship between left frontal lobe to left caudate connectivity and verbal fluency, regardless of which diffusion post-processing method is used. No significant differences were found between the two imaging methods. Further studies are needed to fully assess the reliability and full utility of the DTI and MOW methods.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by PETER T NGUYEN.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Price, Catherine.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043030:00001


This item has the following downloads:


Full Text

PAGE 1

1 RELIABILITY AND VALIDITY OF DIFFUSION IMAGING METHODS ASSESSING CAUDATE TO FRONTAL LOBE FUNCTION By PETER T. NGUYEN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Peter T. Nguyen

PAGE 3

3 To my family

PAGE 4

4 ACKNOWLEDGMENTS I thank my family, friends, and everyone else who has been a part of my l ife. I also thank Dr. Catherine Price for her guidance as a research mentor and her assistance on the current project. I also thank Jared Tanner for all of his assistance and contributions. Finally, I thank my committee members Dr. Vonetta Dotson, Dr. Glenn Ashkanazi, and Dr. David Janicke, for their inputs and suggestions.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 ABSTRACT ................................................................................................................... 10 CHAPTER 1 INTRODUCTION .................................................................................................... 13 Introduction ............................................................................................................. 13 Gray Matter ............................................................................................................. 14 White Matter ........................................................................................................... 15 Diffusion Weighted Imaging (DWI) .......................................................................... 17 Diffusion Tensor Imaging (DTI) ........................................................................ 18 The Mixture of Wishart (MOW) Method ............................................................ 20 White Matter Quantification ..................................................................................... 20 Fractional Anisotropy (FA) ................................................................................ 21 Edge Weight ..................................................................................................... 21 Challenges .............................................................................................................. 22 Reliability .......................................................................................................... 23 Validity .............................................................................................................. 24 2 SPECIFIC RATIONALE AND AIMS ........................................................................ 27 Rationale ................................................................................................................. 27 Aim 1 Reliability of Imaging Methods ................................................................... 28 Aim 2 Cognitive Associations with Diffusion Imaging Methods ............................. 28 Hypothesis 1 ..................................................................................................... 28 Hypothesis 2 ..................................................................................................... 29 3 METHODS .............................................................................................................. 30 Aim 1 Reliability of Imaging Methods ................................................................... 30 Participants ....................................................................................................... 30 Imaging Protocol .............................................................................................. 30 Regions Of Interest (ROI) ................................................................................. 31 MRI Based NeuroAnatomical Quantification ................................................... 31 White Matter Quantification .............................................................................. 32 Stati stical Analysis ............................................................................................ 32 Aim 2 Cognitive Associations with Diffusion Imaging Methods ............................. 33 Participants ....................................................................................................... 33

PAGE 6

6 Imaging Protocol .............................................................................................. 34 Neuropsychological Protocol ............................................................................ 34 Statistical Analysis ............................................................................................ 34 4 RESULTS ............................................................................................................... 36 Aim 1 Reliability of Imaging Methods ................................................................... 36 Aim 2 Cognitive Associati ons with Diffusion Imaging Methods ............................ 36 5 DISCUSSION ......................................................................................................... 41 APPENDIX: GLOSSARY OF TERMS ........................................................................... 45 LIST OF REFERENCES ............................................................................................... 47 BIOGRAPHICAL SKETCH ............................................................................................ 51

PAGE 7

7 LIST OF TABLES Table page 3 1 Aim 2: Mean and Standard Deviation (SD) for Group Demographics ................. 35 4 1 Aim 1: Volumes of Grey and White Matter Regions of Interest across 10 trials 37 4 2 Aim 1: Fractional Anisotropy (FA) Values across 10 trials .................................. 37 4 3 Aim 1: Scanner Temperature Recordings Across 10 Trials ................................ 38 4 4 Aim 1: Left Frontal Lobe Left Caudate Edge Weights Across 10 Trials ........... 38 4 5 Aim 2: Edge Weight Cognitive Correlates .......................................................... 38

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Tensor ellipsoids. More spherical ellipsoids suggest higher degrees of isotropy (adopted from Kindlemann et al., 2004). ............................................... 25 1 2 MOW representations of 1, 2, and 3 fibers (Jian et al., 2007) ............................. 26 4 1 Fractional Anisotropy (FA) values over 10 trials ................................................. 39 4 2 Edge Weights over all 10 reliability trials ............................................................ 40

PAGE 9

9 LIST OF ABBREVIATIONS ADL Activities of daily living COWA Controlled Oral Word Association test CV Coefficient of variation DLPFC Dorsol ateral prefrontal cortex DTI Diffusion Tensor Imaging DWI Diffusion Weighted Imaging FA Fractional anisotropy IADL Instrumental activities of daily living JLO Judgment of Line Orientation test MOW Mixture of Wishart MRI Magnetic Resonance Imaging ROI Reg ion of interest

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science RELIABILITY AND VALIDITY OF DIFFUSION IMAGING METHO DS ASSESSING CAUDATE TO FRONTAL LOBE FUNCTION By Peter T. Nguyen May 2011 Chair: Catherine Price Major: Psychology Diffusion magnetic resonance imaging (MRI) is a novel method of studying the human brain, in vivo There has been a growing interest in ut ilizing diffusion imaging methods to examine white matter in the human brain. However, there is a relative gap in the literature examining the reliability of these methods. Given the rapid proliferation of diffusion imaging studies, it is vital to establis h whether these methods are able to provide stable and consistent results in a single human participant over multiple separate trials. Diffusi on tensor imaging (DTI) is a diffusion imaging post processing method that quantifies white matter integrity and predicts directionality. Although it is effective at tracking highly directional regions such as the corpus callosum, it is limited in regions with crossing fibers, which may result in problems when examining cognitive constructs. The Mixture of Wishart (MO W) method is designed to provide better estimations of white matter integrity The current study was divided into two aims: For the first aim the test retest reliability for white matter integrity and gray matter quantification were examined. Methods inv olved remeasuring an n tridecane phantom and a single healthy participant over 10 separate MRI sessions. Reliability of diffusion data acquisition and post -

PAGE 11

11 processing using traditional DTI measurements and the novel MOW method were evaluated. Reliability of gray matter structure volumes (frontal lobe, caudate) was also examined. It was hypothesized that all white matter and gray matter variables would be acquired reliably across the separate trials. The second aim For aim one, diffusion and gray matter structural volumes were reliable across trials for both methods, although DTI and MOW edge weight values had two time points with noteworthy amounts of variance. For aim 2, l eft hemisphere frontal lobe to caudate connections revealed a dissociation between verbal fluency tasks (COWA) and visuospatial tasks (J LO ) for both DTI and MOW methods investigated the construct validity of a n ovel method of quantifying white matter integrity (edge weights), based on either the traditional DTI or MOW methods, for dissociating verbal fluency from visuospatial abilities as a function of white matter integrity between the left frontal lobe and left caudate. Methods involved examining white matter integrity between the left caudate and left frontal lobe for a sample of 39 older adults (16 nondemented Parkinsons, 17 matchedcontrols) who had completed a set of neuropsychological tests. It was hypoth esized that the MOW method, relative to the traditional DTI method, would provide more robust edge weights and stronger associations with verb fal fluency due to its increased ability to navigate and track crossing fibers. This is one of the first studies comparing the traditional DTI to the novel MOW approach. Regarding reliability, gray matter volumes and fractional anisotropy v alues were consistent across all ten trials although DTI and MOW edge weight values displayed significant variability in two time points In terms of construct validity, r esults suggest a relationship between left frontal lobe to left caudate connectivity and verbal

PAGE 12

12 fluency regardless of which diffusion post processing method is used. No significant differences were found between the two imaging methods. Further studies are needed to fully assess the reliability and full utility of the DTI and MOW methods.

PAGE 13

13 CHAPTER 1 INTRODUCTION Introduction The development of brain imaging techniques has been pivotal to our understanding of the brain and its disorders. Specifically, magnetic resonance imaging (MRI) has provided us with an invivo means of studying brainbehavior relationships. More recently, greater interest has been placed with studying white matter fiber tracks in the brain using diffusion MRI, otherwise known as diffusion weighted imaging (DWI) or diffusion imaging. Despite the proliferation of DWI research, surprisingly few studies have examined the reliability of the imaging methodologies. This is especially troubling considering there are numerous sources of variability in diffusion MRI data acquisition and processing. Changes in MRI scanner characteristics, subject alignment, subject motion, signal noise, and temperature are examples of some of the sources that have been shown to affect variability in MRI data and diffusion fiber tracking (Holodny et al., 2005; Huang et al., 2004; Wakana et al., 2007). Although there is less concern with gray matter structural measurements, the same issues can be stated. As diffusion imaging calculations are contingent upon volumetric masks, it is important to demonstrate reliability with both the diffusion i mage processing as well as acquiring volumetric masks. Much like with a neuropsychological test, or any other measurement tool, it is essential to examine baseline reliability Knowledge of the reliability of the method would allow us to provide a more inf ormed interpretation of the results and whether caution should be exercised. It is imperative to establish that measurements can be made consistently and reliably in an individual over time. The current study examines

PAGE 14

14 the reliability of two diffusion imagi ng measurements and volumetric quantification across 10 sessions. Since the advent of MRI research, a considerable number of studies have examined how brain structures and certain cortical regions affect cognition and behavior. Historically, more attentio n and research was placed on the localization of functions in specific regions of the brain (Filley 2001). Although the associations between cognition and these gray matter regions have been investigated for many years l ittle research has explored the i nteractions between white matter fiber tracks with cognitive functioning (Filley 2001) A possible reason for this discrepancy can be attributed to a limitation in technology (Libon et al., 2004). While current models of diffusion weighted imaging can pr ovide important and useful information on overall white matter integrity, there are inherent weaknesses in the methodology that limit their ability to examine the complex white matter circuits associated with cognition. This is particularly true for regions where there are many crossing fibers (e.g., frontal white matter region). Thus, the second aim of the current study is to examine the construct validity of two different diffusion methods for examining one cognitive function verbal fluency and associ ated connectivity between the caudate and the frontal lobe. This study compares the traditional diffusion tensor imaging (DTI) m ethod (Basser et al., 1994) with the Mixture of Wisharts (MOW) method (Jian et al., 2007); a novel mathematical approach recentl y developed to address the limitations with crossing fibers in diffusion imaging. Gray Matter Brain matter can be divided into two main types of tissue: gray matter, which functions as the information centers of the brain and include the cortical and sub-

PAGE 15

15 cortical structures, and the more fibrous transmission lines known as white matter (Lezak et al., 2004) Gray matter primarily consists of neuronal cell bodies and glial cells, and has long been the topic of much interest and curiosity. This attention i s understandable given that prominent structures and areas in the brain such as the thalamus, basal ganglia, and cerebellum, are all composed of gray matter. In fact, the cerebral cortex, humanitys crowning jewel in the evolutionary ladder, is a gray matt er region. Furthermore, countless numbers of studies have demonstrated that a myriad of functions are produced by these structures. Our memories, emotions, and arguably our very minds stem from these regions of gray tissue in the brain. It therefore comes as no surprise that gray matter has historically taken center stage as the in vogue topic of study. While critical and informative discoveries were gleaned from the meteoric rise of functional localization studies, it also contributed to some inadvertent c onsequences. Brain and behavior relationships were relegated to overly reductionistic models of isolated brain regions, much like a modernday phrenology. However, increasing awareness of the interactions between these brain regions has led to mounting int erest in white matter (Filley 2001). White Matter It has long been established that the brain is composed of functionally distinct regions and structures. It is also understood that these structures do not function in isolation. Current literature supports the notion that the human brain is a system that is dependent upon the interplay of a variety of structures and regions through circuits and networks that are facilitated by white matter fibers (Filley 2001) Despite the discrepancy in the amount of research and attention towards gray matter, white matter plays a pivotal and equally important role in the human brain. In

PAGE 16

16 fact, the gray matter to white matter ratio i n healthy older adults ( 50 years) stands at approximately 1.1:1 (Miller 1980). In other words 47.6% of the adult brain is composed of white matter. For all of the function and ability provided by gray matter, white matter serves as the bottleneck. Proverbially, an old fashioned television cannot operate without a power cord. White matter is c omposed of bundles of myelinated axons in the brain and its primary function is to carry nerve impulses between neurons and facilitate communication between various regions and structures (Filley 2001) White matter dysfunctions have been implicated with i mpairments in cognitive domains associated with frontal/executive functions such as working memory, set shifting, and verbal fluency, while other domains such as memory encoding (Levin et al., 1992; Brown et al., 1988; Cummings et al., 1986) and visuospati al abilities (Pahwa et al., 1998; Bondi et al., 1993; Lichter & Cummings, 2000) remain relatively intact. Additionally, these cognitive impairments have been shown to significantly affect activities of daily living (ADLs) and instrumental activities of dai ly living (IADLs) (Rosenthal et al., 2010), demonstrating the importance of white matter functioning. Frontos ubcortical c ircuits In a classic 1986 study, Alexander, Delong, and Strick suggested that five parallel circuits connecting cortical regions to s ubcortical structures were associated with a range of tasks including motor function, personality, motivation, and cognition. Although separate and independent, each circuit follows a similar pathway and links specific areas of the frontal cortex to the st riatum / caudate, basal ganglia, and thalamus. Each circuit is named in accordance with their function or site of origin in the cortex. The motor circuit, which originates in the supplementary

PAGE 17

17 motor area, and the oculomotor circuit, which originates in the frontal eye fields, are both associated with motor functions. The lateral orbitofrontal circuit originates in the lateral orbitofrontal cortex and is associated with judgment, behavioral inhibition and emotion (Tekin & Cummings, 2002). The anterior cingul ate circuit originates in the anterior cingulate area and is associated with inhibition and creative thinking (Mesulam, 2000; Chow et al., 1999). The dorsolateral prefrontal circuit (DLPFC) is of particular interest to the current study, and originates in the dorsolateral prefrontal cortex and is associated with cognitive executive tasks such as set shifting and word fluency. Additionally, these cognitive executive functions are one of the principal components affected by Parkinsons disease and other disor ders (Tekin and Cummings, 2002) Although past studies have examined the relationship between cognition and neuroanatomical structures in these circuits, few studies have examined the white matter fibers that actually comprise the circuit. Diffusion imagi ng provides a unique window to quantitatively study and measure the integrity of these networks and how they influence cognition and behavior. With new diffusion imaging processing methods, it is possible to quantify the relationship between the physical i ntegrity of these circuits and cognitive functioning Diffusion Weighted Imaging (DWI) Presently, diffusion MRI, or diffusion weighted imaging (DWI), is the only methodology that allows for the in vivo visualization of white matter fiber pathways in the br ain. Conceptually, DWI can be viewed as an umbrella term encompassing a variety of different fiber tracking methodologies such as diffusion tensor imaging (DTI). DWI is based on the act of diffusion and is a naturally occurring transport process that occurs during molecular or particle mixing (JohansenBerg and Behrens 2009) Like

PAGE 18

18 a drop of dye spreading throughout a cup of water, this mixing process functions entirely without stirring or any other motion. It is understood that a particle will traverse a c ertain distance within a particular timeframe, otherwise referred to as displacement. In a completely free and unobstructed environment, this displacement distribution follows a Gaussian distribution. Because this distribution is influenced by the surr ounding microstructure, we can use a priori knowledge of a particular particles properties to quantify the actual observed displacement as a probe of the surrounding geometry (JohansenBerg and Behrens 2009) Using MRI there are acquisition sequences de signed to detect the effects of random particle dispersion. The brain is primarily composed of water, which contains hydrogen molecules. Due to this abundance of water, the dispersion and directionality of water can be quantified (Mosely 1990) As water di spersion is heavily influenced by its surrounding architecture, we are able to infer the existence of white matter fibers and pathways based on the directionality of water. White matter fibers contain a greater degree of directionality, or anisotropy. In tissues such as gray matter, water molecules tend to disperse and propagate independently of its surroundings, resulting in a more isotropic pattern. Diffusion that is entirely isotropic indicates that particles propagate equally in all directions. Conv ersely, dispersion that is propagated entirely along one directional plane can be described as being completely anisotropic (JohansenBerg and Behrens 2009) Diffusion Tensor Imaging (DTI) DTI (Basser et al., 1994) is one method for quantifying diffusion MRI data. It provides two unique insights into tissue microstructure; it quantifies diffusion anisotropy, which is a useful index of white matter integrity, and it estimates the directionality of

PAGE 19

19 axon fibers, which enables tractography. It is also currentl y the most common and widely used method of quantifying diffusion data. To help understand the process behind DTI, consider that displacement of water molecules in the brain operates within a 3D environment. Due to the fact that dispersion patterns may not be equal in all directions, a 3 x 3 symmetric matrix called a diffusion tensor is used to characterize these displacements. Alternatively stated, the diffusion properties of water in biological tissue are better represented by a tensor than a scalar value (Stejskal and Tanner, 1965). Although the underlying principles behind diffusion tensor imaging have been recognized since 1965, it was not until 1992 that Basser and colleagues formulated a mathematical model to estimate the diffusion tensor and provide the ability for tractography However, although DTI has had some success with tracking basic and highly directional white matter pathways such as the corpus callosum (Conturo et al., 1999; Mori et al., 1999; Basser et al., 2000), an intrinsic limitation with the methodology only allows it to reveal a single fiber orientation in each voxel (von dem Hagen & Henk elman, 2002; Tuch et al., 2002), thereby weakening its ability to resolve crossing fibers in the brain. To help understand this, consider that a voxel can contain hundreds of thousands of axon fibers that can adopt a wide range of configurations. The DTI model assumes a Gaussian distribution of the fiber directions in every voxel, which appears as an ellipsoid in 3D space. This model poses a problem for tractography and is especially problematic when dealing with areas such as the frontal lobes, which contain numerous crossing fibers (Schmahmann and Pandya, 2006) Given this limitation with DTI, there are very few studies examining frontal subcortical connectivity.

PAGE 20

20 However, a number of different methods have been developed in an attempt to circumvent the problems associated with crossing fibers such as the Mixture of Wishart (MOW) method. The Mixture of Wishart (MOW) M ethod The MOW method is a novel and mathematically derived diffusion MRI tractography reconstruction method designed to address the crossing fiber limitations found in DTI and other tractography methods (Jian et al., 2007). While DTI and many other reconstruction models assume that each voxel or fiber population is represented by a single tensor, the MOW method assumes that each fiber population possess a distribution of tensors where the mean tensor in this Wishart distribution yields the fiber orientation. In order to address the issue of crossing fibers, an additional formulation was proposed of using a mixture of Wishart distributions. Through the application of a Laplace transform, asymptotic fractals are used to model the MR signal delay and subsequent fiber orientations, in contrast t o the ellipsoids used in DTI. Jian et al. present a rigorous mathematical justification for this formulation in their 2007 study. The investigators compared MOW to traditional DTI tractography on simulated and rat data. On an experiment using simulated cr ossing fibers, the MOW method was able to outperform other DTI based methods in predicting the actual fiber orientations. An additional experiment using actual rat data revealed that the MOW method was more effective at predicting fiber orientations in the central region of the optic chiasm where there are crossing ipsilateral fibers. White Matter Quantification In addition to allowing the visualization of white matter through tractography, DWI can be used to quantify white matter differences in a number of ways.

PAGE 21

21 Fractional Anisotropy (FA) Fractional anisotropy (FA) is the most common index of white matter integrity. Originally described in Basser and Pierpaoli (1996), fractional anisotropy assigns a numerical value, on the scale of 0 to 1, to the degree of water directionality in a voxel or specific region. Higher levels of anisotropy suggest more directionally organized tissue, while lower levels of anisotropy can imply lower degrees of organization. For example, a brain ventricle, which has high water c ontent but no surrounding structure would theoretically have an FA value of 0. The corpous callosum on the other hand, is a highly organized white matter region where water is dispersed in a very parallel and directional fashion, resulting in FA values closer to 1. A limitation with FA is although it provides a summary of diffusion d irectionality in all directions, it provides little information about the specific directionality driving the diffusion differences ( JohansenBerg and Behrens 2009). As a res ult, FA alone is an insufficient means for studying regions with complicated white matter organization. Edge Weight Considering the limitations with FA, additional white matter variables are required to quantitatively measure the connectivity strength of the frontosubcortical circuits. Hagmann and colleagues originally proposed a method of mapping entire brain networks by using tractography methods (Hagmann et al., 2007). Using highpowered diffusion spectrum imaging, they produced a wholebrain diffusio n map, which was then combined with thousands of indiscriminately labeled regions of interest (ROIs). These ROIs functioned as nodes and were used to estimate the density of white matter connections between each and every region of interest A follow up st udy revealed that these brain networks appeared to follow a small world connectivity pattern, whereby

PAGE 22

22 fibers consolidated upon specific central hubs in the brain (Hagmann et al., 2008). These findings mirrored past studies on structural human brain networks (Hagmann et al., 2005) and were further substantiated with functional imaging studies. Although diffusion spectrum imaging is not practical for regular human use due to extensive scanning times, the same principles employed by Hagmann and colleagues can be attempted with the DTI and MOW methods. After constructing a full brain connectivity map, we can use specific ROIs to threshold the fibers and limit the output to include only fibers that connect those regions. For example, the frontal lobe and caud ate could be used as ROIs to examine only the fibers that connect those two regions. These resulting edge weights, or fiber bundles, can be used to characterize the connectivity strength of individual white matter tracts, including the frontosubcortical circuits. Challenges With the continued technological advancements of neuroanatomical imaging methods, we are now able to examine the brain with a degree of resolution that has been previously unattainable. In spite of this, as psychologists and neuropsy chologists, we are not merely interested in what the brain looks like; we are interested in what it does. Without a means of associating the structural anatomy with cognition and behavior, we are left with only half of the story. Although gross disruptions to major fiber circuits could clearly be visualized using tractography, more sensitive measures are required to assess and quantify the multitude of fiber circuits associated with cognition and behavior. Even with the advancements of imaging technology, t here are still numerous obstacles in place that have impeded our ability to associate cognition to white matter integrity using diffusion imaging.

PAGE 23

23 Reliability As a general rule, it is crucial to account for the reliability of a measurement tool before any conclusions can be derived from it. Regardless of whether the instrument is a thermometer, a neuropsychological test, or a neuroimaging method, we need to know whether the observed values are truly reliable and not caused by external variation due to the i nstrument. These principles are especially important for novel and continually developing methodologies such as MRI and DiffusionMRI. Diffusion imaging in particular, is subject to numerous sources of possible variation. In order to acquire MR images, coils in the scanner create magnetic field gradients both inside the scanner bore and outside to the surrounding area. These external magnetic fields can generate electrical eddy currents in nearby conductive surfaces, which in turn can warp and distort a diffusionweighted image. Areas around tissue/air boundaries such as the sinuses often produce magnetic field inhomogeneities, resulting in shifting and warping of the image. Furthermore, diffusion imaging is also quite sensitive to temperature effects and motion artifacts caused by cardiac pulsation and head movement in the scanner. (JohansenBerg and Behrens, 2009) Although less concerning, many of the same issues apply to structural T1 imaging as well. Subject related factors such as hydration status (Wa lters et al., 2001) and head motion may result in variability in MRI derived volumes. Additionally, instrument related factors such as field strengths, scanner software, and post processing parameters are all potential sources of variability. In structural morphometric studies, variability often occurs with designating and extracting regions of interest such as cortical and

PAGE 24

24 subcortical volumes (e.g., caudate, thalamus, frontal lobe). Moreover, variability in the ROI can consequently impact the accuracy of w hite matter connectivity measurements. As mentioned before, there are many obstacles in our goal of using diffusion imaging to study cognition. All of the aforementioned sources of variability can potentially hamper our ability to produce consistent and reliable measurements, thereby preventing us from detecting any real associations with cognition. Therefore, it is imperative that we have an understanding of the expected variability, lest we naively venture forth. Validity The next step after establishing reliability is to examine the construct validity of the method by observing its convergent and discriminant validity. Convergent and discriminant validity are considered subtypes of construct validity and can provide evidence whether our diffusion MRI measurements actually correlate with our hypothesized constructs. Neither convergent nor discriminant validity alone is sufficient; only if evidence for both convergent and discriminant validity can be demonstrated would we have evidence for construct validit y. Specifically, we are examining whether diffusion MRI can effectively associate specific cognitive functions, such as verbal fluency, with implicated white matter pathways, while maintaining no relationship with another cognitive function such as visuosp atial ability. We expect our diffusion MRI measurements to show convergence with a neuropsychological test assessing verbal fluency while being able to diverge from other functions that theoretically should not be involved.

PAGE 25

25 Verbal fluency is one type of cognitive function that has specific interest to neuropsychologists. Verbal fluency has been implicated in a multitude of disorders including dementia, Parkinsons disease (Zgaljardic 2006), epilepsy (Yogarajah 2010), and multiple sclerosis (Henry and Beat ty 2006) and has also been impacted following deep brain stimulation. Alexander et al first theorized that verbal fluency involves activation of the dorsolateral prefrontal cortex, particularly in the left hemisphere, which subsequently implicates subcorti cal structures such as the caudate via the frontosubcortical projections. However, examining this pathway is challenging due to the presence of crossing fibers within the frontal lobe. Consequently, further investigation of this cognitive domain and its i mplicated white matter connections is warranted, especially considering the potential clinical applications associated with it. Neuroanatomical correlates of verbal fluency may prove to be a useful tool to assist with quantifying and differentiating various cognitive disorders. Figure 11. Tensor ellipsoids. More spherical ellipsoids suggest higher degrees of isotropy (adopted from Kindlemann et al., 2004).

PAGE 26

26 Figure 12. MOW representations of 1, 2, and 3 fibers (Jian et al., 2007)

PAGE 27

27 CHAPTER 2 SPECIF IC RATIONALE AND AIMS Rationale White matter fibers serve a critical and essential role in the human brain. Increasing interest in white matter has led to a rapid proliferation of diffusion imaging studies. However, studies on the reliability of these methods have trailed in comparison. There exists a significant lack of informati on regarding the test retest reliability of these methods in a single healthy participant over multiple trials. This is especially troubling, considering that any and all results produced through similar imaging methods must be viewed in the context of the measuring instruments reliability. Consequently, the current study investigated the reliability in two current diffusion imaging methods: the basic and widespread DTI method, as well as the more novel MOW method. Cognition and behavior have been implicated with frontosubcortical white matter fiber circuits that connect subcortical structures to cortical regions. The DLPFC circuit is a frontosubcortical circuit that has been spec ifically implicated with cognitive executive tasks such as set shifting and verbal fluency. Additionally, these cognitive executive functions are one of the principal components affected by Parkinsons disease and other disorders. Neuro anatomical correlat es of these cognitive abilities would provide a useful clinical tool in differential diagnoses and assessing clinical outcomes. However, until recently, inherent limitations in current diffusion imaging methodologies have limited the ability to investigat e the association between cognition and white matter integrity in these circuits. New post processing methods of diffusion imaging, such as the MOW method, have been developed in an attempt to address these limitations. This study also investigated the DTI and MOW methods ability to differentiate verbal fluency

PAGE 28

28 with left fronto subcortical circuit integrity in the human brain from a visuospatial functioning, which is typically allocated to a different region of the brain. Aim 1 Reliability of Imaging Me thods The first aim of the current study is to investigate the test retest reliability of structural and diffusion MRI data acquisition and processing over multiple trials. Hypothesis 1 It is hypothesized that the testretest reliability of diffusion MRI data acquisition and processing will be stable and consistent for both DTI and MOW methods acquired for 10 separate sessions in the same participant. Furthermore, the MOW method is expected to be more reliable than the DTI method. Gray matter volumes invol ving the frontal cortex and the caudate nucleus are expected to be reliable across measurements. DWI variables such as fractional a nisotropy and edge weight are also hypothesized to be reliable over mul tiple trials within regions of high directionality (co rpus callosum) relative to regions of lower directionality (frontal white matter). An ntridecane MRI phantom and one individual will be placed in a scanner at the same time for 10 separate sessions. Imaging software tools (Freesurfer, ITK Snap, TrackTool s) will be used to process the data identically for all 10 trials. Temperature and scanner properties were recorded or noted. Reliability will be calculated using coefficients of variation. Aim 2 Cognitive Associations with Diffusion Imaging Methods Hypo thesis 1 The second aim of the current study is to investigate the construct validity of using the DTI and MOW methods to dissociate verbal fluency from visuospatial functioning for fibers connecting the left frontal lobe to the left caudate. It is hypothesized that white matter fiber tracks connecting the caudate to the frontal lobe will be able to dissociate

PAGE 29

29 cognitive tasks specifically implicated with frontosubcortical circuits involved in verbal fluency A sample of 39 older adults will receive diffusi on MRI scanning and neuropsychological testing. Connectivity strength measurements (edge weights) will be calculated from left frontal lobes to left caudates as a proxy for left frontocaudate circuit integrity. Pearson correlations will be examined between edge weights and scores on neuropsychological tests. It is hypothesized that edge weights will be significantly correlated with a test of verbal fluency, but not with a test of visuospatial functioning. Hypothesis 2 Historically, diffusion tensor imagin g has had difficulties examining regions of complex white matter organization. Due to interference from crossing fibers in the frontal regions associated with verbal fluency, it is hypothesized that the MOW method will provide a stronger and more robust as sociation with verbal fluency compared to the traditional DTI method. Fisher r to z transformations will be used to examine whether there are significant differences in the cognitive correlations between the DTI and MOW methods. Neither method is expected to show an association with visuospatial functioning.

PAGE 30

30 CHAPTER 3 METHODS Aim 1 Reliability of Imaging Methods Participants In order to establish the reliability of our diffusion imaging methodology, it was essential to reduce any external sources of var iation that may have arisen from utilizing different human subjects. As a result, one single healthy participant was scanned and his data was processed across 10 separate trials evenly spaced over the course of 1 month. All scans were conducted at the same time period. Our participant was a 38year old healthy adult male with no reported diseases or abnormal brain status. The participant was scanned using identical parameters during each scan session. Each time point functioned as an independent subject, regardless of time point. (IRB 372 2010). MRI Phantom. In order to assess the reliability of the imaging method, it was important to control for variability that may be inherent from using human participants. An n tridecane solution was utilized because it is relatively homogenous, isotropic, and its properties are sensitive to diffusion imaging sequences. The MRI phantom was scanned immediately before the human participant during each scan session. The phantom was placed in the identical position every tri al using fastening clamps and a straightening device. Imaging Protocol The data were acquired with a Siemens 3T Verio scanner at the same time of day for every session. Single shot EPI diffusion weighted images were acquired with diffusion gradients applied along 6 directions (b = 100/ mm2) and 64 directions

PAGE 31

31 (1000s/mm2). Imaging parameters were set at 73 contiguous axial slices with a slice thickness of 2mm, and TR/TE = 17300/81ms. Separate T1 and T2weighted sequences were acquired with the following param eters: 176 contiguous slices, slicethickness of 1mm, TR/TE = 2500/3.77ms. Past studies have reported inadequate contrast between gray and white matter boundaries when using a single T1 scan (Price et al., 2010), thus t wo T1 weighted volumetric sequences w ere acquired and averaged for each participant to optimize the signal to noise ratio. Regions Of I nterest (R OI) To calculate the strength of white matter connectivity in a particular region, a minimum of two endpoints are required. Thus, In order to study the frontosubcortical circuits, the connectivity strength between the left frontal lobes and left caudates were examined. Although it has been suggested that executive functioning and fluency performance is primarily localized to the left dorsolateral pre frontal cortex, ambiguous morphological boundaries and lack of consensus on the DLPFC resulted in difficulties establishing a consistent and reliable ROI. Consequently, for the current study, the frontal lobe was examined as a proxy of the DLPFC, as its cl ear sulcul boundaries allowed for consistent and reliable structural ROI masks. For increased sensitivity, regions of interest and analyses were constrained to the left hemisphere. MRI Based Neuro Anatomical Quantification Image analyses were performed us ing free or commercially available software. In order to acquire white matter connection measurements, raw T1 volumetric data were converted into ROI masks using Freesurfer (Fischl et al., 2002) (http://surfer.nmr.mgh.harvard.edu), a software tool specific ally developed for the reconstruction of cortical and subcortical structures of the brain from structural MRI

PAGE 32

32 data. All of the automatically segmented caudate and frontal lobe masks were then manually inspected by the investigator for errors using ITK SNA P (Yushkevich et al., 2006) ( http://www.itksnap.org ), a visualization and editing tool that provides the ability to examine ROI masks in three dimensional space, as well as manually cleanup or alter existing masks. Dice similarity coefficients, which meas ure the amount of spatial overlap between threedimensional masks, was greater than .95 in repeated trials, suggesting very high intrarater reliability. White Matter Quantification FA values were acquired from two different white matter regions: a region with highly directional white matter organization (anterior corpus callosum), and another region with relatively complex white matter organization and many crossing fibers (rostral middle frontal white matter region). FA values were also taken from the ntridecane phantom to serve as a baseline. Edge weights were calculated using TrackTools, an inhouse software program developed by our collaborators. First, whole brain networks were calculated using DTI and MOW methods. Caudate and frontal lobe ROI mas ks were then used to filter and threshold the network to output only fibers that connected the left caudates and left frontal lobes. The resulting fibers, or edge weights, were used to correlate with cognitive variables. Statistical Analysis To establish a reliability index of the diffusion method across multiple trials, we calculated means and standard deviations of the edge weights and FA values across all 10 trials. A coefficient of variation (CV) was calculated as a ratio of the measurements standard deviation divided by its mean and multiplied by 100. The CV provides an

PAGE 33

33 intuitive estimate of variance, expressed as a relative percentage, and is the most commonly reported statistical measure in DTI test retest reliability studies. A Fisher r to z trans formation was used to assess the significance of the difference between two correlation coefficients. We calculated the Fisher r to z transformation to compare MOW and DTI edge weight reliability. Aim 2 Cognitive Associations with Diffusion Imaging Methods Participants In order to investigate the construct validity of our diffusion imaging methodologies, a combined sample of 39 individuals that included cognitively intact right side onset Parkinsons disease patients and matche d healthy controls was exam ined from data that were collected as part of a larger study (IRB 4272007). Group demog raphics are presented in table 31 Parkinsons disease patients were included in the sample in order to create a broader range of variation in cognitive performance. A s this study aim seeks to examine the diffusion methods ability to dissociate cognitive function, it was important to include a sample population that was impaired in selective cognitive domains, while remaining intact in others. Stated otherwise, this st udy is not looking to compare Parkinsons disease patients to normal controls; the aim is to correlate cognitive performance with diffusion imaging variables, which is facilitated by having a broad range of cognitive scores. All participants signed a conse nt form approved by the Institutional review board. The study procedures followed principles defined in the Declaration of Helsinki.

PAGE 34

34 Imaging Protocol All participants received the identical imaging protocol as discussed in the first aim. Structural ROIs and edge weight values were individually calculated for each participant. Neuropsychological Protocol Sensitive and reliable cognitive measures with normative data from older populations were used to assess different aspects of cognitive functioning; verbal fluency and visuospatial functioning As previously stated, frontosubcortical circuits including the DLPFC circuit have been specifically implicated with frontal executive functioning. Other cognitive domains such as visuospatial functioning are believed to be independent of frontosubcortical circuit integrity. We assessed the following abilities: 1) Verbal fluency was assessed with the Controlled Oral Word Association test (COWA; Spreen and Strauss 1998). This is a timed auditory verbal fluency task that requires the participant to rapidly generate words starting with the letter F, then the letter A, and finally the letter S. The dependent variable was the number of words produced within one minute for each letter. The number of words produced for each letter was summed for each participant. Raw scores were converted into standardized z scores using published norms. 2) Visuospatial perceptual ability was assessed with the Judgment of Line Orientation test ( JLO; Benton 1983), which requires participants to co rrectly match line orientations. The dependent variable was the number of correctly matched items. Raw scores were converted into standardized z scores using published norms. Statistical Analysis Pearson correlations were conducted between neuropsychological test scores and diffusion imaging variables (edge weights). Fisher r to z transformations were used to compare cognitive correlations between DTI and MOW methods.

PAGE 35

35 Table 31. Aim 2: Mean and Standard Deviation (SD) for Group Demographics Mean SD Age 66.67 5.23 Education 16.03 2.58 MMSE 29.17 .94 WASI IQ 116.83 11.29 MMSE = Mini Mental State Exam (Folstein et al., 1975) WASI IQ = Wechsler Abbreviated Scale of Intelligence

PAGE 36

36 CHAPTER 4 RESULTS Aim 1 Reliability of Imaging Methods Mean v olumetric values and coefficients of variation for gray and white matter ROI are presented in Table 41 CV values ranged from 1.4% to 2%, suggesting very little variability in acquiring volumetric masks across the 10 reliability trials of a single individual FA values for the different regions over 10 trials are graphed in Figure 41 Mean FA values and CVs calculated from all 10 reliability trials are presented in Table 4 2 Temperature recordings are presented in table 43. As expected, regions of higher white matter organization produced the highest FA values while the ntridecane phantom, produced the lowest. Similarly, the CVs for the three regions follow a similar pattern with the anterior corpus callosum producing the most reliable FA values. Mean edge w eight values of frontal lobe to caudate connections were calculated from both DTI and MOW methods and are presented in Table 44 Edge weights over all 10 trials are presented in Figure 42. CVs of edge w eights in both DTI and MOW methods sugges t that the values were variable across the 10 trials. A Fisher r to z transformation revealed no significant differences in reliability between the MOW and DTI methods (z = .17, p = .865). Aim 2 Cognitive Associations with Diffusion Imaging Methods Pea rson correlations were conducted between edge weights from the MOW and DTI methods and neuropsychological test scores and are presented in table 45 Neither DTI nor MOW edge weight calculations correlated significantly with COWA or JOLO variables at a thr eshold of alpha = .05, although positive correlations with COWA

PAGE 37

37 scores appear ed to approach significance for both DTI and MOW methods (p = .084, p = .090, respectively). Edge weights were clearly unrelated to JOLO test scores. Fisher r to z calculations revealed no significant differences between cognitive correlations produced by the DTI and MOW methods (p > .05). Table 41. Aim 1: Volumes of Grey and White Matter Regions of Interest across 10 trials Region Mean (mm^3) SD CV Anterior CC 955.30 13.26 1.4 0 RMF WM 14751.00 260.31 1.8 0 Left Frontal Lobe 166264.80 1602.91 1.0 0 Left Caudate 3385.60 66.91 2.0 0 RMF WM = Rostromedial frontal white matter Anterior CC = Anterior Cingulate Cortex Table 42. Aim 1: Fractional Anisotropy (FA) Values across 10 trials Region Mean SD CV Anterior CC .52 .02 4.5 0 RMF WM .31 .02 8.4 0 n Tridecane Phantom .14 .01 9.2 0 RMF WM = Rostromedial frontal white matter Anterior CC = Anterior Cingulate Cortex

PAGE 38

38 Table 43. Aim 1: Scanner Temperature Recordings A cross 10 Trials Mean Celsius SD CV 18.10 .74 4.00 Table 44. Aim 1: Left Frontal Lobe Left Caudate Edge Weight s Across 10 Trials Mean EW SD CV DTI 1892.78 1046.22 55.3 MOW 2768.69 1386.41 50.1 DTI = Diffusion Tensor Imaging Method MOW = Mixt ure of Wishart Method Table 4 5. Aim 2: Edge Weight Cognitive Correlates DTI MOW Fisher z COWA Z score Pear s on r .28 .27 P = .98 Sig (2 tailed) .08 .09 N 39 39 J LO Z score Pearson r .01 .07 P = .73 Sig (2 tailed) .94 .67 N 39 39 De notes trending towards significance COWA = Controlled Oral Word Association Test JLO = Judgment of Line Orientation Test

PAGE 39

39 Figure 41. Fractional Anisotropy (FA) values over 10 trials

PAGE 40

40 Figure 42. Edge Weights over all 10 reliability trials

PAGE 41

41 CHAPTER 5 DISCUSSION Diffusion weighted imaging is a novel tool that provides the unique ability to investigate white matter integrity in the human brain. The use of diffusion imaging methodologies in research studies has increased rapidly, despite a relative lack of information on overall reliability of the methods. The current study examined the reliability of producing quantitative measures of white matter integrity using diffusion weighted imaging. Structural volumes and regions of interest were acquir ed with high reliability. FA values were calculated from automatically segmented white matter regions in the brain as well as from an ntridecane phantom. With regards to reliability, gray matter volumes of the frontal lobes and caudates were acquired and extracted consistently across all ten sessions. FA values calculated from a complicated white matter region (rostromedial frontal white matter), a simple white matter region (corpus callosum), and an isotropic ntridecane phantom were also reliable over al l ten scanning sessions. Expectedly, FA values from the more complicated white matter regions showed more variability across trials, especially relative to the ntridecane phantom (ref). In order to study white matter regions associated with cognition, com mon diffusion imag ing variables such as FA values a re insufficient due to their inability to examine complicated white matter circuits. The reliability of MOW vs. DTI to quantifying white matter integrity between two regions was also examined. Edge weights were calculated as an index of frontal lobe to caudate connectivity for both DTI and MOW methods. For both methods, edge weights were highly variable over the ten trials, with the much of the variance being attributed to the 4th and 5th trials. No significant differences in reliability were observed between the two

PAGE 42

42 methods. The cause of the fluctuations at the 4th and 5th trials are currently unclear, although current hypotheses include possible unrecorded changes with the scanner or scanner software, erro rs or glitches in data processing, or other factors that have not been accounted for While temperature has been shown to affect diffusion imaging variables (JohansenBerg and Behrens, 2009), temperature recordings were consistent across all ten trials. It is also possible that the variability in the data may be due to inherent problems that are intrinsic to the diffusion imaging methods. Another likely factor contributing to the variability was our choice of ROIs. Primarily a functionally derived region, t here is currently no universally agreed upon demarcation of the dorsolateral prefrontal cortex (Tisserand et al., 2002; Burgmans et al., 2009, Ranta et al., 2009). Due to concerns over acquiring consistent ROIs from poor morphological boundaries of the dor solateral prefrontal cortex, the current study utilized the entire frontal lobe masks as a proxy for the DLPFC. As edge weights are composed of all of the fibers between two regions, it is possible that our frontal lobe ROI is too large and introduces too much external variance in the form of extraneous fibers. With regards to construct validity, the current study examined whether the diffusion imaging methods were able to effectively dissociate a verbal fluency task from a visuospatial task by examining c onnections between the left caudate and left frontal lobe, which encompasses the DLPFC circuit and has been implicated with performance on frontal executive tasks including verbal fluency. Although the findings did not fully reach statistical significance, edge weight showed a mild association with a verbal fluency task, but not on the visuospatial task. This suggests potential cognitivewhite matter specificity in edge weights values were able to correlate with

PAGE 43

43 performance on a verbal fluency t ask, but not on a visuospatial task showing a dissociation of cognitive functioning Our observed effects may have also been limited by our ROI. While the current study utilized an ROI of the entire frontal lobe, verbal fluency has been more specifically localized to the DLPFC region of the frontal lobe. Consequently, the large ROI may have minimized the strength of our findings. Overall, these results suggest that diffusion measurements of FA can be reliable over time; however, there is variability in edg e weight calculations. This pattern was consistent for both DTI and MOW methodologies. It is suggested that edge weights can be useful for assessing a hypothesized cognitiveanatomical construct, although further studies are needed. The clinical implications associated with these findings suggest that diffusion weighted imaging may provide important information on neuroanatomical correlates of cognition, which may potentially aid in differential diagnoses and assessing clinical outcomes. Diseases and disor ders such as Parkinsons disease, multiple sclerosis, and epilepsy are all associated with impairments in verbal fluency and other frontal executive dysfunctions. A neuroanatomical index of white matter integrity could provide important information towards understanding the nature of these cognitive deficits. Furthermore, cognitive decline is often an important indicator of disease progression. Future longitudinal studies may be able to examine white matter integrity before any observable behavior effects, aiding with treatment planning and prognosis. Limitations of the current study primarily involve our ROI selection and power. The large size of the frontal lobe ROI may have contributed to variability in connectivity analyses. The MOW method was designed to address regions of crossing fibers in specific white matter reg ions. Using the entire frontal lobe may include an excess of

PAGE 44

44 extraneous fibers such that the MOW method no longer has the specificity to detect subtle differences in specific fiber circuits associated with just the dorsolateral prefrontal region of the frontal lobe. Future aims should seek to utilize more precise ROIs, despite the ambiguous architectural boundaries of the DLPFC. One possibility would be to apply thresholding rules based on a priori knowledge of where the fiber tracts are hypothesized to connect. Optimally, functional MRI studies could be incorporated to provide a functionally activated ROI for use with diffusion tractography. Additionally, the cognitive associations were limi ted to 39 individuals. Estimates of power suggest that 82 participants would be needed to demonstrate significance at an alpha of .05. In conclusion, diffusion imaging is a novel and exciting tool in understanding how the human brain functions. However, as a novel tool, few studies have examined the basic reliability properties of the methods. In this study we utilized two different methods of diffusion tractography as well as a novel method of quantifying white matter integrity. Our results suggest that f urther studies should attempt to elucidate the nature of the variability, especially in regards to examining white matter regions that are associated with cognition. Nonetheless, the results suggest that the current methods are able to effectively associat e with cognitive functioning.

PAGE 45

45 APPENDIX GLOSSARY OF TERMS Anisotropic Diffusion that is restricted or directionally dependent. Asymptotic Fractal A rough or fragmented geometric shape that can be split into parts. Used by the MOW method to model diff usion directionality. Caudate Sub cortical structure located within the basal ganglia. Diffusion Weighted Imaging A magnetic resonance imaging method that produces in vivo images of biological tissues weighted with the microstructural characteristic s of water diffusion. Diffusion Tensor Imaging A relatively simple method of quantifying diffusion imaging data by estimating directionality and anisotropy of white matter in the brain. Utilizes a tensor matrix to characterize diffusion displacement. D orsolateral prefrontal cortex (DLPFC) Cortical area roughly equivalent to Brodmann areas 9 and 46, although precise structural boundaries are currently nonexistent. Plays an important role with working memory and receives inputs from the basal ganglia. Edge Weight Numerical value used as a measure of connectivity strength between two regions. Fractional anisotropy Scalar value between zero and one that describes the degree of anisotropy of a diffusion process. It is commonly used in diffusion imag ing as a proxy of white matter integrity. Gray Matter Major component of the central nervous system which contains nerve cell bodies as opposed to myelinated nerve fibers. Gaussian Distribution A normal or continuous probability distribution that te nds to cluster around a single mean value. Isotropic Diffusion that is unrestricted (or equally restricted) in all directions. Magnetic Resonance Imaging Medical imaging technique used to visualize detailed internal structures. Laplace Transform An integral transform used to resolve a function into its moments. Mixture of Wishart Novel method of quantifying diffusion imaging data using a distribution of tensors model. Addresses inherent difficulties resolving crossing fibers with diffusion tensor imaging.

PAGE 46

46 Phantom A liquid solution with known MR imaging properties. Often used to calibrate or trouble shoot MRI scanners. Region of Interest (ROI) A selected subset of information within a dataset identified for a particular reason. In imaging studies, ROIs are often used to constrain analyses to specific defined regions or structures of the brain. Sequences Preselected set of defined radiofrequency pulses and gradients. Different sequences are used to favor the signal of a particular tissue. T ensor Arrays of numbers or functions that characterize the properties of a physical system. Tractography Image processing procedure that utilizes MRI and diffusion weighted imaging data to construct and visualize neural tracts and fibers. White Matter Major component of the central nervous system that consists mostly of myelinated axons.

PAGE 47

47 LIST OF REFERENCES Basser, P. J., Mattiello, J., & LeBihan, D. (1994). Estimation of the effective self diffusion tensor from the NMR spin echo. J Magn Reson B, 103(3), 247254. Basser, P. J. & Bihan, D. (1992). Fiber orientation mapping in an anisotropic medium with NMR diffusion spectroscopy. In: Book of abstracts: Eleventh Annual Meeting of the Society of Magnetic Resonance in Medicine, Berkeley, CA, 1221. Basser, P. J., Pajevic, S., Pierpaoli, C., Duda, J., & Aldroubi, A. (2000). In vivo fiber tractography using DT MRI data. Magn Reson Med, 44(4), 625 632. Benton, A. L., Hamsher, K., Varney, N. R., et al. (1983). Contributions to neuropsychological assessment. New York NY : Oxford University Press. Bondi, M. W., Kasniak, A. W., Bayles, K. A., et al. (1993). Contributions of frontal systems dysfunction to memory and perceptual abilities in Parkinson's disease. Neur opsychology. 7, 89 102. Brown, R.G., Marsden, C.D. (1988). An investigation of the phenomenon "set" in Parkinson's disease. Mov Disord, 3, 152 161 Burgmans, S., van Boxtel, M. P. J., Smeets, F., Vuurman, E. F. P. M., Gronenschild, E. H. B. M., Verhey, F. R J., et al. (2009). Prefrontal cortex atrophy predicts dementia over a six year period. Neurobiology of Aging, 30(9), 1413 1419. Chow, T. W. & Cummings, J. L. (1999). Frontal subcortical circuits. In: Miller B., Cummings J. L., editors. The human frontal lobes. New York: Guilford Publications, 326. Conturo, T. E., Lori, N. F., Cull, T. S., Akbudak, E., Snyder, A. Z., Shimony, J. S., et al. (1999). Tracking neuronal fiber pathways in the living human brain. Proc Natl Acad Sci U S A, 96 (18), 1042210427. Cu mmings, J. L. (1986). Subcortical dementia. Neuropsychology, neuropsychiatry, and pathophysiology. Br J Psychiatry, 149 682 697. Fick, A. (1855). Concerns diffusion and concentration gradiant. Ann Phys, 170. Filley, C. M. (2001). The Behavioral Neurolog y of White Matter New York, NY: Oxford University Press. Fischl, B., Salat, D. H., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., et al. (2002). Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron, 33, 341 355.

PAGE 48

48 Folstein, M., Folstein, S., & McHugh, P. (1975). Mini mental state: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatry Research, 12, 189 198. Hagmann, P. (2005). From diffusion MRI to brain connectomics (PhD Thesis). Lausanne: Ecole Polytechnique. 127. Hagmann, P., Cammoun, L., Gigandet, X., Meuli, R., Honey, C. J., Wedeen, V. J., et al. (2008). Mapping the structural core of human cerebral cortex. PLoS Biol, 6 (7), e159. Henry, J. D., & Beatty, W. W. (2006). Verbal fluency deficits in multiple sclerosis. Neuropsychologia, 44(7), 11661174. Holodny, A. I., Watts, R., Korneinko, V. N., Pronin, I. N., Zhukovskiy, M. E., Gor, D. M., et al. (2005). Diffusion tensor tractography of the motor white m atter tracts in man: Current controversies and future directions. Ann N Y Acad Sci, 1064, 88 97. Huang, H., Zhang, J., van Zijl, P. C., & Mori, S. (2004). Analysis of noise effects on DTI based tractography using the bruteforce and multi ROI approach. Mag n Reson Med, 52(3), 559 565. Jian, B., Vemuri, B. C., Ozarslan, E., Carney, P. R., & Mareci, T. H. (2007). A novel tensor distribution model for the diffusionweighted MR signal. Neuroimage, 37(1), 164 176. JohansenBerg, H. & Behrens, T. E. J. (2009). Dif fusion MRI. London, UK: Academic Press. Levin, B. E., Tomer, R., & Rey, G. J. (1992). Cognitive impairments in Parkinson's disease. Neurol Clin, 10 (2), 471 485. Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsychological Assessment. New York, NY: Oxford University Press. Libon, D. J., Price, C., Garrett, K. D., & Giovannetti, T. (2004). From Binswanger's disease to Leukoaraiosis: What we have learned about subcortical vascular dementia. The Clinical Neuropsychologist 18(1), 83100. Lichter, D .G. (2000). Movement disorders and frontal subcortical circuits. In: Lichter DJ, Cummings JL eds. Frontal subcortical Circuits in Psychiatric and Neurological Disorders. New York, NY: The Guilford Press, 260313. Mesulam, M. M. (2000). Behavioral neuroanat omy, Large scale networks, association cortex, frontal syndromes, the limbic system and hemispheric specializations. In: Mesulam, MM, editor. Principles of behavioral and cognitive neurology. New York, NY: Oxford Univ.

PAGE 49

49 Miller, A. K., Alston, R. L., & Corse llis, J. A. (1980). Variation with age in the volumes of grey and white matter in the cerebral hemispheres of man: measurements with an image analyser. Neuropathol Appl Neurobiol, 6(2), 119132. Mori, S., Crain, B. J., Chacko, V. P., & van Zijl, P. C. (199 9). Threedimensional tracking of axonal projections in the brain by magnetic resonance imaging. Ann Neurol, 45(2), 265 269. Moseley, M. E., Cohen, Y., Kucharczyk, J., Mintorovitch, J., Asgari, H. S., Wendland, M. F., et al. (1990). Diffusion weighted MR i maging of anisotropic water diffusion in cat central nervous system. Radiology, 176(2), 439 445. Pahwa, R., Paolo, A., Troster, A., & Koller, W. (1998). Cognitive impairment in Parkinson's disease. Eur J Neurol, 5(5), 431 441. Price, C. C., Wood, M. F., Le onard, C. M., Towler, S., Ward, J., Montijo, H., et al. (2010). Entorhinal cortex volume in older adults: reliability and validity considerations for three published measurement protocols. J Int Neuropsychol Soc, 16 (5), 846 855. Ranta, M. E., Crocetti, D., Clauss, J. A., Kraut, M. A., Mostofsky, S. H., Kaufmann, W. E. (2009). Manual MRI parcellation of the frontal lobe. Psychiatry Res, 172(2) 14754. Rosenthal, E., Brennan, L., Xie, S., et al. (2010). Association between cognition and function in patients with Parkinson Disease with and without dementia. Movement Disorder Society, 25(9), 11701176. Schmahmann, J. D., & Pandya, D. N. (2006). Fiber Pathways of the Brain. New York: Oxford University Press. Spreen, O. & Strauss, E. (1998). A compendium of neuro psychological tests: Ad ministration, norms, and comment a ry New York: Oxford University Press. Stejskal E. O. & Tanner J. E. (1965). Spin diffusion measurements: spin echoes in the presence of a timedependent field gradient. J Chem Phys, 42, 288 292. Teki n, S., & Cummings, J. L. (2002). Frontal subcortical neuronal circuits and clinical neuropsychiatry: an update. J Psychosom Res, 53 (2), 647 654. Tisserand, D. J., Pruessner, J. C., Sanz Arigita, E. J., van Boxtel, M. P. J., Evans, A. C., Jolles, J., et al. (2002). Regional frontal cortical volumes decrease differentially in aging: an MRI study to compare volumetric approaches and voxel based morphometry. Neuroimage, 17(2) 657 659. Tuch, D. S., Reese, T. G., Wiegell, M. R., Makris, N., Belliveau, J. W., & W edeen, V. J. (2002). High angular resolution diffusion imaging reveals intravoxel white matter fiber heterogeneity. Magn Reson Med, 48(4), 577582.

PAGE 50

50 von dem Hagen, E. A., & Henkelman, R. M. (2002). Orientational diffusion reflects fiber structure within a v oxel. Magn Reson Med, 48 (3), 454459. Wakana, S., Caprihan, A., Panzenboeck, M. M., Fallon, J. H., Perry, M., Gollub, R. L., et al. (2007). Reproducibility of quantitative tractography methods applied to cerebral white matter. Neuroimage, 36(3), 630 644. W alters, R. J. C., Fox, N. C., Crum, W. R. et al. (2001). Hemodialysis and cerebral edema. Nephron, 87, 143 147. Yogarajah, M., Focke, N. K., Bonelli, S. B., Thompson, P., Vollmar, C., McEvoy, A. W., et al. (2010). The structural plasticity of white matter networks following anterior temporal lobe resection. Brain, 133 (8), 2348 2364. Yushkevich, P. A., Piven, J., Hazlett, H. C., Smith, R. G., Ho, S., Gee, J. C., et al. (2006). User guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability. Neuroimage, 31(3), 11161128. Zgaljardic, D. J., Borod, J. C., Foldi, N. S., & Mattis, P. J. (2003). A Review of the cognitive and behavioral sequelae of Parkinson's Disease: Relationship to Frontostriatal Circuitry. Cognitive and Behavioral Neurology, 16(4) Zgaljardic, D. J., Borod, J. C., Foldi, N. S., Mattis, P. J., Gordon, M. F., Feigin, A., et al. (2006). An examination of executive dysfunction associated with frontostriatal circuitry in Parkinson's disease. J C lin Exp Neuropsychol, 28(7), 1127 1144.

PAGE 51

51 BIOGRAPHICAL SKETCH Peter T. Nguyen was born and raised in San Jose, California and completed his bachelors degree in psychology at the University of California at Los Angeles. After graduating, he spent two year s working as a full time research assistant at the UCLA Semel Institute and held a volunteer research position at the UCLA Alzheimers Disease Research Center. During his time at UCLA, he also worked as a research assistant at the Brain Injury Research Center and Dr. Greens schizophrenia lab. He subsequently completed his Master of Science degree in clinical psychology with an emphasis on neuropsychology at the University of Florida Department of Clinical and Health Psychology. His current research interes ts include examining brainbehavior relationships using structural magnetic resonance imaging and diffusion weighted imaging.