{"http:\/\/dx.doi.org\/10.14288\/1.0085643":{"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool":[{"value":"Science, Faculty of","type":"literal","lang":"en"},{"value":"Physics and Astronomy, Department of","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/dataProvider":[{"value":"DSpace","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeCampus":[{"value":"UBCV","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/creator":[{"value":"Vavasour, Irene Margaret","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/issued":[{"value":"2009-06-02T19:33:29Z","type":"literal","lang":"en"},{"value":"1998","type":"literal","lang":"en"}],"http:\/\/vivoweb.org\/ontology\/core#relatedDegree":[{"value":"Doctor of Philosophy - PhD","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeGrantor":[{"value":"University of British Columbia","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/description":[{"value":"Magnetic resonance imaging (MRI) has become an invaluable tool for studying brain\r\nand its associated pathologies. Multiple sclerosis (MS) is one such pathology and attempts\r\nare being made to use MRI to characterise the myelination state of MS lesions.\r\nTwo techniques have been proposed which appear to be sensitive to myelination:\r\nmagnetization transfer (MT) and T\u2082 relaxation. Quantification of these techniques uses\r\nmagnetization transfer ratios (MTR) for MT and myelin water percentages for T\u2082 relaxation.\r\nIf the two techniques are both related to myelin content then they are expected to\r\nbe related to each other. It was found by in vivo MRI measurements that white matter\r\nfrom normal volunteers and normal appearing white matter from MS patients had significantly\r\nlarger MTRs and myelin water percentages than grey matter. However, only a\r\nweak correlation was found between MTRs and myelin water percentages in MS lesions\r\n(R=0.5,P=0.005) indicating that each technique provides an independent measure of M S\r\npathology.\r\nSince water in white matter resides in two main compartments, in intra\/extracellular\r\nspaces and between myelin bilayers, it was thought that MT would have a different effect\r\non each water pool. This was examined by combining a T\u2082 relaxation sequence, which\r\nseparates the two water pools, with an MT pulse. It was found using in vivo MRI measurements\r\non normal human white matter that the myelin water pool was significantly\r\nmore affected by an MT pulse than the intra\/extracellular water pool (P=0.00001 to\r\np=0.04 for different white matter structures). It was also found that small offset frequencies\r\ncaused more direct saturation of the myelin water pool than the intra\/extracellular pool resulting in different contrast. Finally, at long delay times between the MT pulse\r\nand the initiation of the T\u2082 relaxation sequence (>500 ms), the difference in MT between\r\nthe two pools was eliminated indicating exchange within that timescale.\r\nIn vitro experiments on bovine brain were performed on a \u00b9H - NMR spectrometer. A\r\n4-pool model was proposed to explain the different relaxation times measured in bovine\r\nwhite matter. These pools included intra\/extracellular water, myelin water, non-myelin\r\nmolecules and myelin molecules. Exchange between the myelin water and myelin, and\r\nthe intra\/extracellular water and non-myelin molecules were rapid with the former being\r\nslightly faster than the latter. There was no evidence for exchange between the two water\r\npools within the timescale of 1 s.\r\nFor human brain, a diffusion model was proposed to investigate exchange between\r\nthe water pools. Results showed that variations in parameters associated with the intra\/\r\nextracellular water pool affected only that pool. Variations in the myelin water pool,\r\nhowever, influenced the relaxation times and amplitudes of both water pools. Finally, it\r\nwas found that changes in the axonal diameter and myelin thickness resulted in changes\r\nin the myelin water percentages and T\u2082 relaxation times. This could account for some\r\nof the differences in myelin water percentages and T\u2082 times measured in different white\r\nmatter structures in the human brain.","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/aggregatedCHO":[{"value":"https:\/\/circle.library.ubc.ca\/rest\/handle\/2429\/8600?expand=metadata","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/extent":[{"value":"7510032 bytes","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/elements\/1.1\/format":[{"value":"application\/pdf","type":"literal","lang":"en"}],"http:\/\/www.w3.org\/2009\/08\/skos-reference\/skos.html#note":[{"value":"M A G N E T I C R E S O N A N C E O F H U M A N A N D B O V I N E B R A I N Irene Margaret Vavasour B . Sc. (Physics) University of New Brunswick, 1991 M . Sc. (Physics) University of Br i t i sh Columbia, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E D E G R E E OF D O C T O R OF PHILOSOPHY in T H E FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHYSICS We accept this thesis as conforming to the required standard T H E UNIVERSITY *6F BRITISH COLUMBIA February 1998 @ Irene Margaret Vavasour in presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2\/88) Abstract Magnetic resonance imaging (MRI) has become an invaluable tool for studying brain and its associated pathologies. Mul t ip le sclerosis (MS) is one such pathology and at-tempts are being made to use M R I to characterise the myelination state of M S lesions. Two techniques have been proposed which appear to be sensitive to myelination: magnetization transfer ( M T ) and T 2 relaxation. Quantification of these techniques uses magnetization transfer ratios ( M T R ) for M T and myelin water percentages for T 2 relax-ation. If the two techniques are both related to myelin content then they are expected to be related to each other. It was found by in vivo M R I measurements that white matter from normal volunteers and normal appearing white matter from M S patients had sig-nificantly larger M T R s and myelin water percentages than grey matter. However, only a weak correlation was found between M T R s and myelin water percentages in M S lesions (R=0.5,P=0.005) indicating that each technique provides an independent measure of M S pathology. Since water in white matter resides in two main compartments, in intra\/extracellular spaces and between myelin bilayers, it was thought that M T would have a different effect on each water pool. This was examined by combining a T 2 relaxation sequence, which separates the two water pools, wi th an M T pulse. It was found using in vivo M R I mea-surements on normal human white matter that the myelin water pool was significantly more affected by an M T pulse than the intra\/extracellular water pool (P=0.00001 to p=0.04 for different white matter structures). It was also found that small offset frequen-cies caused more direct saturation of the myelin water pool than the intra\/extracellular i i pool resulting in different contrast. Finally, at long delay times between the M T pulse and the ini t iat ion of the T 2 relaxation sequence (>500 ms), the difference in M T between the two pools was eliminated indicating exchange within that timescale. In vitro experiments on bovine brain were performed on a 1 H - N M R spectrometer. A 4-pool model was proposed to explain the different relaxation times measured in bovine white matter. These pools included intra\/extracellular water, myelin water, non-myelin molecules and myelin molecules. Exchange between the myelin water and myelin, and the intra\/extracellular water and non-myelin molecules were rapid wi th the former being slightly faster than the latter. There was no evidence for exchange between the two water pools wi th in the timescale of 1 s. For human brain, a diffusion model was proposed to investigate exchange between the water pools. Results showed that variations in parameters associated wi th the in-tra\/extracellular water pool affected only that pool. Variations in the myelin water pool, however, influenced the relaxation times and amplitudes of both water pools. Finally, it was found that changes in the axonal diameter and myelin thickness resulted in changes in the myelin water percentages and T 2 relaxation times. This could account for some of the differences in myelin water percentages and T 2 times measured in different white matter structures in the human brain. i i i Table of Contents Abstract ii List of Tables ix List of Figures x Acknowledgements xii 1 Introduction 1 1.1 Bra in and Mye l in 1 1.2 Mul t ip le Sclerosis : 5 1.3 Mot ivat ion 7 1.4 Review of N M R and M R I work on Bra in 9 1.4.1 T 2 Relaxation 9 1.4.2 T x Relaxation . 11 1.4.3 Magnetization Transfer 12 1.5 Overview of Thesis 13 2 General Theory 14 2.1 Relaxation 14 2.2 Second Moment 17 2.3 Cross Relaxation and Exchange 18 2.4 Magnetization Transfer 19 iv 3 General Materials and Methods 25 3.1 Samples 25 3.1.1 Bovine brain 25 3.1.2 Human brain 26 3.2 N M R and M R I Equipment 27 3.3 N M R Pulse Sequences and Analyses 27 3.3.1 Free Induction Decay 27 3.3.2 Spin-Spin Relaxation 28 3.3.3 Spin-Lattice Relaxation 28 3.3.4 Cross Relaxation 29 3.3.5 T 1 - T 2 Relaxation Dependence 30 3.3.6 Cross-T2 Relaxation Dependence 31 3.4 M R I Pulse Sequences and Analyses 31 3.4.1 Magnetization Transfer 33 3.4.2 T 2 Relaxation 33 3.4.3 T 2 Relaxation wi th M T 33 3.5 Non-negative Least Squares, (NNLS) Analysis of Relaxation 34 4 Comparison of M T R s and Myel in Water Percentages 37 4.1 Summary 37 4.2 Introduction . . . 37 4.3 Materials and Methods 38 4.4 Results . .' 39 4.4.1 Normal Volunteers 39 4.4.2 M S Patients 41 4.4.3 Normal Volunteers vs M S Patients 44 v 4.4.4 Lesions 44 4.5 Discussion 46 4.5.1 M T R Values 46 4.5.2 Comparison of Our M T R s and Mye l in Water Percentages wi th Other Studies 46 4.5.3 Other Comparisons of M T and T 2 Results 48 4.6 Concluding Remarks 48 5 M T Effects on the Short and Long T 2 Relaxation Components of Bra in 50 5.1 Summary 50 5.2 Introduction 50 5.3 Mater ia l and Methods . . . 51 5.3.1 Reproducibil i ty 52 5.3.2 Binomia l M T pulse 53 5.3.3 Sine M T pulse 53 5.3.4 M T Frequency Offset 53 5.3.5 Delay Between M T Pulse and 32 Echo Sequence 53 5.4 Results 54 5.4.1 Reproducibil i ty 54 5.4.2 Binomia l M T pulse 57 5.4.3 Sine M T pulse 58 5.4.4 M T Frequency Offset 58 5.4.5 Delay Between M T Pulse and 32 Echo Sequence 58 5.5 Discussion 63 5.6 Concluding Remarks 66 v i 6 Relaxation Measurements of Bovine Brain using Magnetic Resonance 68 6.1 Summary 68 6.2 Introduction 69 6.3 Mater ia l and Methods 70 6.3.1 Samples 70 6.3.2 N M R Experiments . 70 6.4 Results 70 6.4.1 Free Induction Decay 70 6.4.2 C P M G \u2022 \u2022 \u2022 \u2022 7 3 6.4.3 T i Relaxation 77 6.4.4 Cross Relaxation 77 6.4.5 T i - T 2 Dependence 84 6.4.6 Cross-T 2 Dependence 84 6.5 Discussion 88 6.6 Conclusion 91 7 Diffusion model of T 2 and T i relaxation in two brain water pools 92 7.1 Summary 92 7.2 Introduction 92 7.3 Numerical Methods 93 7.4 Numerical Applications 96 7.5 T 2 simulations 98 7.6 T i simulations 102 7.7 Discussion 102 7.8 Concluding Remarks 105 v i i 8 Conclusions 106 8.1 T h i s w o r k 106 8.2 O n - g o i n g a n d future work 108 Bibliography 109 Appendix A Source Code for T 2 Simulations 121 v i i i List of Tables 3.1 Composit ion of bovine brain samples 26 4.1 Comparison of myelin water percentages and M T R s for M S lesions . . . . 45 5.1 Ampl i tude of T 2 components for smooth and 4 T 2 model distributions . . 54 5.2 M T R s from binomial and sine pulse 57 5.3 M T R s using different off-resonance offsets 59 5.4 M T R s using different M T delays 61 6.1 M y e l i n water percentages for bovine white and grey matter 75 6.2 Ts and T ^ for white and grey matter 81 6.3 Cross relaxation times for white and grey matter 87 7.1 A x o n diameter and myelin thickness versus myelin water percentage . . . 99 ix List of Figures 1.1 Schematic of a nerve cell 2 1.2 Electron micrograph of myelin 3 1.3 A c t i o n potential 5 1.4 Different water components from T 2 relaxation 10 2.1 Relaxation times and amplitudes as a function of cross relaxation time . 20 2.2 Lineshape for two proton pools in brain 21 2.3 Power spectrum of different M T pulses 23 3.1 Image wi th R O I regions shown 32 3.2 Typ ica l C P M G decay curve and T 2 distribution 35 4.1 Images from a normal volunteer 40 4.2 M T R s versus myelin water percentages in white and grey matter 42 4.3 Duncan's multiple range test for myelin water percentages 43 4.4 Images from an M S patient . 43 4.5 M T R vs myelin water percentage in M S lesions 45 5.1 Difference between smooth and 4 T 2 model 55 5.2 Typ ica l no M T and M T T 2 distributions 56 5.3 Average M T R as a function of off-resonance offset 60 5.4 Average M T R as a function of delay 62 6.1 Free induction decays for white and grey samples 71 6.2 Second moment as a function of moisture content 72 x 6.3 T 2 distributions for white and grey samples 73 6.4 T 2 distributions at different r values 74 6.5 T 2 distributions at different moisture contents 76 6.6 T i distributions for white and grey samples 78 6.7 T i distributions at different moisture contents \". 79 6.8 Saturation recovery curves for white and grey samples . . 80 6.9 Solid signal recovery from cross relaxation 82 6.10 Model of proton pools in bovine brain 83 6.11 Plot of T i - T 2 dependence 85 6.12 Cross relaxation dependence of T 2 components 86 7.1 Cyl indr ica l model of water pools in human brain 94 7.2 Effect of cell radii and myelin thickness on T 2 and T i times and amplitudes 98 7.3 Effect of in i t ia l T 2 and T i on T 2 and T i times and amplitudes 100 7.4 Effect of diffusion coefficient on T 2 and T i times and amplitudes 101 x i A c k n o w l e d g e m e n t s I would first like to thank my supervisor Alex since none of this would be possible without h im. He was a wonderful supervisor and also a great role model since he's shown me that although research is fun, there are other things in life that are just as rewarding (not to mention that he's a really nice guy). I would also like to thank my committee of Myer , San and David for all the helpful suggestions and discussions. In conjunction wi th David , a thank you must go out to the technologists and the Wendy Morrison at the M S clinic who were instrumental in getting M S patients as well as scanning both patients and normal volunteers. Next would be K e n who has been invaluable in everything from programming to analysis to life. (He's also a Babylon 5 fan which puts h im in my good books.) There are al l the people from room 100 (past and present) who have made the past 4.5 years unbelievably great. In particular, I've enjoyed pottery wi th Elana and Cornelia, talks wi th Elana, Cornelia, Jamie and Traci, Frank is a never ending source of information and always wil l ing to lend a hand (how do you think my thesis got printed?), Jenifer (although she has gone on to bigger and better things) is always ready wi th great conversation, Denis (who has also moved up in the world) brought fun and girl guide cookies, Reza and X i n were usually ready for a game of fooseball and Sophia who is wi l l ing to listen to me go on about my research. I also need to thank my parents who have insti l led in me a love of physics (as well as the correct physics genes). I guess the apple never falls far from the tree. Last but c e r t a i n l y not least, I need to thank my husband, Jeff, who, from his proposal at the beginning of my P h D to now has been incredibly supportive and loving. (He was also invaluable help in getting my simulations going.) These past few years have beentruely wonderful. Thanks everyone! x i i Chapter 1 Introduction 1.1 Bra in and Mye l in The brain is made up of neurons and glial cells which work together to carry out brain function. The neurons form the communication network in the body while the glial cells act as a support structure. A sketch of a typical nerve cell is shown in Figure 1.1. The nerve cell is made up of four different structures, the cell body (which is the metabolic centre of the cell), the axon (which transmits nerve impulses also known as action potentials over long distances), the dendrites (which receive signals from other neurons) and the pre-synaptic terminal (which releases neurotransmitter in response to action potential in order to pass on information to the next neuron). Axons can propagate for distances up to 1 m and range from 0.2-20 jim in diameter. The axon connects to the cell body at a region called the action hillock. A t this location, the composition of the neuron is unique in order to allow the init iat ion of action potentials v i a an integration of al l signals received by the cell. In order for signals to propagate quickly along the length of the axon, an insulating structure called the myelin sheath is present. Mye l in is described in more detail below. A t its terminus, the axon divides into small branches each wi th a pre-synaptic terminal. These terminals end in proximity to other dendrites or cell bodies in order to transmit the signal, through release of neurotransmitters, to the next cell. Therefore, signal propagation in a neuron begins wi th neurotransmitters binding to specialised receptors in the dendrites or cell body where these chemical signals 1 Chapter 1. Introduction 2 are transformed into electrical signals. These electrical signals then transmit passively to the axon hillock where they are integrated and, if a certain threshold is reached, an action potential is generated. This action potential flows uni-directionally down the length of the axon unt i l it reaches the pre-synaptic terminals. Modulated by the number and frequency of the action potentials, neurotransmitter is released from the pre-synaptic terminal. The neurotransmitter then diffuses to the next cell and binds to receptors. The whole process begins again. Depending on the type of neuron being fired, an inhibitory or excitatory signal can be delivered to the post-synaptic cell. Figure 1.1: A sketch of a typical nerve cell in the brain. From left to right are the pre-synaptic terminals, the myelinated axon, the cell body and the dendrites. [7] The myelin sheath is a lipid-protein membrane found in the central nervous system (CNS) and peripheral nervous system (PNS) of vertebrates [1, 2]. In the C N S , it is created by specialised glial cells called oligodendrocytes which wind themselves tightly around the axon (Figure 1.2) [3, 4]. The resulting multilayer is composed of repeating units of membrane-cytoplasmic space-membrane-extracellular space wi th a thickness of 150-160 A[5]. Dehydrated myelin is composed of 75-80% l ip id and 20-25% protein [6]. This is quite unusual since other membranes are generally 50% l ip id and 50% protein. > Chapter 1. Introduction 3 Myel in makes up 50% of the dry weight of white matter and has a relatively low water content of 40% [6]. The myelin is present to act as an electrical insulator for neurons and allow conduction of nerve signals to propagate about 100 x faster. There are gaps along the axon where no myelin is present known as the nodes of Ranvier which are important in signal conduction. Figure 1.2: A n electron micrograph of myelin. The bar represents O.lpmx 150000. [1] Since nerve impulses consist of electrical signals, it is important to know how these are created and how they propagate. A nerve cell membrane has a membrane resting potential of -65 m V . (The negative sign indicates that the inside of the cell is negatively Chapter 1. Introduction 4 charged wi th respect to the outside of the cell.) This potential is due to the uneven distr ibution of ions, namely N a + , K + , C I \" and C a 2 + , across the membrane. In signal transduction, the two most important ions are N a + and K + . The N a + concentration is about 10 times lower inside the cell than outside while the K + concentration is about 50 times higher inside than outside [7]. These gradients are maintained by a N a + \/ K + pump and also by the leakiness of the plasma membrane to K + but not N a + . In the membrane, there are voltage gated ion channels which only open when the membrane becomes depolarised. The interplay between these channels creates the action potential which propagates along the axon (shown in Figure 1.3). When the membrane potential is increased by about 10 m V (from -65 to -55), an action potential is fired. The in i -t ia l depolarisation causes voltage gated N a + channels to open and N a + rushes into the cell. This causes a large depolarisation to about +110 m V . The N a + channels have a self-regulating feature which automatically closes them after a short time. A t this time, voltage gated K + channels open and K + rushes out of the cell causing the membrane to become repolarised. This efflux of K + actually overshoots the ini t ia l resting potential and the membrane becomes hyperpolarised. The K + channels close and the N a + \/ K + pump restores the ion gradients. The overall duration of an action potential is about 1 ms. C I \" is often passively distributed across the membrane although there are sometimes pumps which pump the CT~ out of the cell. This ion wi l l also help contribute to the action po-tential but to a lesser extent. A s the membrane becomes depolarised, it dissipates quickly along the axon. This passive conduction of the depolarisation can cause neighbouring regions of the axon to reach the threshold and a new action potential is triggered. If myelin is present, the rate of dissipation of the depolarisation is much slower. A t the nodes of Ranvier, N a + channels are found in large concentrations making the triggering of action potentials very easy. Therefore, the action potential can hop along the axon from node to node which allows it to travel much faster. This propagation of action Chapter 1. Introduction 5 potentials is called saltatory conduction. 0 1 2 3 4 Time (msec) Figure 1.3: A sketch of an action potential. The two solid lines show the flux of N a + and K + through ion channels and the dashed line represents the resulting action potential. The horizontal reference line is at -65 m V and the top of the action potential reaches about +110 m V . [7] 1.2 Mult iple Sclerosis In Canada, the number of people afflicted wi th multiple sclerosis (MS) is one of the highest in the world. It is thought that over 50000 Canadians have the disease. M S is usually diagnosed at an early age, between 20 and 40 years and the ratio of women to men affected is 1.8:1. M S is characterised by the destruction of C N S myelin. The disease is thought to be brought on by both environmental and genetic factors. A virus is thought Chapter 1. Introduction \u20226 to play a role but none has been definitively linked [8, 9]. There are two forms of M S : a chronic form (90% of cases) which manifests as a continuous relapsing\/remitting cycle of demyelination and an acute form (10% of cases) where there is rapid and progressive de-myelination. The chronic form usually consists of relapses and then subsequent recovery of most disabilities. After several years, complete recovery no longer occurs and the dis-ease becomes known as secondary progressive. Demyelination arises from an autoimmune response targeting myelin antigens which are thought to derive from myelin basic protein or some other myelin specific protein [1]. These areas of demyelination are referred to as lesions. Ear ly lesions have been correlated wi th the breakdown of the blood brain barrier near small vessels. The lesions are then believed to become inflamed followed by demyelination, gliosis and finally axonal loss. Lesions can be visualised by computerised tomography ( C T ) and magnetic resonance imaging (MRI) [10, 11, 12] although the im-ages are generally non-specific to the state of the lesion. M S lesions become infiltrated by immune cells such as macrophages, neutrophils and lymphocytes. The macrophages appear to destroy myelin sheaths by digesting the outer layers unt i l al l myelin is gone. Strangely, this cellular attack wi l l stop at a boundary and the white matter beyond looks relatively normal. There is no known cause for this behaviour. Normal ly the demyeli-nated tissue would be repaired but the disease also affects the body's abili ty to replenish the myelin. Oligodendrocytes do attempt to remyelinate some neurons but they are both slow and inefficient. Al though lesions are the main locations for disease, there is evidence that normal appearing white matter may also have small abnormalities which are difficult to see wi th M R I (hence the term normal appearing). These abnormalities may add to the overall lesion load and contribute to disabilities acquired by the patient. Chapter 1. Introduction 7 1.3 Motivat ion M y e l i n is found wrapped around axons in brain white matter and helps speed up nerve signal conduction. Unfortunately in certain diseases such as multiple sclerosis (MS) , myelin breaks down. Magnetic resonance imaging (MRI) is a fairly new technique which has proven useful in detecting M S lesions. However, M R I is not capable of detecting the pathological state of the lesion and correlation between lesion load and disability has been small [13, 14, 15]. Therefore efforts are being made to produce an M R I sequence which wi l l be able to probe the state of the lesion and be able to quantify the amount of myelin. Dr . A l e x M a c K a y has developed ah in vivo T 2 relaxation pulse sequence which allows the separation of different water pools in the brain. One of these pools is thought to be associated wi th myelin. Ini t ial experiments on brain [16, 17] showed that white matter had larger quantities of this \"myelin water\" than grey matter. Also, M S lesions were found to have greatly reduced amounts of myelin water compared to the surrounding tissue. These experiments were very exciting since they showed that one component from T 2 relaxation seemed to be related to water associated wi th myelin which may be related to myelin content. Elsewhere, the group of Dousset et al. began applying the technique of magnetization transfer ( M T ) to lesions [18]. Their ini t ia l findings showed that mainly edematous lesions produced a different M T effect than demyelinated lesions. These results indicated that M T might also be able to differentiate between different types of lesions. Neither technique has fully validated its ability to measure myelin. A study was carried out where an MS-l ike disease (experimental allergic encephalomyelitis or E A E ) was induced in primates and M T was subsequently performed [19]. From histological data, lesions wi th more demyelination were indeed found to have a larger change in Chapter 1. Introduction 8 M T than non-demyelinated lesions. T 2 relaxation has been performed on guinea pigs wi th E A E [20] and histological samples showed that the lesions were demyelinated which correlated well wi th their reduced myelin water peak. Therefore, there is evidence that both these techniques are influenced by myelin content. Since both techniques are thought to be related to myelination, we decided to compare them. If they are indeed proportional to the amount of myelin, then a linear correlation between them would be expected. This study was done on normal volunteers and M S patients. It was hoped that we could determine which technique was better at quantifying myelin. From the results of T 2 relaxation, it was evident that there were different water pools in the brain. These water pools have different interactions wi th the macromolecules and therefore each pool should have a different M T effect. The T 2 relaxation pulse sequence was combined wi th the M T sequence in order to monitor the effects of M T on each water pool separately. Since most groups assume that there is only one water pool in the brain, this work would improve modelling in this area and also help in understanding the mechanisms involved wi th M T in the brain. In vitro 1 H - N M R experiments were performed to measure interactions between the different water pools as well as the surrounding macromolecules in order to provide a better understanding of the mechanisms of relaxation present in brain tissue. This in turn could lead to improvements in M R pulse sequences which could provide novel information in the diagnosis of different diseases. Final ly , a model for relaxation was proposed that involved diffusion between the two water pools. Numerical simulations were carried out to try and determine the effect of different parameters (such as diffusion coefficients and cell size) on the relaxation amplitude and time of each pool. Chapter 1. Introduction 9 1.4 Review of N M R and M R I work on Brain 1.4.1 T 2 Relaxation The T 2 relaxation decay curve from brain arises from all the water in the brain [20, 21]. Previous studies [16, 17, 22, 23] have shown that three water compartments may be distinguished on the basis of T 2 time: a long T 2 component assigned to cerebrospinal fluid, an intermediate T 2 component assigned to extracellular water and cytoplasm, and a short T 2 component assigned to myelin water (Figure 1.4). The myelin water is thought to be trapped between myelin bilayers where the short T 2 time derives from interactions between the water and the molecules in the myelin bilayers. The amount of water between myelin bilayers is expected to be proportional to the amount of myelin. In general, we ignore C S F which does not contribute much to the signal from white and grey matter so that there are two water pools in normal C N S tissue and they are resolvable on the basis of their T 2 relaxation. Measurements of T 2 relaxation times in brain are not always done rigorously which leads to conflicting results in the literature [24]. T 2 relaxation pulse sequences must include more than 4 echoes and have echo spacings of 10 ms or less in order to measure both the infra\/extracellular and myelin water. In white matter, the intra\/extracellular water T 2 varies for the different structures (e.g. 86 ms for the internal capsules and 71 ms for the minor forceps) which is likely due to differences in myelin content [17]. In certain cases, the T 2 distribution for white matter shows two peaks in the 80-100 ms region which may be separation of the intracellular and extracellular water pools. The T 2 of infra\/extracellular water in grey matter is also found to be of the order of 70-90 ms. Some groups induced cerebral edema in animals in order to determine the effect on T i and T 2 relaxation [25, 26, 27]. When edema was induced, the longer T 2 component Chapter 1. Introduction 10 Figure 1.4: A T 2 distribution of white matter showing the different water components in the brain. The area under each peak is proportional to the number of protons in that environment. Chapter 1. Introduction 11 split into two peaks representing intracellular and extracellular water. The extracellular water peak was larger and at longer T 2 which was attributed to the extra water present from edema. The intracellular peak remained largely unaffected by the edema. Another experiment on cats studied, the effect of gliosis [28]. T 2 times were found to be largely unchanged even in the presence of edema as well as white matter packed wi th glial fibrils and other structures. This lack of change in T 2 was attributed to the efficient cross relaxation between the water and the extra cytoplasmic structures resulting in no net T 2 change. Finally, in animals induced with E A E , edema in lesions caused the intra\/extracellular water T 2 to increase [29, 30, 31]. In multiple sclerosis, the T 2 of normal appearing white matter was found to be slightly elevated compared to normal controls [32]. In lesions, large variations in T 2 were found which were thought to be due to different underlying tissue composition [33, 34, 35, 36]. In some cases, the large T 2 component had split into two peaks and was thought to represent axonal loss leading to a larger extracellular space [37, 38]. However, other studies have not been able to confirm this result [39, 40]. 1.4.2 T i Relaxation Unlike T 2 relaxation which is able to differentiate different water pools, T i relaxation of human brain appears to yield only one relaxation time. In a myelinated crayfish nerve, the T i was found to be about 1.2 s [22] which is slightly less than that expected for pure water. In the squid giant axon, a similar T i of 1.5 s was found [41]. In human brain however, the T i of white matter was about 600 ms whereas the T i of grey matter was 1 s [42]. Again , only mono-exponential relaxation was found. The difference in T i between white and grey matter is thought to be due to myelin and in particular interactions between myelin water and myelin molecules [21, 43, 44]. Supporting this theory are experiments measuring T i from newborns and adolescents which showed T i times of 1.6 Chapter 1. Introduction 12 s and 500 ms respectively for white matter and 1.6s and 800 ms for grey matter [45]. T i relaxation times were linearly correlated wi th water content in brain tumour sam-ples [46]. W i t h edema, the T i of white matter was found to increase more rapidly than grey matter. The suggested reason i s that excess water in grey matter i s taken up by cells where the T i is reduced by the proteins present in the cytoplasm. In white matter, the excess water remains in the extracellular spaces which is free of such molecules. In animal models of edema, T i was also increased [25, 26, 27, 29, 30, 31]. In gliosis, T i was increased as opposed to T 2 which d id not change resulting in a decrease of T 2 relative to T i [28]. In M S patients, T i increased in lesions but was sti l l mono-exponential and therefore probably not as useful as multi-exponential T 2 measurements [34, 35, 36]. There were however, large variations in T i between lesions attributed to different underlying tissue structure. N A W M was also found to have an increased T i [32]. 1.4 .3 M a g n e t i z a t i o n T r a n s f e r Magnetization transfer ( M T ) is a relatively new M R technique which provides a novel form of contrast. It was developed by Wolff and Balaban [47] and based on a technique discovered by Forsen and Hoffman [48]. M T has become widely used to study many diseases, in particular multiple sclerosis [16, 18, 47, 49, 50, 51, 52, 53, 54, 55, 56]. M T [47] utilises the fact that there is continuous magnetization exchange between two proton pools in the brain: the motionally restricted pool which arises from non-aqueous tissue and the mobile pool from water [21]. M R I can only directly detect signal from the mobile pool. If the magnetization from the motionally restricted pool is disturbed by an M T pulse, then the effect of exchange can be seen on the mobile pool as a decrease in signal [48]. This effect is called magnetization transfer and is quantified through a magnetization transfer ratio ( M T R ) (further defined in Chapter 2). It is expected that Chapter 1. Introduction 13 brain volumes which have a larger number of motionally restricted protons wi l l show a greater signal decrease upon application of an M T pulse and therefore a greater M T R . Protons associated wi th myelin would be part of the motionally restricted pool. In this way, the state of myelination of a lesion can be probed using M T . 1.5 Overview of Thesis First , the groundwork wi l l be set by reviewing some general theory in Chapter 2. This w i l l be followed by the general materials and methods used in al l experiments (Chapter 3). Further details on the materials and methods is given in the individual results chapters. Three different studies are presented i n the next three chapters. Chapter 4 compares M T R s and myelin water percentages in normal volunteers and M S patients. Chapter 5 extends the work by incorporating both M T and T 2 relaxation into one sequence and studying the effect on normal volunteers. In Chapter 6, in vitro ^ - N M R studies are presented from bovine brain. M a n y different \u2022 experiments were carried out in order to determine how different proton pools interacted and over what timescale. Simulations of T 2 and T i relaxation are presented in Chapter 7 for human brain by assuming that diffusion occurs between the different water pools. F ina l ly in Chapter 8, conclusions are made about how al l the studies relate to each other. Also , future experiments are suggested as well as some that are already underway. Chapter 2 General Theory 2.1 Relaxation 1 H nuclei are spin 1\/2 particles and possess magnetic moment y and angular momentum J . These two quantities are related by A i = 7 J (2.1) where 7 is known as the gyromagnetic ratio. The sum of all the magnetic moments in a sample gives the total magnetization denoted by M . If a magnetic field, B , is applied to the system, a torque results such that j = 7 ( M x B ) . (2.2) This equation produces precession of the magnetization around the external magnetic field at a rate known as the Larmor frequency given by UJ0 = 7 5 o ; (2.3) If a radio frequency (rf) pulse, B\\, is applied at this frequency, the energy will be absorbed and the net magnetization will tip by an angle 9 = 7-Bi\u00a3p where tp is the length of time that the rf field is applied. When the magnetization is perturbed from its equilibrium state, it tries to return via a process called relaxation.. One form of relaxation, denoted T i , brings the magnetization back to equilibrium along the direction of the external magnetic field (normally assigned 14 Chapter 2. General Theory- lb to the z-axis). The other, denoted T 2 , destroys any net magnetization in a plane perpen-dicular to the external field (x-y plane). These terms are added to equation 2.2 in order to obtain the Bloch equation 7 = 7 ( M < B ) t ^ - - - | (2.4) where M0 is the net magnetization at equilibrium. For spin 1\/2 particles, the most important relaxation mechanism is the dipole-dipole interaction. The dipolar Hamiltonian is given by #12 = 2 ^ [ I i \u2022 h - 3(Ix \u2022 n)(I 2 \u2022 h)] (2.5) where I is the angular momentum, r is the separation between the two spins and n is the unit vector in the direction joining them. If we transform n into polar coordinates with angle 6 and
\u00a3 = - sin2 9e~2i* F = ~(Jf J2-) sin2 9e2i