**4. Functional imaging**

While structural imaging is important to assess brain atrophy, the hope is that AD patholo‐ gy will be identified before neuronal death so that atrophy can be prevented. One current theory is that one of the major components leading to amyloid and tau pathologies could be vascular changes [42]. Two of the risk factors for AD are mutated forms of APP, and the ApoE ε4 isoform and both of these factors are involved in cholesterol processing. The inabil‐ ity of a neuron to clear amyloid plaques may be prognostic and indicate impaired blood flow as a risk factor for AD. While it is not immediately apparent how blood flow is contri‐ buting to AD, some vascular changes are being evaluated through the use of hemodynamicbased functional imaging techniques.

#### **4.1. BOLD fMRI**

Functional magnetic resonance imaging, or fMRI is a way to gain insight into the functional processes occurring in the brain. Most fMRI modalities are based on the blood oxygenation level-dependent (BOLD) effect. This is an indirect method of tracking the activation or inac‐ tivation of brain regions relative to a baseline state, and is based on the idea that an active area will need more energy and consume more glucose and oxygen and therefore more blood will need to be directed to that area. More specifically, oxygenated and deoxygenated blood water have different intrinsic magnetic properties (oxygenated blood is diamagnetic and deoxygenated blood is paramagnetic) and therefore affect the T2 and T2\* relaxation times of surrounding water in blood and tissue in different ways. Deoxygenated blood has a strong enough magnetic affect (paramagnetic) that it will distort the local field and decrease the signal intensity (i.e. shorten T2) of surrounding water for that region. Oxygenated blood will not have the same effect, and therefore regions containing more oxygenated blood will have higher signal intensity (longer T2). Importantly, during functional activation the cere‐ bral blood flow increases by a large amount (20-100%) relative to the cerebral metabolic rate of oxygen consumption (CMRO2), resulting in a relative decrease in the concentration of de‐ oxyhemoglobin in capillaries and veins. By comparing the signal intensities of regions at baseline (Figure 4a) and during a task (Figure 4b), the regions that have an increase in capil‐ lary and venous oxygenation can be visualized.

*Evoked BOLD fMRI*

There are several established testing paradigms that have been designed to study memo‐ ry. The most commonly used paradigms look specifically at either episodic or semantic memory. Episodic memories involve the recognition of autobiographical or cued informa‐ tion (e.g., faces, words, other visual stimuli) while semantic memory involves the recog‐ nition of a fact or information regardless of personal context (e.g., famous faces, geographical locations). Because episodic memory is highly affected by AD, many fMRI paradigms use an episodic memory task to elicit functional differences between patients and controls. While encoding a new memory, activation of the hippocampal and para‐ hippocampal regions is decreased in mild AD patients compared to healthy controls [43]. In a block design face name paradigm, AD patients also show decreased hippocampal

Using Magnetic Resonance Imaging in the Early Detection of Alzheimer's Disease

http://dx.doi.org/10.5772/54445

235

A multitude of studies have been performed in asymptomatic carriers of ApoE ε4 with mixed results. In an extensive review of the literature by Trachtenberg et al, some claim that carriers have increased activation across brain structures while others claim the oppo‐ site[44]. Moreover, there have been reports of both increases and decreases of activation or that there is no significant effect at all of carrying the ApoE ε4 allele[44]. In each case, inves‐ tigators have provided hypotheses to explain both increased and decreased activation in ApoE ε4 carriers: decreased activation can be easily explained by the fact that presympto‐ matic carriers are already accumulating AD pathology hallmarks before cognitive decline is experienced. These pathologies may be hindering the BOLD response in the specific areas that experience a decrease in activation, or they may be inhibiting areas that lie functionally upstream. In contrast, an increase in activation can be explained in two ways, which take into account AD pathology. For one, the accumulation of pathology may lead to the dedif‐ ferentiation of neural network such that many networks become involved in a specific proc‐ ess. This may in fact be a part of healthy aging [45] and could be found in young, presymptomatic carriers of ApoE ε4 because their brains are aging more rapidly. Alterna‐ tively, the brain may have a cognitive reserve that needs to "work harder" during a difficult task to perform at a normal level, and thus would have a higher amount of activation. Trachtenberg et al [44] argue that the populations tested in these studies are very young (20s and 30s) and have a great deal of time before they will begin to experience cognitive decline. He suggests instead that the possession of an ApoE ε4 allele leads to a fundamental differ‐

A growing body of evidence suggests that an episodic memory task may not be the best way to characterize memory loss because episodic memory declines as a part of healthy aging as well. Episodic memory tasks are also more difficult than semantic memory tasks, The work may therefore be experiencing a basement effect[45]. Semantic memory, in contrast, is affect‐ ed very early in AD, but remains relatively untouched in the healthy aging process [45]. Most semantic memory tests involve the recognition of famous names and faces [45-47] or categorizing word lists [1,3]. These types of studies have shown an increase in activation

and a decrease in deactivation the MTL regions of carriers of the ApoE ε4 allele.

activation to novel stimuli compared to repeated comparisons [43].

ence in neurophysiology that could be contributing to this effect.

*Spontaneous BOLD fMRI*

**Figure 4.** Blood flow at rest (a) and during activation (b)

BOLD imaging involves very fast sequences in order to visualize changes in functional acti‐ vation on the timescale of the hemodynamic response. This rapid sequencing allows for a time resolution of approximately 2s. Total time required to perform a BOLD scan varies with the task being performed, but typically scans take 5-15 minutes.

There are two main types of fMRI: evoked (task-related) and spontaneous ("resting" state). Evoked fMRI is the more commonly performed test in which the same task is repeated many times with a baseline measurement taken between trials. Statistical tests (Z- and ttests) are used to differentiate the regions activated during the task from those active at baseline. By contrast, spontaneous BOLD specifically measures synchrony of baseline signal fluctuations to determine how the brain is functionally connected.

#### *Evoked BOLD fMRI*

bral blood flow increases by a large amount (20-100%) relative to the cerebral metabolic rate of oxygen consumption (CMRO2), resulting in a relative decrease in the concentration of de‐ oxyhemoglobin in capillaries and veins. By comparing the signal intensities of regions at baseline (Figure 4a) and during a task (Figure 4b), the regions that have an increase in capil‐

BOLD imaging involves very fast sequences in order to visualize changes in functional acti‐ vation on the timescale of the hemodynamic response. This rapid sequencing allows for a time resolution of approximately 2s. Total time required to perform a BOLD scan varies

There are two main types of fMRI: evoked (task-related) and spontaneous ("resting" state). Evoked fMRI is the more commonly performed test in which the same task is repeated many times with a baseline measurement taken between trials. Statistical tests (Z- and ttests) are used to differentiate the regions activated during the task from those active at baseline. By contrast, spontaneous BOLD specifically measures synchrony of baseline signal

with the task being performed, but typically scans take 5-15 minutes.

fluctuations to determine how the brain is functionally connected.

lary and venous oxygenation can be visualized.

234 Understanding Alzheimer's Disease

**Figure 4.** Blood flow at rest (a) and during activation (b)

There are several established testing paradigms that have been designed to study memo‐ ry. The most commonly used paradigms look specifically at either episodic or semantic memory. Episodic memories involve the recognition of autobiographical or cued informa‐ tion (e.g., faces, words, other visual stimuli) while semantic memory involves the recog‐ nition of a fact or information regardless of personal context (e.g., famous faces, geographical locations). Because episodic memory is highly affected by AD, many fMRI paradigms use an episodic memory task to elicit functional differences between patients and controls. While encoding a new memory, activation of the hippocampal and para‐ hippocampal regions is decreased in mild AD patients compared to healthy controls [43]. In a block design face name paradigm, AD patients also show decreased hippocampal activation to novel stimuli compared to repeated comparisons [43].

A multitude of studies have been performed in asymptomatic carriers of ApoE ε4 with mixed results. In an extensive review of the literature by Trachtenberg et al, some claim that carriers have increased activation across brain structures while others claim the oppo‐ site[44]. Moreover, there have been reports of both increases and decreases of activation or that there is no significant effect at all of carrying the ApoE ε4 allele[44]. In each case, inves‐ tigators have provided hypotheses to explain both increased and decreased activation in ApoE ε4 carriers: decreased activation can be easily explained by the fact that presympto‐ matic carriers are already accumulating AD pathology hallmarks before cognitive decline is experienced. These pathologies may be hindering the BOLD response in the specific areas that experience a decrease in activation, or they may be inhibiting areas that lie functionally upstream. In contrast, an increase in activation can be explained in two ways, which take into account AD pathology. For one, the accumulation of pathology may lead to the dedif‐ ferentiation of neural network such that many networks become involved in a specific proc‐ ess. This may in fact be a part of healthy aging [45] and could be found in young, presymptomatic carriers of ApoE ε4 because their brains are aging more rapidly. Alterna‐ tively, the brain may have a cognitive reserve that needs to "work harder" during a difficult task to perform at a normal level, and thus would have a higher amount of activation. Trachtenberg et al [44] argue that the populations tested in these studies are very young (20s and 30s) and have a great deal of time before they will begin to experience cognitive decline. He suggests instead that the possession of an ApoE ε4 allele leads to a fundamental differ‐ ence in neurophysiology that could be contributing to this effect.

A growing body of evidence suggests that an episodic memory task may not be the best way to characterize memory loss because episodic memory declines as a part of healthy aging as well. Episodic memory tasks are also more difficult than semantic memory tasks, The work may therefore be experiencing a basement effect[45]. Semantic memory, in contrast, is affect‐ ed very early in AD, but remains relatively untouched in the healthy aging process [45]. Most semantic memory tests involve the recognition of famous names and faces [45-47] or categorizing word lists [1,3]. These types of studies have shown an increase in activation and a decrease in deactivation the MTL regions of carriers of the ApoE ε4 allele. *Spontaneous BOLD fMRI*

Resting state, or functional connectivity MRI (fcMRI) is a task-independent measurement of brain regions that fluctuate in their BOLD signal together, indicating that they are function‐ ally connected. The Default Mode Network (DMN) is a collection of brain regions that seem to activate together while the brain is at rest, and are deactivated while the brain is engaged in a cognitive task. The DMN is composed of MTL and lateral frontal regions, particularly the posterior cingulate complex [4,6,7]. This network is altered in AD and is a potent bio‐ marker for separating patients with AD from healthy controls [8], patients with aMCI from healthy controls [48], and genetically at-risk individuals from healthy controls [4]. *Caveats to BOLD fMRI*

**Figure 5.** Positive and negative BOLD effects are influenced by CBF, CBV, and CMRO2 and it is not possible to distin‐

Using Magnetic Resonance Imaging in the Early Detection of Alzheimer's Disease

http://dx.doi.org/10.5772/54445

237

Cerebral blood flow is a measurement of the rate of tissue perfusion, usually measured by the amount of blood that reaches a tissue per unit time (mL blood per 100 g tissue per mi‐ nute) [52]. CBF has been quantified by Positron Emission Tomography (PET) [53,54] and Single Photon Emission Computed Tomography (SPECT) [10,55,56] but today it can also be quantified noninvasively using a technique called Arterial Spin Labeling (ASL). ASL uses a radiofrequency pulse to label blood water in an area outside of the region of interest, usually in the neck. After 1-2s, the labeled blood water flows into the imaging region and exchanges with tissue water and a tagged image can be obtained [51,57]. This image is compared with an image where the blood water is not labeled, and the difference between the two images provides a map proportional to CBF. As can be seen, ASL is analogous to tracer-based ap‐ proaches such as 15O PET and Gadolinium-MRI, however the tracer is endogenous blood water as opposed to an exogenous contrast agents. Whole-brain ASL scans can be per‐

In AD patients, deficits in CBF have been seen in the temporoparietal cortex, posterior cin‐ gulate cortex, and frontal cortex [57-59]. CBF as measured by ASL has been shown to be in‐ creased in aMCI patients but decreased pre-symptomatic carriers of ApoE ε4 [11,60]. The

Often, changes in blood flow precede structural changes, but reduced CBF is not necessarily an indicator of vascular dysfunction. For instance, CBF alterations may be due to a lower metabolic demand, cardiac output, or blood pressure [10,61,62]. Longitudinal analysis of CBF in at-risk populations should be developed for its potential as a method for tracking

Cerebral blood volume measures the amount of blood per 100 mL brain tissue. It is an indi‐ rect measurement of the vascularization of brain regions, and is less dependent on the sub‐ ject's respiration than CBF[11,15,63,64]. There are currently two major techniques that measure CBV: Dynamic Susceptibility Contrast MRI (DSC-MRI) and Vascular Space Occu‐ pancy MRI (VASO). DSC-MRI involves the injection of gadolinium as a contrast agent, and is the best validated measure. Unfortunately, the injection of gadolinium is dose-restricted because of its toxic effect on kidneys which limits its potential for longitudinal studies and older patient populations [17,65]. VASO is a completely non-invasive method of measuring CBV changes and has been gaining popularity in recent years. Unlike DSC-MRI, VASO uses endogenous blood water as a contrast agent. VASO can be performed by measuring the tis‐ sue signal with and without blood water nulled, and subtracting one image from the other. Although VASO is correlated with DSC-MRI there are some minor variations in the two

measurements, suggesting that the underlying physiology may be different [19,63].

increase that is seen in aMCI has been attributed to compensatory mechanisms [60].

guish which factor is contributing by only measuring BOLD.

formed in less than 5 minutes with a spatial resolution of 3-5 mm.

disease progress or recognizing it before cognitive symptoms begin.

**4.2. Cerebral blood flow**

**4.3. Cerebral blood volume**

Although BOLD fMRI is an important tool for research, there are some limitations to its clin‐ ical feasibility as a biomarker for future AD. To date, it has not successfully been used in predicting patient prognosis or trajectory. In terms of practicality, fMRI is expensive and re‐ quires extensive image processing, which will drive up the cost of any tests. It is also not completely reproducible across testing sites or days. Different equipment and software can create variables in data analysis across testing sites. Longitudinal studies can present diffi‐ culties because as they age, patients may develop comorbidities, or begin taking drugs that will interfere with the BOLD signal in a way unrelated to AD pathology. Even subtle changes can influence the BOLD signal such as recent alcohol [49] or caffeine [50] intake.

The biggest difficulty with BOLD fMRI is that it is generally not quantitative. Changes in blood oxygenation are based on three individual components: Cerebral Blood Flow (CBF), Cerebral Blood Volume (CBV), and the Cerebral Metabolic Rate of Oxygen (CMRO2) [51]. Figure 5 represents the many ways that CBF, CBV, and CMRO2 can contribute to the BOLD effect. It is impossible to determine which of these is contributing to a BOLD fluctuation with fMRI alone. For this reason, vascular imaging techniques are being developed that are able to quantitatively determine the physiological changes that are contributing to the BOLD signal. Techniques to quantify CBF and CBV have been validated and are gaining populari‐ ty. CMRO2 methods are still in development and have not been used in an AD population and will therefore not be covered.


contributing by only measuring BOLD.

recognizing it before cognitive symptoms begin.

**10. Cerebral blood volume** 

may be different [19,63].

**11. Chemical imaging** 

**9. Cerebral blood flow** 

Figure 5. Positive and negative BOLD effects are influenced by CBF, CBV, and CMRO2 and it is not possible to distinguish which factor is

Cerebral blood flow is a measurement of the rate of tissue perfusion, usually measured by the amount of blood that reaches a tissue per unit time (mL blood per 100 g tissue per minute) [52]. CBF has been quantified by Positron Emission Tomography (PET) [53,54] and Single Photon Emission Computed Tomography (SPECT) [10,55,56] but today it can also be quantified noninvasively using a technique called Arterial Spin Labeling (ASL). ASL uses a radiofrequency pulse to label blood water in an area outside of the region of interest, usually in the neck. After 1-2s, the labeled blood water flows into the imaging region and exchanges with tissue water and a tagged image can be obtained [51,57]. This image is compared with an image where the blood water is not labeled, and the difference between the two images provides a map proportional to CBF. As can be seen, ASL is analogous to tracer-based approaches such as 15O PET and Gadolinium-MRI, however the tracer is endogenous blood water as opposed to an exogenous

In AD patients, deficits in CBF have been seen in the temporoparietal cortex, posterior cingulate cortex, and frontal cortex [57-59]. CBF as measured by ASL has been shown to be increased in aMCI patients but decreased pre-symptomatic carriers of ApoE ε4

Often, changes in blood flow precede structural changes, but reduced CBF is not necessarily an indicator of vascular dysfunction. For instance, CBF alterations may be due to a lower metabolic demand, cardiac output, or blood pressure [10,61,62]. Longitudinal analysis of CBF in at-risk populations should be developed for its potential as a method for tracking disease progress or

Cerebral blood volume measures the amount of blood per 100 mL brain tissue. It is an indirect measurement of the vascularization of brain regions, and is less dependent on the subject's respiration than CBF[11,15,63,64]. There are currently two major techniques that measure CBV: Dynamic Susceptibility Contrast MRI (DSC-MRI) and Vascular Space Occupancy MRI (VASO). DSC-MRI involves the injection of gadolinium as a contrast agent, and is the best validated measure. Unfortunately, the injection of gadolinium is dose-restricted because of its toxic effect on kidneys which limits its potential for longitudinal studies and older patient populations [17,65] . VASO is a completely non-invasive method of measuring CBV changes and has been gaining popularity in recent years. Unlike DSC-MRI, VASO uses endogenous blood water as a contrast agent. VASO can be performed by measuring the tissue signal with and without blood water nulled, and subtracting one image from the other. Although VASO is correlated with DSC-MRI there are some minor variations in the two measurements, suggesting that the underlying physiology

VASO has been applied to a mixed group of patients with aMCI and AD and found that there are CBV reductions in the frontal and parietal lobes. These reductions were most striking in white matter which suggests that any vascular component of AD is especially damaging to white matter compared to gray matter [21,22,61]. In the future, longitudinal studies should be performed in

carriers of ApoE ε4 to determine if these white matter vascular deficiencies can be recognized at a young age.

contrast agents. Whole-brain ASL scans can be performed in less than 5 minutes with a spatial resolution of 3-5 mm.

[11,60]. The increase that is seen in aMCI has been attributed to compensatory mechanisms [60].

**Figure 5.** Positive and negative BOLD effects are influenced by CBF, CBV, and CMRO2 and it is not possible to distin‐ guish which factor is contributing by only measuring BOLD.

#### **4.2. Cerebral blood flow**

Resting state, or functional connectivity MRI (fcMRI) is a task-independent measurement of brain regions that fluctuate in their BOLD signal together, indicating that they are function‐ ally connected. The Default Mode Network (DMN) is a collection of brain regions that seem to activate together while the brain is at rest, and are deactivated while the brain is engaged in a cognitive task. The DMN is composed of MTL and lateral frontal regions, particularly the posterior cingulate complex [4,6,7]. This network is altered in AD and is a potent bio‐ marker for separating patients with AD from healthy controls [8], patients with aMCI from

Although BOLD fMRI is an important tool for research, there are some limitations to its clin‐ ical feasibility as a biomarker for future AD. To date, it has not successfully been used in predicting patient prognosis or trajectory. In terms of practicality, fMRI is expensive and re‐ quires extensive image processing, which will drive up the cost of any tests. It is also not completely reproducible across testing sites or days. Different equipment and software can create variables in data analysis across testing sites. Longitudinal studies can present diffi‐ culties because as they age, patients may develop comorbidities, or begin taking drugs that will interfere with the BOLD signal in a way unrelated to AD pathology. Even subtle changes can influence the BOLD signal such as recent alcohol [49] or caffeine [50] intake.

The biggest difficulty with BOLD fMRI is that it is generally not quantitative. Changes in blood oxygenation are based on three individual components: Cerebral Blood Flow (CBF), Cerebral Blood Volume (CBV), and the Cerebral Metabolic Rate of Oxygen (CMRO2) [51]. Figure 5 represents the many ways that CBF, CBV, and CMRO2 can contribute to the BOLD effect. It is impossible to determine which of these is contributing to a BOLD fluctuation with fMRI alone. For this reason, vascular imaging techniques are being developed that are able to quantitatively determine the physiological changes that are contributing to the BOLD signal. Techniques to quantify CBF and CBV have been validated and are gaining populari‐ ty. CMRO2 methods are still in development and have not been used in an AD population

Figure 5. Positive and negative BOLD effects are influenced by CBF, CBV, and CMRO2 and it is not possible to distinguish which factor is

CBF CBV CMRO2

CBF CBV CMRO2 CBF CBV CMRO2

 CBF CBV CMRO2 CBF CBV CMRO2 CBF CBV CMRO2 CBF CBV CMRO2

Cerebral blood flow is a measurement of the rate of tissue perfusion, usually measured by the amount of blood that reaches a tissue per unit time (mL blood per 100 g tissue per minute) [52]. CBF has been quantified by Positron Emission Tomography (PET) [53,54] and Single Photon Emission Computed Tomography (SPECT) [10,55,56] but today it can also be quantified noninvasively using a technique called Arterial Spin Labeling (ASL). ASL uses a radiofrequency pulse to label blood water in an area outside of the region of interest, usually in the neck. After 1-2s, the labeled blood water flows into the imaging region and exchanges with tissue water and a tagged image can be obtained [51,57]. This image is compared with an image where the blood water is not labeled, and the difference between the two images provides a map proportional to CBF. As can be seen, ASL is analogous to tracer-based approaches such as 15O PET and Gadolinium-MRI, however the tracer is endogenous blood water as opposed to an exogenous

In AD patients, deficits in CBF have been seen in the temporoparietal cortex, posterior cingulate cortex, and frontal cortex [57-59]. CBF as measured by ASL has been shown to be increased in aMCI patients but decreased pre-symptomatic carriers of ApoE ε4

Often, changes in blood flow precede structural changes, but reduced CBF is not necessarily an indicator of vascular dysfunction. For instance, CBF alterations may be due to a lower metabolic demand, cardiac output, or blood pressure [10,61,62]. Longitudinal analysis of CBF in at-risk populations should be developed for its potential as a method for tracking disease progress or

Cerebral blood volume measures the amount of blood per 100 mL brain tissue. It is an indirect measurement of the vascularization of brain regions, and is less dependent on the subject's respiration than CBF[11,15,63,64]. There are currently two major techniques that measure CBV: Dynamic Susceptibility Contrast MRI (DSC-MRI) and Vascular Space Occupancy MRI (VASO). DSC-MRI involves the injection of gadolinium as a contrast agent, and is the best validated measure. Unfortunately, the injection of gadolinium is dose-restricted because of its toxic effect on kidneys which limits its potential for longitudinal studies and older patient populations [17,65] . VASO is a completely non-invasive method of measuring CBV changes and has been gaining popularity in recent years. Unlike DSC-MRI, VASO uses endogenous blood water as a contrast agent. VASO can be performed by measuring the tissue signal with and without blood water nulled, and subtracting one image from the other. Although VASO is correlated with DSC-MRI there are some minor variations in the two measurements, suggesting that the underlying physiology

VASO has been applied to a mixed group of patients with aMCI and AD and found that there are CBV reductions in the frontal and parietal lobes. These reductions were most striking in white matter which suggests that any vascular component of AD is especially damaging to white matter compared to gray matter [21,22,61]. In the future, longitudinal studies should be performed in

carriers of ApoE ε4 to determine if these white matter vascular deficiencies can be recognized at a young age.

contrast agents. Whole-brain ASL scans can be performed in less than 5 minutes with a spatial resolution of 3-5 mm.

[11,60]. The increase that is seen in aMCI has been attributed to compensatory mechanisms [60].

healthy controls [48], and genetically at-risk individuals from healthy controls [4].

*Caveats to BOLD fMRI*

236 Understanding Alzheimer's Disease

and will therefore not be covered.

contributing by only measuring BOLD.

recognizing it before cognitive symptoms begin.

**10. Cerebral blood volume** 

may be different [19,63].

**11. Chemical imaging** 

**9. Cerebral blood flow** 

Positive BOLD Effect

Negative BOLD Effect

Cerebral blood flow is a measurement of the rate of tissue perfusion, usually measured by the amount of blood that reaches a tissue per unit time (mL blood per 100 g tissue per mi‐ nute) [52]. CBF has been quantified by Positron Emission Tomography (PET) [53,54] and Single Photon Emission Computed Tomography (SPECT) [10,55,56] but today it can also be quantified noninvasively using a technique called Arterial Spin Labeling (ASL). ASL uses a radiofrequency pulse to label blood water in an area outside of the region of interest, usually in the neck. After 1-2s, the labeled blood water flows into the imaging region and exchanges with tissue water and a tagged image can be obtained [51,57]. This image is compared with an image where the blood water is not labeled, and the difference between the two images provides a map proportional to CBF. As can be seen, ASL is analogous to tracer-based ap‐ proaches such as 15O PET and Gadolinium-MRI, however the tracer is endogenous blood water as opposed to an exogenous contrast agents. Whole-brain ASL scans can be per‐ formed in less than 5 minutes with a spatial resolution of 3-5 mm.

In AD patients, deficits in CBF have been seen in the temporoparietal cortex, posterior cin‐ gulate cortex, and frontal cortex [57-59]. CBF as measured by ASL has been shown to be in‐ creased in aMCI patients but decreased pre-symptomatic carriers of ApoE ε4 [11,60]. The increase that is seen in aMCI has been attributed to compensatory mechanisms [60].

Often, changes in blood flow precede structural changes, but reduced CBF is not necessarily an indicator of vascular dysfunction. For instance, CBF alterations may be due to a lower metabolic demand, cardiac output, or blood pressure [10,61,62]. Longitudinal analysis of CBF in at-risk populations should be developed for its potential as a method for tracking disease progress or recognizing it before cognitive symptoms begin.

#### **4.3. Cerebral blood volume**

Cerebral blood volume measures the amount of blood per 100 mL brain tissue. It is an indi‐ rect measurement of the vascularization of brain regions, and is less dependent on the sub‐ ject's respiration than CBF[11,15,63,64]. There are currently two major techniques that measure CBV: Dynamic Susceptibility Contrast MRI (DSC-MRI) and Vascular Space Occu‐ pancy MRI (VASO). DSC-MRI involves the injection of gadolinium as a contrast agent, and is the best validated measure. Unfortunately, the injection of gadolinium is dose-restricted because of its toxic effect on kidneys which limits its potential for longitudinal studies and older patient populations [17,65]. VASO is a completely non-invasive method of measuring CBV changes and has been gaining popularity in recent years. Unlike DSC-MRI, VASO uses endogenous blood water as a contrast agent. VASO can be performed by measuring the tis‐ sue signal with and without blood water nulled, and subtracting one image from the other. Although VASO is correlated with DSC-MRI there are some minor variations in the two measurements, suggesting that the underlying physiology may be different [19,63].

VASO has been applied to a mixed group of patients with aMCI and AD and found that there are CBV reductions in the frontal and parietal lobes. These reductions were most strik‐ ing in white matter which suggests that any vascular component of AD is especially damag‐ ing to white matter compared to gray matter [21,22,61]. In the future, longitudinal studies should be performed in carriers of ApoE ε4 to determine if these white matter vascular defi‐ ciencies can be recognized at a young age.

#### **5. Chemical imaging**

Structural and functional imaging are important for assessing the damage caused by AD, but for designing therapeutics the ability to view changes at the macromolecular level would be highly beneficial. New techniques are being developed that can do just that. Mag‐ netic Resonance Spectroscopy (MRS) can be done in a single voxel or across multiple voxels (MRS imaging, MRSi) to assess macromolecular concentration. Both are new techniques that are still being optimized, but will be extremely useful in understanding AD.

**Figure 6.** Example of chemical shift spectrum from a normal brain (from the University of Missouri-Kansas City Radiol‐

Using Magnetic Resonance Imaging in the Early Detection of Alzheimer's Disease

http://dx.doi.org/10.5772/54445

239

The most common macromolecules studied in neuroimaging are creatine (Cr) which is usually unaffected by disease and can act as an internal standard, *N-*acetyl-aspartate (NAA), a marker for neuronal health, myo-inositol (mI), a marker of gliosis [32,66], and choline (Cho). In AD, NAA is typically decreased in AD and mI is typically increased so NAA/Cr, mI/Cr and NAA/mI ratios are good markers of the disease with the ratio of NAA to mI being the strongest. The mI concentration has been found to be elevated in aMCI [33,34,67]. There is a trend that NAA is decreased in aMCI, however the effect is much more mild if it exists [35-37,68]. Glutamate (Glu) is the primary excitatory neuro‐ transmitter, and is significantly reduced in AD [69] Gamma-amino butyric acid (GABA) is an inhibitory neurotransmitter, and its concentration may be decreased in AD [70]. It is possible to use MRS to estimate relative concentrations of both Glu and GABA *in vivo*, but due to their low concentrations compared to NAA and other metabolites, and the fact that signal from these metabolites is very close in frequency space to other metabo‐ lites of larger concentration, it is more difficult to identify them without suppressing or "editing" other signals. One of the common methods used to quantify GABA is a PRESS or MEGA-PRESS sequence which suppresses or edits signals from water, creatine, and other nearby metabolites so that the characteristic GABA peaks can be identified. For

more information on the MEGA-PRESS sequence, see Waddell 2007 [71].

The importance of MRS research is clear, but there are some difficulties associated with it. To begin with, the scans take a long time to complete—more than ten minutes in

ogy Resident Resource Webpage

#### **5.1. Magnetic resonance spectroscopy**

MR imaging primarily measures signal from water protons, but in MR spectroscopy protons of various metabolites can be assessed at one time. Quantification is achieved by exciting a single voxel with a combination of RF pulses, and obtaining a free induction decay (FID) spectrum. When this spectrum is Fourier transformed, metabolites can be visualized due to their variability in chemical shift (Figure 6). Because the chemical shift of a single metabolite is constant, it will always peak at the same frequency (measured in parts per million, ppm). By calculating the area under the peak, the concentration of a metabolite relative to an inter‐ nal standard can be obtained.

VASO has been applied to a mixed group of patients with aMCI and AD and found that

there are CBV reductions in the frontal and parietal lobes. These reductions were most strik‐

ing in white matter which suggests that any vascular component of AD is especially damag‐

ing to white matter compared to gray matter [21,22,61]. In the future, longitudinal studies

should be performed in carriers of ApoE ε4 to determine if these white matter vascular defi‐

Structural and functional imaging are important for assessing the damage caused by AD,

but for designing therapeutics the ability to view changes at the macromolecular level

would be highly beneficial. New techniques are being developed that can do just that. Mag‐

netic Resonance Spectroscopy (MRS) can be done in a single voxel or across multiple voxels

(MRS imaging, MRSi) to assess macromolecular concentration. Both are new techniques that

MR imaging primarily measures signal from water protons, but in MR spectroscopy protons

of various metabolites can be assessed at one time. Quantification is achieved by exciting a

single voxel with a combination of RF pulses, and obtaining a free induction decay (FID)

spectrum. When this spectrum is Fourier transformed, metabolites can be visualized due to

their variability in chemical shift (Figure 6). Because the chemical shift of a single metabolite

is constant, it will always peak at the same frequency (measured in parts per million, ppm).

By calculating the area under the peak, the concentration of a metabolite relative to an inter‐

are still being optimized, but will be extremely useful in understanding AD.

ciencies can be recognized at a young age.

**5.1. Magnetic resonance spectroscopy**

nal standard can be obtained.

**5. Chemical imaging**

238 Understanding Alzheimer's Disease

**Figure 6.** Example of chemical shift spectrum from a normal brain (from the University of Missouri-Kansas City Radiol‐ ogy Resident Resource Webpage

The most common macromolecules studied in neuroimaging are creatine (Cr) which is usually unaffected by disease and can act as an internal standard, *N-*acetyl-aspartate (NAA), a marker for neuronal health, myo-inositol (mI), a marker of gliosis [32,66], and choline (Cho). In AD, NAA is typically decreased in AD and mI is typically increased so NAA/Cr, mI/Cr and NAA/mI ratios are good markers of the disease with the ratio of NAA to mI being the strongest. The mI concentration has been found to be elevated in aMCI [33,34,67]. There is a trend that NAA is decreased in aMCI, however the effect is much more mild if it exists [35-37,68]. Glutamate (Glu) is the primary excitatory neuro‐ transmitter, and is significantly reduced in AD [69] Gamma-amino butyric acid (GABA) is an inhibitory neurotransmitter, and its concentration may be decreased in AD [70]. It is possible to use MRS to estimate relative concentrations of both Glu and GABA *in vivo*, but due to their low concentrations compared to NAA and other metabolites, and the fact that signal from these metabolites is very close in frequency space to other metabo‐ lites of larger concentration, it is more difficult to identify them without suppressing or "editing" other signals. One of the common methods used to quantify GABA is a PRESS or MEGA-PRESS sequence which suppresses or edits signals from water, creatine, and other nearby metabolites so that the characteristic GABA peaks can be identified. For more information on the MEGA-PRESS sequence, see Waddell 2007 [71].

The importance of MRS research is clear, but there are some difficulties associated with it. To begin with, the scans take a long time to complete—more than ten minutes in some cases—and because the measurements are taken in a single voxel the subject must stay absolutely still throughout the scan. This is very difficult for young healthy subjects, and may be nearly impossible in older, demented subjects. Common sedation drugs such as propofol will change the levels of brain metabolites and should be avoided[72]. In premenopausal women GABA levels also vary depending on the stage of the menstrual cycle, and may introduce variability[73].

1 Department of Neurology, Vanderbilt University, Nashville, TN, USA

2 Department of Radiology, Vanderbilt University, Nashville, TN, USA

riers. Brain. 2006 May;129(Pt 5):1240–8.

gia. 2008;46(6):1679–87.

[1] Lind J, Persson J, Ingvar M, Larsson A, Cruts M, Van Broeckhoven C, et al. Reduced functional brain activity response in cognitively intact apolipoprotein E epsilon4 car‐

Using Magnetic Resonance Imaging in the Early Detection of Alzheimer's Disease

http://dx.doi.org/10.5772/54445

241

[2] Wilcock DM. The usefulness and challenges of transgenic mouse models in the study of Alzheimer's disease. CNS Neurol Disord Drug Targets. 2010 Aug.;9(4):386–94.

[3] Persson J, Lind J, Larsson A, Ingvar M, Sleegers K, Van Broeckhoven C, et al. Altered deactivation in individuals with genetic risk for Alzheimer's disease. Neuropsycholo‐

[4] Fleisher AS, Sherzai A, Taylor C, Langbaum JBS, Chen K, Buxton RB. Resting-state BOLD networks versus task-associated functional MRI for distinguishing Alzheim‐

[6] Filippini N, Macintosh BJ, Hough MG, Goodwin GM, Frisoni GB, Smith SM, et al. Distinct patterns of brain activity in young carriers of the APOE-epsilon4 allele. Proc

[7] Boly M, Phillips C, Tshibanda L, Vanhaudenhuyse A, Schabus M, Dang-Vu TT, et al. Intrinsic brain activity in altered states of consciousness: how conscious is the default

[8] Greicius MD, Srivastava G, Reiss AL, Menon V. Default-mode network activity dis‐ tinguishes Alzheimer's disease from healthy aging: evidence from functional MRI.

[9] Alzheimer's Association. 2012 Alzheimer's disease facts and figures. Alzheimers De‐

[10] Plewes DB, Kucharczyk W. Physics of MRI: a primer. J Magn Reson Imaging. 2012

[11] Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta

[12] Jack CR Jr., Bernstein MA, Borowski BJ, Gunter JL, Fox NC, Thompson PM, et al. Up‐ date on the Magnetic Resonance Imaging core of the Alzheimer's Disease Neuroi‐

maging Initiative. Alzheimer's and Dementia. 2010 May;6(3):212–20.

er's disease risk groups. NeuroImage. 2009 Oct. 1;47(4):1678–90.

mode of brain function? Ann N Y Acad Sci. 2008;1129:119–29.

Proc Natl Acad Sci USA. 2004 Mar. 30;101(13):4637–42.

ment. 8th ed. 2012 Mar. 5;:131–68.

Neuropathol. 1991;82(4):239–59.

May;35(5):1038–54.

Natl Acad Sci USA. 2009 Apr. 28;106(17):7209–14.

[5] Sinha G. Peering inside Alzheimer's brains. Nat. Biotechnol. 2011 May;:384–7.

**References**

Typically, spectroscopy is done in the posterior cingulate or medial temporal cortices, but these are only affected by AD in late stages of the disease. It would be more helpful to study the smaller limbic areas that are affected sooner, but the voxel sizes typically used in spec‐ troscopy are larger than many of these areas [66]. Falini et al developed a technique to per‐ form spectroscopy across the entire brain and found that NAA levels are reduced in those with AD, however whole-brain spectroscopy is a non-specific marker [41,74]. These limita‐ tions will be overcome with higher field strength, advances in shimming algorithms, and improvements to computerized registration techniques [42,68].

#### **5.2. Magnetic Resonance Spectroscopy imaging (MRSi)**

MRSi is a technique that uses spectroscopy but applies it to voxels across the entire brain. The concentration of the chemical of interest corresponds to the brightness or color of the voxel in the image produced. It can achieve high spatial resolution (up to 0.25 cm3 ), and when optimized can produce a wealth of information [44,75]. This technique has largely been developed for breast cancer imaging, and can identify chemical "hot spots" that are of use when categorizing a tumor. It has great potential as a technique for understanding AD.
