**3. Characteristics of main neurodegenerative diseases**

#### **3.1 Diseases of the central and peripheral nervous system and muscles 3.1.1 Motor Neuron Diseases (MND)**

Motor neuron diseases describe pathologies of the motor system. The most common is Amyotrophic Lateral Sclerosis, which evolves mostly in midlife and is characterised by a fast progression of immobility and loss of verbal communication. Cognition is mostly unaffected; yet, there is an on-going debate on the proportion of patients with cognitive impairments. There are more benign forms of motor degenerative diseases e.g. affecting just the upper or just the lower motor neurons (Ludolph and Dengler, 1999). ALS, which affects upper and lower motor neurons, is usually fast progressing causing death within 3-5 years. The fatal event is usually respiratory failure. There is a gender ratio of men: women of about 1.5: 1.

The cause of ALS is mostly unknown. In the early 1990s there was evidence provided that some patients with a familial form of ALS have a defined mutation in the Superoxide-Dismutase 1 (SOD1). Since then various genetic changes have been detected as a possible cause for ALS. However, for most cases the cause is yet unknown. Riluzole as a glutamate agonist is the only applicable drug in ALS that prolongs life of ALS patients. Numerous drug trials are ongoing to provide new therapeutic targets in ALS.

#### **3.2 Disease of the central nervous system 3.2.1 Dementias**

16 Neuroimaging for Clinicians – Combining Research and Practice

Neurodegenerative diseases encompass different pathologies with the common feature of

Over the last years, evidence evolved of common pathological processes in different neurodegenerative diseases like accumulation of neurofilaments, protein degradation and induced cell death. For example, accumulated TDP-43 is found in sporadic ALS cases, which links this disease to fronto-temporal dementias (FTD) (Neumann et al., 2006). Furthermore, on behavioural level there is accumulating evidence for a common cognitive decline in some

Genetics provide further evidence for an association between different neurodegenerative diseases like polyglutamine repeats in Huntington's disease and spinocerebellar ataxias and mutations in the alpha-synuclein in Parkinson's Disease, dementia with Lewy bodies and multiple system atrophy. Furthermore, mitochondrial dysfunction, disturbed axonal transport and endothelial dysfunction are common pathological hallmarks found in different neurodegenerative diseases. In clinical routine, differential diagnosis is usually done on clinical basis. According to clinical criteria, Amyotrophic lateral sclerosis is regarded as a disease of the peripheral and central nervous system and the other most common neurodegenerative diseases like dementias, Parkinson's disease and Huntington's disease are primarily regarded as diseases of the central nervous system (Hacke, 2007). However, evidence for overlaps in molecular or cellular pathways question this classification and suggests a major overlap of different neurodegenerative pathologies in central and peripheral nervous system (Braak et al., 2006b; Braak & Del Tredici, 2011a, 2011b; Grammas et al., 2011; Neumann et al., 2006). New imaging techniques carry the hope of revolutionizing the diagnosis of neurodegenerative disease to improve staging of patients

**2. Clinical understanding and differential diagnosis among** 

**neurodegenerative diseases** 

ALS patients and in patients with FTD.

**3.1.1 Motor Neuron Diseases (MND)** 

1.5: 1.

progressive loss of structure or function of neurons.

and follow disease progression and treatment trial efficacy.

**3. Characteristics of main neurodegenerative diseases** 

drug trials are ongoing to provide new therapeutic targets in ALS.

**3.1 Diseases of the central and peripheral nervous system and muscles** 

Motor neuron diseases describe pathologies of the motor system. The most common is Amyotrophic Lateral Sclerosis, which evolves mostly in midlife and is characterised by a fast progression of immobility and loss of verbal communication. Cognition is mostly unaffected; yet, there is an on-going debate on the proportion of patients with cognitive impairments. There are more benign forms of motor degenerative diseases e.g. affecting just the upper or just the lower motor neurons (Ludolph and Dengler, 1999). ALS, which affects upper and lower motor neurons, is usually fast progressing causing death within 3-5 years. The fatal event is usually respiratory failure. There is a gender ratio of men: women of about

The cause of ALS is mostly unknown. In the early 1990s there was evidence provided that some patients with a familial form of ALS have a defined mutation in the Superoxide-Dismutase 1 (SOD1). Since then various genetic changes have been detected as a possible cause for ALS. However, for most cases the cause is yet unknown. Riluzole as a glutamate agonist is the only applicable drug in ALS that prolongs life of ALS patients. Numerous Alzheimer's disease (AD) is the most common form of dementia. It is named after the German psychiatrist and neuropathologist Alois Alzheimer who first described the disease in the early 20th century. The other most common form of dementia is vascular dementia. Some other forms are dementia with Lewy bodies, and frontotemporal lobal degeneration, which is subdivided into the behavioural variant (fronto-temporal dementia, FTD), the semantic dementia, primary progredient aphasia, corticobasal degeneration, progressive supranuclear palsy and amyotrophic lateral sclerosis with frontotemporal dementia. Numerous other neurodegenerative illnesses have an associated dementia, including Creutzfeldt–Jakob disease, Huntington's disease, multiple system atrophy, and Parkinson's disease dementia (Tartaglia et al.,2011).

dementias are acquired diseases that clinically affect cognitive abilities and daily activities. Classification of dementias can be done according to different criteria: cortical (memory, language, thinking and social skills are affected) and subcortical pathology (emotional processing, movement and memory are primarily affected). Furthermore, it can be classified according to whether it is a progressive form (cognitive abilities worsen over time), and whether it is primary (results from a specific disease such as Alzheimer's disease) or secondary (occurs because of disease or injury like vascular dementia). Patho-anatomical hallmark is the degeneration of the brain (mainly frontal and temporal areas). Early stages of dementia are often mistakenly considered as normal aging problems like forgetfulness and memory storage problems. With the means of standardised diagnostic tools problems in memory, language (aphasia), attention, planning and concept formation, psychomotor function and personality problems can be detected.

Due to the changing demographics in western countries, the incidence of dementias constantly increases. About 5-10% of the people >65 years and 30-40% of those above 80 develop dementias. Incidence increases exponentially with age. Women are more often affected than men.

The cause of Alzheimer's disease is not fully understood. There are several hypotheses, which have different supporters. The most widely used hypothesis is the amyloid hypothesis. Amyloid beta deposits are found in the brain of AD patients preceding the onset of clinical dementia. However, amyloid beta is not pathological per se and is found in healthy aged people. Furthermore, amyloid plaque deposition do not correlate with neuron loss and also not with clinical symptoms. Abnormally phosphorylated tau protein may start quite early, i.e., before puberty or in early young adulthood and therefore decades before clinical onset of the disease (Braak & Del Tredici, 2011b).

#### **3.2.2 Parkinson's Disease (PD)**

Parkinson's disease is the second most common degenerative disease of the central nervous system. Pathological hallmark of the idiopathic Parkinson's syndrome are movement related symptoms like slowness of movements, rigidity and tremor. Subtypes distinguish between the predominance of symptoms in a patient. Changes in mood and cognitive deficits are described as non-motor symptoms in PD. In late stage PD dementia is a common hallmark. Like ALS, PD affects people in midlife. However, there is evidence that disease process starts early in life (Braak & Del Tredici 2011b). The disease is caused by death of

Degeneration of the Human Nervous System and Magnetic Resonance Neuroimaging 19

structures in the scull and in the spinal canal. Furthermore, in functional MRI it is applicable

Whereas MRI gives only information about the structure of the body like the distribution of water and fat, magnetic resonance spectroscopy is suitable to give information of property and chemical structure of tissue. MRS provides information on the nuclei of atoms, which allows deduction on the chemical properties of a tissue. It is a non-invasive technique, which is used in some clinical fields like tumour diagnostic. It can as well provide evidence of longitudinal change in cerebral function using proton-based metabolites (among others choline, creatine, lactate, N-acetylaspartate (NAA), glutamate). NAA is thought to be a marker of neuronal integrity and is therefore used as a diagnostic marker in neurodegenerative diseases. MRS has been used in research but fields of clinical application are expanding i.e. longitudinal change of metabolites in therapeutic intervention (Turner et

Voxel based morphometry (VBM) in MRI quantifies white or gray matter volume in the CNS. It is a non-invasive technique to detect brain volume changes *in vivo* and compare it between groups e.g. patients with neurodegenerative diseases and healthies. VBM registers every brain to a brain template to provide for main anatomical differences between brains. Statistical comparison of each subvolume of the brain (so called voxels) allows fast

Diffusion tensor imaging (DTI) is based on the physics of diffusion of all molecules in e.g. the human body according to Brownian motion theory. Membranes, fibres and other molecules restrict the movement of molecules and the molecules align along barriers. The stronger the aligned diffusion, the higher is the anisotropy. In living tissue, fractional anisotropy (FA) is at its minimum (0) where there are no barriers and diffusion is not directed. FA is at its maximum (1) if alignment of diffusion is highest. DTI is applicable in white matter pathology in neurodegenerative diseases. DTI based fibre tracking gives additional information on directionality and can therefore be used to visualise e.g. direction and strength of bundles of white matter fibres within the brain. In neurodegenerative

Functional neuroimaging in neurodegenerative disease aims to explore the functional state of the brain as well as the capacity of the adult brain to functionally compensate for progressive loss of neurons. In order to map brain functions, non-invasive neuroimaging techniques have been available for almost 80 years. Since different techniques have different shortcomings, the development and implementation of new functional imaging techniques have been complementary over these years (for review see Lulé et al., 2009). Electroencephalography (EEG), already developed in the 1920s by Hans Berger, is a technique for directly measuring electrical activity of cortical neurons on the surface of the head, thus providing a high temporal but a very low spatial resolution (Berger, 1929). In

to non-invasively visualise brain function *in vivo*.

**4.1.3 Voxel based morphometry MRI** 

quantification of brain volume alterations.

diseases it's applicable to detect white and grey matter loss.

**4.1.4 Diffusion tensor imaging** 

**4.2 Functional neuroimaging** 

**4.1.2 Spectroscopy** 

al., 2011).

dopaminergic cells in substantia nigra of the midbrain that project to the striatum. The pathological process of PD (formation of proteinaceous intraneuronal Lewy bodies and Lewy neurites) begins at two sites and continues in a topographically predictable sequence in six stages, during which components of the olfactory, autonomic, limbic, and somatomotor systems become progressively involved (Braak et al., 2006a). PD cannot be cured but dopaminergic medication is an effective treatment for the disease. However, medication may become ineffective in the cause of disease. Deep brain stimulation is an available tool in PD if no other therapy is applicable.

### **3.2.3 Huntington's Disease**

Huntington's disease (HD) is a movement disorder with inability to control movements: involuntary, sudden, fast and erratic movements of distal extremities, face, neck and trunk are seen. In early stages of HD slightly exaggerated movements might be considered as nervousness, however, in the course of the disease control of coordinated body movements are becoming more and more difficult, especially affecting walking. Men and women are similarly affected. Disease onset is usually between 30 and 50 years.

Huntington described the disease already in 1872. It has a prevalence of 2-10 per 100 000 inhabitants. It is an autosomal dominant genetic cause with full penetrance. There is a 50% chance of diseased offspring and therefore presymptomatic enrolment in clinical trials is possible.

## **4. Common neuroimaging techniques**

Neuroimaging techniques in degenerative diseases are used to investigate structural or functional changes in the brain and spinal chord. Those techniques may be used to support clinical diagnosis or, as for most of the functional techniques, are used in research to enrich our understanding of pathophysiological outcomes of neurodegenerative processes.

### **4.1 Structural techniques**

#### **4.1.1 Anatomical scans with CT and MRI**

Conventional radiography like computerized tomography (CT) is still widely used in traumatology and other fields of medicine. For CT, x-rays pass through the body and are attenuated in the tissue. The denser a tissue is, the more the x-rays are attenuated. Detectors pick up the signals and digital geometry processing generates three-dimensional images of e.g. the brain. CT has mainly lost its importance in the clinical evaluation of neurodegenerative diseases.

With the finding of magnetic resonance imaging (MRI) in the 1970s, a new era of medical neuroimaging evolved to visualize structures of the central nervous system. MRI uses the property of magnetic resonance in atoms with uneven number of nuclei. In living organisms, protons are mostly used, but any other atom with the according properties can similarly be used. An object (e.g. a human) is placed within a permanent magnetic field. A proportion of the protons align within this field. Gradient pulses in a high radio frequency are pulsed to deflect the spins of the protons. Inbetween those pulsed gradients, spins return to their original position and emit energy. This emission can be detected in MRI.

MRI images provide images of soft tissue e.g. the brain with high contrasts and without bone artefacts. Therefore, MRI is applicable to visualize anatomical and pathological structures in the scull and in the spinal canal. Furthermore, in functional MRI it is applicable to non-invasively visualise brain function *in vivo*.

#### **4.1.2 Spectroscopy**

18 Neuroimaging for Clinicians – Combining Research and Practice

dopaminergic cells in substantia nigra of the midbrain that project to the striatum. The pathological process of PD (formation of proteinaceous intraneuronal Lewy bodies and Lewy neurites) begins at two sites and continues in a topographically predictable sequence in six stages, during which components of the olfactory, autonomic, limbic, and somatomotor systems become progressively involved (Braak et al., 2006a). PD cannot be cured but dopaminergic medication is an effective treatment for the disease. However, medication may become ineffective in the cause of disease. Deep brain stimulation is an

Huntington's disease (HD) is a movement disorder with inability to control movements: involuntary, sudden, fast and erratic movements of distal extremities, face, neck and trunk are seen. In early stages of HD slightly exaggerated movements might be considered as nervousness, however, in the course of the disease control of coordinated body movements are becoming more and more difficult, especially affecting walking. Men and women are

Huntington described the disease already in 1872. It has a prevalence of 2-10 per 100 000 inhabitants. It is an autosomal dominant genetic cause with full penetrance. There is a 50% chance of diseased offspring and therefore presymptomatic enrolment in clinical trials is

Neuroimaging techniques in degenerative diseases are used to investigate structural or functional changes in the brain and spinal chord. Those techniques may be used to support clinical diagnosis or, as for most of the functional techniques, are used in research to enrich

Conventional radiography like computerized tomography (CT) is still widely used in traumatology and other fields of medicine. For CT, x-rays pass through the body and are attenuated in the tissue. The denser a tissue is, the more the x-rays are attenuated. Detectors pick up the signals and digital geometry processing generates three-dimensional images of e.g. the brain. CT has mainly lost its importance in the clinical evaluation of

With the finding of magnetic resonance imaging (MRI) in the 1970s, a new era of medical neuroimaging evolved to visualize structures of the central nervous system. MRI uses the property of magnetic resonance in atoms with uneven number of nuclei. In living organisms, protons are mostly used, but any other atom with the according properties can similarly be used. An object (e.g. a human) is placed within a permanent magnetic field. A proportion of the protons align within this field. Gradient pulses in a high radio frequency are pulsed to deflect the spins of the protons. Inbetween those pulsed gradients, spins return

MRI images provide images of soft tissue e.g. the brain with high contrasts and without bone artefacts. Therefore, MRI is applicable to visualize anatomical and pathological

to their original position and emit energy. This emission can be detected in MRI.

our understanding of pathophysiological outcomes of neurodegenerative processes.

available tool in PD if no other therapy is applicable.

**4. Common neuroimaging techniques** 

**4.1.1 Anatomical scans with CT and MRI** 

similarly affected. Disease onset is usually between 30 and 50 years.

**3.2.3 Huntington's Disease** 

**4.1 Structural techniques** 

neurodegenerative diseases.

possible.

Whereas MRI gives only information about the structure of the body like the distribution of water and fat, magnetic resonance spectroscopy is suitable to give information of property and chemical structure of tissue. MRS provides information on the nuclei of atoms, which allows deduction on the chemical properties of a tissue. It is a non-invasive technique, which is used in some clinical fields like tumour diagnostic. It can as well provide evidence of longitudinal change in cerebral function using proton-based metabolites (among others choline, creatine, lactate, N-acetylaspartate (NAA), glutamate). NAA is thought to be a marker of neuronal integrity and is therefore used as a diagnostic marker in neurodegenerative diseases. MRS has been used in research but fields of clinical application are expanding i.e. longitudinal change of metabolites in therapeutic intervention (Turner et al., 2011).

#### **4.1.3 Voxel based morphometry MRI**

Voxel based morphometry (VBM) in MRI quantifies white or gray matter volume in the CNS. It is a non-invasive technique to detect brain volume changes *in vivo* and compare it between groups e.g. patients with neurodegenerative diseases and healthies. VBM registers every brain to a brain template to provide for main anatomical differences between brains. Statistical comparison of each subvolume of the brain (so called voxels) allows fast quantification of brain volume alterations.

#### **4.1.4 Diffusion tensor imaging**

Diffusion tensor imaging (DTI) is based on the physics of diffusion of all molecules in e.g. the human body according to Brownian motion theory. Membranes, fibres and other molecules restrict the movement of molecules and the molecules align along barriers. The stronger the aligned diffusion, the higher is the anisotropy. In living tissue, fractional anisotropy (FA) is at its minimum (0) where there are no barriers and diffusion is not directed. FA is at its maximum (1) if alignment of diffusion is highest. DTI is applicable in white matter pathology in neurodegenerative diseases. DTI based fibre tracking gives additional information on directionality and can therefore be used to visualise e.g. direction and strength of bundles of white matter fibres within the brain. In neurodegenerative diseases it's applicable to detect white and grey matter loss.

#### **4.2 Functional neuroimaging**

Functional neuroimaging in neurodegenerative disease aims to explore the functional state of the brain as well as the capacity of the adult brain to functionally compensate for progressive loss of neurons. In order to map brain functions, non-invasive neuroimaging techniques have been available for almost 80 years. Since different techniques have different shortcomings, the development and implementation of new functional imaging techniques have been complementary over these years (for review see Lulé et al., 2009). Electroencephalography (EEG), already developed in the 1920s by Hans Berger, is a technique for directly measuring electrical activity of cortical neurons on the surface of the head, thus providing a high temporal but a very low spatial resolution (Berger, 1929). In

Degeneration of the Human Nervous System and Magnetic Resonance Neuroimaging 21

et al., 1992). The increase in blood flow and oxygenation decreases deoxyhaemoglobin concentration and leads to an increase in the corresponding MRI signal intensity and the effective spin-spin relaxation time T2\*. This can be measured as a BOLD signal change in

The time curve of the measured BOLD signal in fMRI may be explained as follows: The local increase in deoxygenated blood level following neuronal activity in the corresponding brain area is assumed to be represented by the "initial dip" in the relevant BOLD signal. Oxygenated blood invades brain areas shortly after to compensate the increased metabolic rate in the neurons which results in a BOLD signal increase that peaks after around 6 s and

**Time series of BOLD response**

**b**

**0 10 20 30**

**c**

**undershoot"**

**time [s]**

**"initial dip" "post stimulus** 

Fig. 1. Time series of BOLD response in fMRI. BOLD signal decreases as oxygen expenditure

The BOLD in fMRI does not directly measure neural activity but relies on a surrogate 'secondary' signal resulting from changes in oxygenation, perfusion (blood volume and flow), and metabolism (e.g. glucose and oxygen consumption) (Logothesis et al., 2004; Kim et al., 1997b; DiSalle et al., 1999). Neurovascular and metabolic correlates associated with brain activation are not yet fully understood, but there is evidence for a correlation between neuronal activity (or activation) in the brain and the fMRI signal (Logothesis et al., 2001). Functional MRI gives an approximation of neuronal activity, detecting, for example, taskinduced changes in local brain function (DiSalle et al., 1999; van Geuns et al., 1999). Because

in the brain tissue (a: initial dip) increases; in the following phase extensive flow of oxygenated blood leads to increase in BOLD signal, reaching its maximum at about 6s (b).

The signal returns to baseline within 20-30 s, often combined with a post-stimulus

fMRI (Kim et al., 1997a; Logothetis et al., 2004; Ogawa et al., 1990b).

returns to baseline after 20–30 s (Figure 1).

**signal intensity**

undershoot of the BOLD response (c).

**a**

1968, the first measurements of the magnetic equivalents of EEG recordings (electrical activity of cortical neurons induce magnetic fields) signalled the beginning of magneto encephalography (MEG), which complemented the field of non-invasive imaging of neuronal activity in the brain (Cohen, 1968). However, the interpretation of signals in the spatial domain remains challenging and is not suitable for subcortical structures.

Since the mid 1970s, methods for measuring brain metabolism have been established. Changes in metabolism in the brain are a consequence of energy expenditure following neuronal activity in the brain. These data facilitate indirect measurement of overall brain activity. Accordingly, measurements of metabolism have improved spatial resolution especially for subcortical structures. Positron Emission Tomography (PET) and Single Photon Emission Computed Tomography (SPECT) record the dynamic distribution in the human brain of isotopes administered to a subject when assigned to a specific task (Phelps et al., 1975; Ter-Pogossian et al., 1975). Since metabolism is a second order effect following the electrical activity of neurons with latency, the temporal resolution of such measurements is low. Furthermore, radioactive substances have to be applied, limiting the application for scientific research.

In the 1990s, functional MRI was developed. Functional MRI can be used to measure physiological changes not only of metabolism but also of e.g. blood flow in the brain. Like metabolism, blood flow is stimulated following oxygen expenditure of the cortical substrate. Like PET and SPECT, fMRI is a second-order signal with the problem of low temporal resolution but high spatial resolution, with the potential to indirectly measure cortical and subcortical activity easily during the performance of a given task. Accordingly, fMRI combines advantages of different non-invasive functional neuroimaging techniques. However, the indirect nature by which brain activity is currently measured by fMRI continues to limit its role as a "front-line" imaging tool.

Notwithstanding, the clinical potential of a non-invasive probe of brain function with the option of repeated measures over time (e.g. due to a lack of radiation charge) in addition to the wide-spread availability of MRI scanners in many hospitals and research centres have extended its application in clinical science and contributed to an exponential increase in scientific publications on fMRI over the last decade (Jezzard & Buxton, 2006).

#### **4.2.1 Principles of fMRI**

Experimental work in animals first demonstrated that oxygenated blood and deoxygenated blood present different properties in a magnetic field, as noted by Linus Pauling as early as the 1930s (Pauling, 1936). Because of its unshielded iron, deoxygenated blood has paramagnetic properties whereas oxygenated blood has diamagnetic properties. Deoxyhaemoglobin as an endogenous paramagnetic contrast agent dephases nuclear spins of water protons in its vicinity with a physical effect of signal intensity change in T2\* weighted MR images (Frahms et al., 1999). Ogawa and co-workers realised that those differences in magnetic properties in blood could have implications in the visualisation of local brain function (Ogawa et al., 1990a; 1990b).

The most commonly used fMRI approach is to measure mainly blood flow changes using the blood-oxygenation-level-dependency (BOLD) effect (Kwong et al., 1992; Ogawa et al., 1992), although other parameters can also be measured. A change in neuronal activity causes a decreased local blood oxygenation and an increased demand for oxygen. The local increase in deoxygenated blood level in the corresponding brain area is followed by a rise in cerebral blood flow that at least transiently 'uncouples' from oxygen consumption (Frahms

1968, the first measurements of the magnetic equivalents of EEG recordings (electrical activity of cortical neurons induce magnetic fields) signalled the beginning of magneto encephalography (MEG), which complemented the field of non-invasive imaging of neuronal activity in the brain (Cohen, 1968). However, the interpretation of signals in the

Since the mid 1970s, methods for measuring brain metabolism have been established. Changes in metabolism in the brain are a consequence of energy expenditure following neuronal activity in the brain. These data facilitate indirect measurement of overall brain activity. Accordingly, measurements of metabolism have improved spatial resolution especially for subcortical structures. Positron Emission Tomography (PET) and Single Photon Emission Computed Tomography (SPECT) record the dynamic distribution in the human brain of isotopes administered to a subject when assigned to a specific task (Phelps et al., 1975; Ter-Pogossian et al., 1975). Since metabolism is a second order effect following the electrical activity of neurons with latency, the temporal resolution of such measurements is low. Furthermore, radioactive substances have to be applied, limiting the application for

In the 1990s, functional MRI was developed. Functional MRI can be used to measure physiological changes not only of metabolism but also of e.g. blood flow in the brain. Like metabolism, blood flow is stimulated following oxygen expenditure of the cortical substrate. Like PET and SPECT, fMRI is a second-order signal with the problem of low temporal resolution but high spatial resolution, with the potential to indirectly measure cortical and subcortical activity easily during the performance of a given task. Accordingly, fMRI combines advantages of different non-invasive functional neuroimaging techniques. However, the indirect nature by which brain activity is currently measured by fMRI

Notwithstanding, the clinical potential of a non-invasive probe of brain function with the option of repeated measures over time (e.g. due to a lack of radiation charge) in addition to the wide-spread availability of MRI scanners in many hospitals and research centres have extended its application in clinical science and contributed to an exponential increase in

Experimental work in animals first demonstrated that oxygenated blood and deoxygenated blood present different properties in a magnetic field, as noted by Linus Pauling as early as the 1930s (Pauling, 1936). Because of its unshielded iron, deoxygenated blood has paramagnetic properties whereas oxygenated blood has diamagnetic properties. Deoxyhaemoglobin as an endogenous paramagnetic contrast agent dephases nuclear spins of water protons in its vicinity with a physical effect of signal intensity change in T2\* weighted MR images (Frahms et al., 1999). Ogawa and co-workers realised that those differences in magnetic properties in blood could have implications in the visualisation of

The most commonly used fMRI approach is to measure mainly blood flow changes using the blood-oxygenation-level-dependency (BOLD) effect (Kwong et al., 1992; Ogawa et al., 1992), although other parameters can also be measured. A change in neuronal activity causes a decreased local blood oxygenation and an increased demand for oxygen. The local increase in deoxygenated blood level in the corresponding brain area is followed by a rise in cerebral blood flow that at least transiently 'uncouples' from oxygen consumption (Frahms

scientific publications on fMRI over the last decade (Jezzard & Buxton, 2006).

continues to limit its role as a "front-line" imaging tool.

local brain function (Ogawa et al., 1990a; 1990b).

spatial domain remains challenging and is not suitable for subcortical structures.

scientific research.

**4.2.1 Principles of fMRI** 

et al., 1992). The increase in blood flow and oxygenation decreases deoxyhaemoglobin concentration and leads to an increase in the corresponding MRI signal intensity and the effective spin-spin relaxation time T2\*. This can be measured as a BOLD signal change in fMRI (Kim et al., 1997a; Logothetis et al., 2004; Ogawa et al., 1990b).

The time curve of the measured BOLD signal in fMRI may be explained as follows: The local increase in deoxygenated blood level following neuronal activity in the corresponding brain area is assumed to be represented by the "initial dip" in the relevant BOLD signal. Oxygenated blood invades brain areas shortly after to compensate the increased metabolic rate in the neurons which results in a BOLD signal increase that peaks after around 6 s and returns to baseline after 20–30 s (Figure 1).

**Time series of BOLD response**

Fig. 1. Time series of BOLD response in fMRI. BOLD signal decreases as oxygen expenditure in the brain tissue (a: initial dip) increases; in the following phase extensive flow of oxygenated blood leads to increase in BOLD signal, reaching its maximum at about 6s (b). The signal returns to baseline within 20-30 s, often combined with a post-stimulus undershoot of the BOLD response (c).

The BOLD in fMRI does not directly measure neural activity but relies on a surrogate 'secondary' signal resulting from changes in oxygenation, perfusion (blood volume and flow), and metabolism (e.g. glucose and oxygen consumption) (Logothesis et al., 2004; Kim et al., 1997b; DiSalle et al., 1999). Neurovascular and metabolic correlates associated with brain activation are not yet fully understood, but there is evidence for a correlation between neuronal activity (or activation) in the brain and the fMRI signal (Logothesis et al., 2001). Functional MRI gives an approximation of neuronal activity, detecting, for example, taskinduced changes in local brain function (DiSalle et al., 1999; van Geuns et al., 1999). Because

Degeneration of the Human Nervous System and Magnetic Resonance Neuroimaging 23

increased activity in motor areas was observed in fifteen ALS patients compared to fifteen healthy controls during a sequential finger tapping movement task (Han et al., 2006) and in ALS patients compared both to patients with upper limb weakness due to peripheral nerve lesions and to controls during freely selected random joystick movements of the right hand (Stanton et al., 2007a). It has been proposed that these changes may represent cortical plasticity, as new synapses and pathways are developed to compensate for the selective loss of pyramidal cells in the motor cortex (Schoenfeld et al., 2005). A shift of activity to more anterior regions of the premotor cortex, i.e. Brodmann area (BA) 6, during upper limb movement has been observed in ALS patients (Konrad et al., 2002; Han et al., 2006), such findings being supported by previous functional imaging studies with PET (Kew et al., 1993a; 1994). Furthermore, there is longitudinal fMRI evidence of progressive involvement of the premotor area in upper limb motor tasks in the course of the disease (Lulé et al., 2007a). Thirteen patients with sporadic ALS and 14 healthy controls were asked to perform tasks involving a grip movement of the left, right, and both hands and to imagine the same without any overt movement of the hand. Motor imagery is known to involve similar areas as motor execution without being affected by confounding factors of effort and strain. In two consecutive fMRI measurements at a six-month interval, evidence for progressive recruitment of premotor areas in motor imagery was found in the course of the disease (Lulé

Furthermore, a changed pattern and an anterior shift of activity in ALS were also observed in further cortical areas besides the premotor cortex for various motor tasks. For instance, increased involvement of supplementary motor areas (SMA) (Konrad et al., 2002; 2006; Han et al., 2006) and sensorimotor cortices has been seen (Brooks et al., 2000; Han et al., 2006; Stanton et al., 2007a; Mohammadi et al., 2011). Activity in contralateral sensorimotor cortex activity was increased the stronger the physical impairments were in patients (Mohammadi et al., 2011). Furthermore, activity in adjacent areas such as the bilateral inferior parietal lobe (BA 40) and bilateral superior temporal gyrus (BA 22) was increased in ALS patients compared to healthy controls during upper limb motor task performance in different fMRI studies (Stanton et al., 2007a). Altered somatotopy in the sensorimotor cortices was not observed in patients with exclusive lower motor neuron involvement (Kew et al., 1994), but only in ALS patients with clinical and functional involvement of both upper and lower motor neurons (Han et al., 2006; Kew et al., 1993a; 1994) or upper motor neuron only (Stanton et al., 2007a). This suggests that this changed pattern of activity might represent the loss of the pyramidal tract (Kew et al., 1994). A similar shift of activity in motor tasks into more anterior regions of sensorimotor and premotor areas and the SMA has been demonstrated for different neuropathologies with distinct aetiology such as stroke (Weiller et al., 2006). Thus, it may be assumed that this anterior shift represents a general pattern of plasticity as a response to neuronal loss in primary motor areas as a more or less efficient way to compensate motor function rather than an ALS-specific pattern of altered motor

Increasing activity in ipsilateral cortical areas such as the sensorimotor cortex (Han et al., 2006; Stanton et al., 2007a) and primary motor areas (Schoenfeld et al., 2005) has been observed in ALS patients. In a motor task of upper limb movement with varying task difficulty, six ALS patients presented an increased activity in ipsilateral primary motor areas compared to six healthy controls, corresponding to the degree of difficulty (Schoenfeld et al., 2005). The fact that healthy controls recruit ipsilateral areas with increasing complexity of a

et al., 2007a).

activation (Weiller et al., 2006).

the BOLD signal, however, derives from the interaction of multiple parameters (e.g. perfusion, metabolic turnover of neurons, density of venous vasculature of tissue, medication etc.) and may vary between brain areas and individuals as well as experimental and clinical settings, quantitative analysis in absolute terms is precluded (Kim et al., 1997a; Di Salle et al., 1999; Logothesis et al., 2001).

MRI sequences that are best suited for functional neuroimaging should be both fast and sensitive to changes in the deoxyhaemoglobin concentration (Frahms et al., 1999). The MRI sequence that is generally considered to be the first choice for measuring BOLD in fMRI is a T2\*-weighted echo-planar imaging (EPI) sequence, with its high speed yielding imaging times which translate into a maximum temporal resolution (Edelman et al., 1994; Kwong et al., 1995; Schmitt et al., 1998; Roberts et al., 2007). Temporal and spatial parameters of MRI scanning such as repetition time (TR) or slice thickness are limited by the T2\* signal decay of the MRI sequence and determined according to study-specific factors, e.g. region of interest and field strength (Schmitt et al., 1998; Turner et al., 1998; Triantafyllou et al., 2005; Norris et al., 2006; Figure 2). Other MRI sequences such as fast low angle shot (FLASH) techniques facilitate access to higher spatial resolution at the expense of temporal resolution and volume coverage of larger volumes leading to higher scanning times (Frahms et al., 1999), which is not always favourable in clinical settings.

Fig. 2. Example of orientation of EPI-sequence (white lines) along the anterior-posterior commissure overlaid onto a t1 weighted image (mprage) of a head (sagittal view)
