Quantitative functional magnetic resonance imaging in...
128
1 Quantitative functional magnetic resonance imaging in cerebral small vessel disease Thesis submitted in accordance with the requirements of the University of Liverpool for the degree of Master of Philosophy by Guy Lumley. April 2011
Transcript of Quantitative functional magnetic resonance imaging in...
1
Quantitative functional magnetic resonance
imaging in cerebral small vessel disease
Thesis submitted in accordance with the requirements of the University of Liverpool for
the degree of Master of Philosophy by Guy Lumley.
April 2011
2
Abstract
Introduction: Cerebral small vessel disease (cSVD) is an important, but relatively
poorly understood cause of both lacunar strokes and vascular dementia. Structural
magnetic resonance imaging (MRI) markers of cSVD, including lacunes, white matter
lesions (WML) and microbleeds, have been shown not to correlate consistently with
clinical severity, as gauged by cognitive decline, and might offer little more than
endpoint markers of disease. However, alternative developing MR techniques,
including functional MRI (fMRI) using the blood-oxygen-level-dependent (BOLD)
signal, offer a promising approach to charting disease severity.
Aims: The primary aim is to determine whether „n‟, a measure of neurovascular
coupling (NVC) which underpins interpretation of the BOLD signal, differs between
patients with cSVD and healthy matched controls. If „n‟ does differ, a secondary aim is
to determine whether „n‟ correlates with tests of cognitive function.
Methods: Eleven patients with cSVD and sixteen age-, education- and gender-matched
healthy controls were recruited. Participants underwent a battery of cognitive tests
focused upon executive functions and a series of MRI scans. These included structural
scans, arterial spin labelling (ASL) to measure cerebral blood flow and BOLD signal.
Oxygen calibrated fMRI was used with a modified Stroop Interference Task.
Results: The cSVD group performed worse on the digit symbol substitution test
(DSST) (p = 0.00005) than the control group. There was a significantly different
BOLD response in 11 regions between patient and control groups, which were
aggregated into frontal, parietal, motor, insular and total regions. „n‟ was reduced
across total regions (p = 0.02) in the patient group. „M‟ was increased in the patient
group and correlated inversely with „n‟. DSST did not correlate with „n‟ in patients.
Conclusion: The results suggest an uncoupling of the neurovascular response in
patients with cSVD, possibly associated with an increase in the oxygen extraction
fraction. A larger sample size would be needed to investigate whether altered
neurovascular coupling might highlight at-risk subjects who have not yet had a stroke.
3
Contents
Abstract............................................................................................................................2
List of Figures..................................................................................................................7
List of Tables....................................................................................................................9
List of Boxes.....................................................................................................................9
Chapter 1: Introduction................................................................................................12
1.1 Cerebral small vessel disease........................................................................12
1.1.1 White matter lesions.......................................................................12
1.1.2 Cerebral microbleeds.....................................................................15
1.1.3 Lacunar infarcts.............................................................................16
1.1.4 Silent subcortical infarcts...............................................................17
1.1.5 Dilated perivascular spaces...........................................................19
1.1.6 Vascular dementia..........................................................................19
1.1.7 Pathology of cSVD..........................................................................20
1.1.8 Diagnosis of cSVD..........................................................................23
1.1.9 Risk factors for cSVD.....................................................................23
1.1.10 Management of cSVD...................................................................24
1.1.11 Prognosis of cSVD........................................................................26
1.2 Cognitive disturbances in cSVD...................................................................27
Chapter 2: Review of imaging markers of cSVD........................................................30
2.1 WML.............................................................................................................31
2.2 Lacunes..........................................................................................................32
2.3 Atrophy..........................................................................................................33
2.4 DTI measures.................................................................................................34
4
2.5 MRS...............................................................................................................35
2.6 Cerebral perfusion and glucose metabolism..................................................35
2.7 Amplitude of vascular response to Stroop Interference Task........................36
2.8 Discussion......................................................................................................36
Chapter 3: MRI.............................................................................................................39
3.1 MRI Theory...................................................................................................39
3.2 Functional MRI..............................................................................................42
3.2.1 CBF measurement using ASL.........................................................42
3.2.2 BOLD signal...................................................................................44
3.2.3 Oxygen calibration.........................................................................47
3.3 Aims...............................................................................................................49
Chapter 4: Methods.......................................................................................................50
4.1 Subjects..........................................................................................................50
4.1.1 Sample size calculation..................................................................52
4.2 Neuropsychological tests...............................................................................53
4.3 Imaging..........................................................................................................55
4.3.1 Image acquisition...........................................................................55
4.3.2 Stimulus paradigm..........................................................................56
4.3.3 Hyperoxia paradigm.......................................................................56
4.4 Data analysis..................................................................................................57
4.4.1 Identifying the active regions of interest........................................57
4.4.2 Quantification of neural and vascular parameters........................58
4.4.3 Assessment of cerebral atrophy and ventricular enlargement.......61
4.4.4 Assessment of WML........................................................................62
4.4.5 Assessment of microbleeds.............................................................62
4.4.6 Angiography...................................................................................62
4.5 Statistical analysis..........................................................................................63
5
Chapter 5: Results.........................................................................................................64
5.1 Recruitment...................................................................................................64
5.2 Participant demographics..............................................................................65
5.3 Neuropsychological test results.....................................................................66
5.4 Stroop Interference Task...............................................................................67
5.5 Neurovascular coupling parameter „n‟..........................................................67
5.6 Additional functional measures.....................................................................72
5.6.1 BOLD..............................................................................................72
5.6.2 CMRO2............................................................................................74
5.6.3 CBF.................................................................................................74
5.6.4 CBF in response to Stroop Interference Task…………….……...75
5.6.5 ‘M’..................................................................................................76
5.7 Structural measures........................................................................................79
5.7.1 Cerebral volumes............................................................................79
3.7.2 WML...............................................................................................81
3.7.3 Microbleeds....................................................................................81
Chapter 6: Discussion....................................................................................................82
6.1 Discussion of findings...................................................................................82
6.2 Strengths of the study....................................................................................88
6.3 Limitations of the study.................................................................................89
6.3.1 Limitations of studying cSVD.........................................................91
Chapter 7: Conclusions.................................................................................................94
Appendix A: Table of reviewed studies......................................................................95
Appendix B: Participant consent form.......................................................................99
Appendix C: Heading letter for patients..................................................................101
6
Appendix D: Patient information sheet....................................................................102
Appendix E: Presentations resulting from this work..............................................106
Glossary........................................................................................................................107
References.....................................................................................................................110
7
List of Figures
1. WML (arrows) on FLAIR (Fluid attenuated inversion recovery) image, axial
view.
2. Susceptibility-weighted image of a cerebral microbleed (arrow), axial view.
3. Lacunar infarct (arrow) on FLAIR image, axial view.
4. Axial view of T1-weighted images from two individuals, one with (right) and one
without (left) evidence of cerebral atrophy.
5. Diagram to show the movements of a proton (shown as a thin arrow): spin
about its axis (a); and precession (b) with the proton shown at three different
time-points. The vertical block arrow represents the direction of the applied
external magnetic field.
6. Diagram showing protons (each represented by a single arrow) in (a) and out of
(b) phase.
7. Examples of, from left to right, T1-, T2-, and FLAIR-weighted images.
8. Diagram of BOLD signal, for a stimulus of duration 30 seconds.
9. Summary flow chart of events leading to BOLD signal.
10. Slice coverage.
11. Example slide from Stroop Interference Task.
12. Hyperoxia paradigm.
13. Example coronal view showing highlighted regions of interest in the left and
right precentral gyri (circled).
14. Example grid used to measure volumes of brain structures with EasyMeasure.
15. Chart to demonstrate the difficulties in achieving the desired sample size; an
example taken from individuals presenting to UHA between 01/02/2008 and
01/02/2009.
16. Bar charts with standard error bars to demonstrate the difference in accuracy
(left) and response time (right) to the Stroop Interference Task between controls
and patients.
17. ROI selected for analysis are circled.
18. Bar chart with standard error bars showing a reduction in ‘n’ in patients with
cSVD compared with controls.
8
19. Bar chart to show differences in 'n' between patient and control groups across
regions, with standard error bars
20. Comparison of DSST and ‘n’ in patient and control groups.
21. Graph showing a correlation between DSST and ‘n’ in the control group, once
an apparently anomalous result had been removed.
22. Bar chart to show the difference in BOLD signal change between patient and
control groups across selected regions, with standard error bars.
23. Averaged BOLD curves during the Stroop Interference Task across total (top)
and frontal (bottom) regions.
24. Averaged BOLD curves during the Stroop Interference Task across (from top to
bottom) motor, parietal and insular regions.
25. Bar chart to show the difference in CMRO2 between patient and control groups
across selected regions, with standard error bars.
26. Bar chart with standard error bars to show differences in CBF in response to
the Stroop Interference Task between patient and control groups.
27. Bar chart with standard error bars to show differences in CBF in response to
the Stroop Interference Task between patient and control groups.
28. Graph showing no significant correlation between CBF and ‘n’ in the patient
group.
29. Graph showing no significant correlation between CBF and ‘n’ in the control
group.
30. Bar chart with standard error bars to show the differences in M between control
and patient groups across.
31. Bar chart with standard error bars to show regional variation in M.
32. Scatter graph to show the correlation between ‘M’ and ‘n’.
33. Bar charts with standard error bars demonstrating the differences in normalised
brain structure volumes between patient and control groups.
34. Scatter graphs to show individual volume measurements plotted against DSST
score.
9
List of Tables
1. Inclusion criteria.
2. Exclusion criteria.
3. List of MRI scans and their purpose.
4. Derivation of included patients.
5. Participant demographics.
6. Mean cognitive scores ± standard deviation for patients and controls.
7. Mean normalised volumes (%) ± standard deviation.
8. Sum WML by region for patient and control groups.
9. Sum microbleeds by region for patient and control groups.
List of Boxes
1. Lacunar stroke syndromes in descending order of frequency.
2. Limitations of research into imaging markers for cSVD.
3. Advantages and disadvantages of BOLD and ASL techniques for fMRI.
4. Neuropsychological test battery.
5. One-way ANOVA with repeated measures comparing ‘n’ in ROI.
6. One-way ANOVA with repeated measures comparing ‘M’ in ROI.
7. Ideas for further study.
10
Contributions to the study
My thanks to all other contributors to the study. Ethical approval had already been
gained by Dr Hedley Emsley and Dr Laura Parkes prior to my joining the study. Final
approval of patient suitability was done by Dr Emsley. MRI scanning was done by Dr
Parkes, Ms Valerie Adams, Dr Jonathan Goodwin and Mr Rafat Mohtasib.
Neuropsychological tests and MRI safety screening were carried out by Mr Andrew
Irwin. Ten controls had been recruited prior to my arrival. All other work, including
recruitment, stroke scale (NIHSS), pre-processing, lesion measurements, analysis,
interpretation and write-up was done by myself.
11
Acknowledgements
I would like to thank my supervisors Dr Laura Parkes, Dr Hedley Emsley and Dr
Graham Kemp for all the advice and support they have given me throughout the
duration of this study. Especial thanks must be given to Dr Parkes for teaching me how
to analyse the MR images and providing custom-written MATLAB programmes. I am
grateful for the use of, and instruction on, their microbleed scale provided ahead of print
by Dr Simone Gregoire and Dr David Werring. I would also like to thank Dr Helen
Martin, Sharon Dealing, Christine Kelly and Alan Ledger for help in identifying
appropriate patients. I am indebted to all participants for agreeing to take part in the
study and also, to my parents without whose financial support my involvement in this
project would not have been possible.
12
Chapter 1: Introduction
1.1 Cerebral small vessel disease
Cerebral small vessel disease (cSVD) is an increasingly important problem in an ageing
population (1). It is responsible for a quarter of all first-ever ischaemic strokes (2-4)
and is the most common subgroup of vascular dementia (VaD) (5). cSVD can result in
a picture of dementia, gait abnormalities, urinary incontinence, dysarthria and mood
disturbances such as depression and emotional lability (6-8). The gait is classically
termed „marche à petit pas’, characterised by small shuffling steps with the arms
abducted and elbows flexed (6,7).
Patients with cSVD represent a clinically and aetiologically relatively homogeneous
population (7). It is a disease of the small perforating end arteries in the brain (5,9).
cSVD manifests itself as a number of often coexisting conditions, which may be
additive in their negative effect on the patient‟s long term prognosis (1). These
conditions include white matter ischaemic lesions (WML), microbleeds, dilated or
enlarged perivascular spaces (EPVS), infarcts in a lacunar territory either with a
concomitant clinical stroke syndrome or clinically „silent‟, plus other silent subcortical
infarcts (1,10-12). All of these can be seen on magnetic resonance imaging (MRI) and
can be considered structural markers of underlying cSVD. This study primarily aims to
evaluate the use of a functional MRI (fMRI) marker in patients with cSVD.
1.1.1 White matter lesions
WML are areas of ischaemic injury of the white matter (13). Leukoaraiosis (LA), seen
as hyperintense areas on T2-weighted MRI, is the radiological correlate of the WML
(9). Ischaemic leukoaraiosis is a term that assumes cSVD as the likely pathological
basis of the observed WML, and is usually reserved for radiological LA in a patient
with a clinical lacunar stroke (9).
WML are caused by cSVD, but are also associated with inflammation, oedema, trauma,
and toxic and metabolic disorders (13). WML are also commonly seen in the
13
apparently healthy elderly: the Cardiovascular Health Study (CHS) found that 95.6% of
3301 community dwelling elderly, aged 65 years or older, without a history of stroke or
transient ischaemic attack, had evidence of WML (14). They have been reported in
64%–100% of patients with VaD and also in 7.5%–100% of those with Alzheimer‟s
disease, though usually to a less severe extent (15).
The severity of WML in the CHS was
independently associated with increasing
age, clinically silent stroke evidenced on
MRI, raised systolic blood pressure (BP), a
reduced forced expiratory volume in one
second, and a lower income (14).
Longitudinal data from the same study
found that worsening WML were
associated with a reduction in the mini-
mental state examination (MMSE) score
and digit symbol substitution test (DSST)
score (both measures of cognitive
function) (16). The risk factors associated
with worsening lesions were complexly
related, but for a low initial severity grade
worsening WML were associated with
increased age, diastolic BP and high-density lipoprotein (HDL) cholesterol, and with
reduced low-density lipoprotein (LDL) cholesterol (16). The presence of severe WML
is linked to an increased risk of developing dementia, stroke and disability (17).
Histopathological examination of WML reveals partial loss of myelin sheaths,
oligodendroglial cells and axons (13). Pathologically, WML found in the apparently
healthy elderly, VaD, Alzheimer‟s disease or the damaged tissue between a completed
infarct and the surrounding normal tissue, are the same (13). They are thought to be
caused by hypoperfusion resulting in ischaemic injury of deep white matter (7).
Figure 1: WML (arrows) on FLAIR (Fluid
attenuated inversion recovery) image,
axial view.
14
WML may arise from periodic hypotension in chronically hypertensive patients (18).
Autoregulation of CBF allows a relatively constant supply of blood within vascular
beds during variations in arterial pressure (19). Cerebral resistance arteries dilate in
response to a reduction in perfusion pressure and constrict in response to an increase in
pressure (19). The range of perfusion pressures associated with a constant CBF is
known as the „autoregulatory plateau‟ (19). Beyond the perfusion pressure limits of the
autoregulatory plateau cerebral resistance vessels respond passively (19). Thus a further
increase in perfusion pressure beyond the plateau limit would result in increased CBF.
Chronically elevated BP leads to a rightward shift of the autoregulatory plateau (20).
This shift is associated with vascular remodelling that results in a reduced external
vessel diameter and vessel wall hypertrophy, which together lead to a narrower vascular
lumen (19). It is believed that chronically hypertensive individuals may be susceptible
to rapid drops in BP, and may develop ischaemic change in white matter even within the
BP range considered normal in the non-diseased population (15,21).
Resting CBF has been shown to be reduced in white matter, but not in normal appearing
grey matter in a study of patients with ischaemic leukoaraiosis (22). Isaka et al. (23)
found no reduction in resting CBF in subjects with asymptomatic LA. However,
cerebral vasomotor reactivity (VMR; cerebrovascular reserve) was reduced in this
group. Marstrand et al. (24) also demonstrated reduced CBF and VMR in WML
compared with normal appearing white matter. VMR represents the response of the
cerebral vessels to a vasodilatory stimulus (25). VMR has been found to be inversely
associated with WML (25)(26). This appears to follow on from the observation of a
rightward shift of the autoregulatory plateau in chronically hypertensive individuals
with damage to small cerebral vessels. In these subjects, cerebral resistance vessels
would have reduced capacity to further dilate. Thus with a drop in perfusion pressure or
administration of a vasodilatory stimulus these vessels would have a more limited
response than in an asymptomatic normotensive individual.
15
The normal cerebral endothelium is involved in CBF regulation, maintaining the blood-
brain barrier (BBB) and preventing clot formation (27). In cSVD the endothelium is
„activated‟ to participate in the inflammatory response (28). As part of this process
vascular permeability is increased, resulting in potentially toxic serum proteins entering
the perivascular neural parenchyma (29). Chronic endothelial dysfunction has been
suggested as a cause of WML, resulting from a breakdown of the BBB, impaired
cerebral autoregulation and/ or prothrombotic changes (27). Hassan et al. (27) showed
increased expression of plasma markers associated with endothelial dysfunction in
patients with cSVD compared to non-stroke controls. However, a recent systematic
review does not present convincing evidence that endothelial dysfunction is specific to
lacunar stroke, having been observed in cortical stroke also (30).
1.1.2 Cerebral microbleeds
Cerebral microbleeds are seen as
small hypointense lesions on T2*-
weighted MR images of the brain
(31). They correspond to the
breakdown products of blood, such
as haemosiderin, that persist in the
cerebral matter for many years
following a bleed from damaged
arterioles (31). Hypertensive
lipohyalinotic changes (as seen in
cSVD; described later) and cerebral
amyloid angiopathy both lead to
cerebral microbleeds but with
differing lesion distributions (32).
The reported prevalence of
microbleeds is 11.1%–23.5% in the community-dwelling elderly population (32). They
are found in a quarter of patients with ischaemic stroke (31). Greenberg et al. (32) cite
evidence to suggest that they might directly damage surrounding tissues leading to
Figure 2: Susceptibility-weighted image of a
cerebral microbleed (arrow), axial view.
16
cognitive decline, as well as acting as markers of disease. A small case-control study
showed microbleeds to be associated with executive dysfunction in stroke patients,
independent of WML (31). Microbleeds are also associated with an increased risk of
haemorrhagic transformation following ischaemic stroke (31).
1.1.3 Lacunar infarcts
Lacunar infarcts (LI) are small subcortical
infarcts that result from occlusion of a
single penetrating artery of the brain
(1,2,33). They tend to occur in the basal
ganglia, thalamus, internal capsule, corona
radiata or brainstem (1,33,34). When the
infarcts heal, they leave behind a small (3-
15mm) cerebrospinal fluid (CSF)-filled
cavity, known as a “lacune” (33,35).
Fisher argues that the term LI ought to be
reserved only for those lesions believed to
be of vascular origin (33). The
terminology of LI, lacunar strokes and
lacunes is inconsistently used in the
literature (35). In this report the definitions
set out by Wardlaw (35) shall be used:
Lacune: CSF-filled cavity. Radiological infarcts in a lacunar territory shall be
referred to as lacunes throughout.
Lacunar stroke: A clinical stroke syndrome with symptoms suggestive of a small
subcortical or brainstem infarct.
LI: A clinical stroke syndrome with a corresponding lacune (radiological
infarct).
Figure 3: Lacunar infarct (arrow) on
FLAIR image, axial view.
17
Lacunes have been classified pathologically into three types by Poirier and Derouesne
(36), with a further sub-category subsequently added by Lammie et al. (12):
I. a. Old, small infarct
b. Old, small “incomplete” infarct
II. Old, small haemorrhage
III. Dilated perivascular space
Lacunes cause symptoms if they are located in clinically eloquent sites. Although small
lesions, in the region of 3mm to 15mm (however, the definition varies) (1,33), they can
cause a large deficit if they occur within the long ascending or descending tracts, where
multiple fibres serving a large proportion of the body are found in close proximity, for
example a LI might result in weakness of one side of the body (1). Fisher described
over 20 varieties of lacunar syndromes (33). The five best documented are described in
descending order of frequency in Box 1 (33,37-40).
1.1.4 Silent subcortical infarcts
Many subcortical infarcts are not evident clinically as a discrete stroke. They are
known as „silent‟ infarcts (1). They are five times more common than LI presenting as
stroke, being present in over a quarter of all people above seventy years old (1). Silent
infarcts are found more often in patients with a LI rather than other ischaemic strokes
(1). Although they do not cause a stroke syndrome, silent infarcts are associated with
cognitive decline, depressive symptoms and a reduced likelihood of functional recovery
(1,4,41). Furthermore, their presence doubles the risk of dementia (41). One theory is
that „silent‟ lacunar infarcts reduce the capacity for neuroplasticity and recovery
mechanisms of the brain (1,42). Any future stroke would therefore lead to reduced
recovery and so, increased disability.
18
Box 1: Lacunar stroke syndromes in descending order of frequency.
Pure sensory stroke
o Numbness, with or without severe deficit of other sensory modalities,
of the face, arm and leg on one side, in the absence of weakness,
homonymous hemianopia, aphasia, agnosia or apraxia.
Although usual, not all three of face, arm and leg need be
affected.
o Usual infarct site: sensory (posteroventral) nucleus of the thalamus
Pure motor hemiparesis
o Pure motor stroke of the face, arm and leg on one side, in the absence
of sensory deficit, homonymous hemianopia, aphasia, agnosia or
apraxia
All three of face, arm and leg need to be affected.
o Usual infarct site: Posterior limb of the internal capsule; lower basis
pontis where the corticospinal tracts congregate to form the pyramids;
or, rarely, the mid-portion of cerebral peduncle.
Sensorimotor stroke
o Combination of pure sensory stroke and pure motor hemiparesis.
o Usual infarct site: Posterolateral thalamus, extending into the adjacent
posterior limb of the internal capsule.
Ataxic hemiparesis
o Pure motor hemiparesis with cerebellar dysmetria, not due to
weakness, in the affected limbs. Dysarthria, nystagmus and falling to
one side may be present.
o Usual infarct site: Pons
Dysarthria-clumsy hand syndrome
o Facial weakness, dysarthria and dysphagia are combined with slight
weakness and clumsiness of the hand. No sensory deficit.
o Usual infarct site: Basis pontis
o Considered a variant of ataxic hemiparesis.
19
1.1.5 Dilated perivascular spaces
In the central nervous system the CSF-filled cavities between the vessel wall and the
cerebral parenchyma are known as perivascular, or Virchow-Robin, spaces (PVS) (43).
These invaginations of subarachnoid space surround arteries, arterioles, veins and
venules, but not capillaries (43). PVS allow drainage of the interstitial fluid and also act
as a channel for inflammatory processes (44). Pathologically enlarged PVS (EPVS) are
associated with reduced cognitive function in normal ageing (45), hypertension (44),
inflammation such as multiple sclerosis (46), lacunar strokes (47) and LA (47). EPVS
are also associated with worsening BBB function (48).
EPVS are classified as type III lacunes (12). On T2-weighted MRI EPVS appear as
small high-signal areas running in the same direction as penetrating arterioles (47).
1.1.6 Vascular dementia
Of the population over 65 years, 8% have dementia (18). VaD is the second commonest
dementia type, after Alzheimer‟s disease (49,50). VaD makes up 9%–39% of
dementias, though a further 11%–43% may have a mixed dementia with a vascular
contribution to an Alzheimer‟s picture (18). The annual incidence of VaD is 3.8/1000
(18). Vascular cognitive impairment is a new term, as yet without consensus on
defining criteria, that refers to VaD and other cognitive impairment resulting from
vascular disease (18,51). It is associated with stroke and death from stroke (18).
Patients with cognitive impairment or dementia resulting from vascular disease
represent an important group as vascular disease has the potential to be prevented and
treated (10,18).
VaD itself is a broad term that has many causes, including large and small vessel
disease, and embolism (52,53). The most common subtype is cSVD which accounts for
36%–37% of VaD (49,54). This group has a much more homogeneous clinical picture
than other subtypes (5). cSVD incorporates the concepts of Binswanger‟s disease,
which is dominated by WML, and the lacunar state, which is dominated by lacunes,
both of which have been shown to be aetiologically and clinically similar (7,55).
20
1.1.7 Pathology of cSVD
In order to understand the pathology of cSVD and its propensity to affect certain brain
regions, it is helpful to consider the relevant normal anatomy of the cerebral
vasculature. Large cerebral arteries lead off the circle of Willis and branch into pial
arteries on the surface of the brain, which give off penetrating arteries that supply the
deep brain structures (20,21,56). These are end-arteries with no collateral supply (21).
The white matter is supplied by distributing vessels that branch off perpendicular to the
penetrating vessel (20). Each distributing vessel irrigates a single cylindrical metabolic
unit (21).
Ageing is associated with particular changes in the cerebral blood vessels, namely
tortuosity and narrowing of the lumen (13). Tortuosity refers to the lengthening,
twisting and coiling of the vessels, which increases the minimum BP threshold
necessary to maintain perfusion (13). The deep white matter is supplied by long
penetrating arteries, which are particularly affected by this tortuosity, and it is thus
vulnerable to hypoperfusion, even at „normal‟ pressures in the elderly (13). Senile
arteriosclerosis begins towards the end of the fourth decade and leads to thickened walls
with progressively stenosed lumens (7). Hypertension can exaggerate the
arteriosclerotic changes (7).
A grading system exists for arteriosclerotic microvascular disease (57):
I. Increased tortuosity and irregularity in small arteries and arterioles
II. Progressive sclerosis, hyalinosis, lipid deposits, loss of vascular wall smooth
muscle and lacunes:
a. Old small infarcts
b. Old haemorrhages
c. EPVS
III. Multiple lacunes, diffuse WML, fibrotic thickening of vessel wall with „onion-
skinning‟
21
The pathological evidence for cSVD is relatively weak. Since Fisher‟s meticulous
dissections in the 1960s, there have been few studies of this type (4). Lacunar strokes
are rarely fatal and so post mortem examination may lag years after the initiating events
leading to infarction (4,58). Furthermore, cSVD requires more detailed dissection of
the perforating arteries than is routinely performed at autopsy (4). In a small series of
pathological studies, involving mostly asymptomatic lacunes often months or years after
the acute event, in a small number of patients (59), Fisher suggested two distinct
vascular pathologies other than embolism as the cause of most lacunes (2,4,33). The
first is intracranial atherosclerosis, which involves occlusion of a penetrating artery by
an atheromatous plaque (2,33). The second he named “segmental arteriolar
disorganisation”, which encompasses lipohyalinosis and fibrinoid necrosis (48).
Lipohyalinosis is characterised by disorganisation and thickening of the arterial wall
with either luminal narrowing and subsequent occlusion leading to lacunar infarction or,
with damage to the wall resulting in microaneurysms and haemorrhage (4,7,33,58).
Fibrinoid necrosis, associated with severe hypertension, refers to narrowing, dilatation
and necrosis (7). These hypertensive arteriopathies are thought to lead to either vessel
occlusion and complete infarction or, to hypoperfusion, ischaemic injury and WML (7).
However, Wardlaw et al. (4) apply different interpretations to Fisher‟s pathological
findings. They hypothesise that cSVD may be the result of dysfunction of the blood-
brain barrier (BBB). They argue that microvascular endothelial damage and
dysfunction, perhaps as a result of hypertension, might allow toxic substances in the
blood to enter the arteriolar wall and subsequently the brain interstitial space. Resultant
thickening and disorganisation of the arteriolar wall might further reduce the barrier to
said toxic substances, and lead to direct damage to the neural tissue (4). CBF might
then be reduced either due to luminal narrowing or secondary to damaged neural tissue
(4). Farrall and Wardlaw (58) recently reviewed, and performed a meta-analysis upon,
papers investigating the BBB in ageing and microvascular disease. BBB permeability
was seen to increase with normal ageing, further in patients with VaD, and also in
patients with a larger WML load (58).
22
There are three limitations to the findings of Farrall and Wardlaw (58) that ought to be
noted. There were no studies of BBB permeability in patients with lacunar stroke, nor
indeed, probable cSVD (other causes of lacunar stroke having been excluded). The
reviewed papers did not all report congruent results, though where significant findings
from the meta-analysis were presented, the majority of papers evidently did agree.
Finally, although the reviewed evidence shows increased BBB permeability, it does not
follow that this is the causative process. Farrall and Wardlaw (58) have tried to address
this last issue, citing evidence derived from an animal model. Sironi et al. (60) describe
a spontaneously hypertensive stroke-prone rat that develops a cerebral microvascular
disease similar to the human cSVD, in which the initial brain abnormality is leakage of
plasma components across the BBB into perivascular tissue.
There is a rapidly growing body of evidence associating BBB dysfunction with cSVD,
reviewed recently by Professor Wardlaw (44). However, it is not yet clear whether
BBB dysfunction is the cause or result of cSVD.
The neurological deficit associated with a LI often progresses over the first 24–36 hours
(1,61). Such progression occurs more often with infarcts situated in the basal ganglia
(40%) and brainstem (50%) compared to those in white matter (9.5%) (61). This
possibly relates to the lower density of glutamatergic neurons in white matter (61).
High glutamate and low gamma-aminobutyric acid (GABA) concentrations in blood
within 24 hours of the onset of stroke symptoms are a strong predictor of subsequent
neurological deterioration (61). On the basis of these two findings, Serena et al. (61)
hypothesised that glutamate, released in high concentrations into the infarct core and
surrounding penumbra, has a neurotoxic effect resulting from a massive influx of
calcium ions initiating cell death. This is compounded by the reduction in the
neuroprotective GABA concentrations, which would normally increase chloride influx
to counterbalance the effects of glutamate during cerebral ischaemia (61). Interestingly,
that same observation of LI progression is cited by Wardlaw et al. (4) as evidence of
leakage of the blood-brain barrier. Another possible explanation for progressive
neurological deterioration following lacunar infarction is inflammation (62,63).
23
1.1.8 Diagnosis of cSVD
As previously discussed, the diagnosis of cSVD is complex. Roman (7) presents the
following brain imaging criteria for subcortical ischaemic vascular dementia, which is
caused by disease of the small arteries and hypoperfusion: extensive WML or multiple
lacunes, in the absence of haemorrhages, cortical and non-lacunar territorial infarcts,
watershed infarcts, normal pressure hydrocephalus, or specific causes of WML.
O‟Sullivan et al. (10) use a combination of radiological findings with a history of
lacunar stroke to ensure a homogeneous study group with cSVD.
1.1.9 Risk factors for cSVD
The risk factors associated with cSVD include increasing age (7), hypertension (7,64),
diabetes mellitus (7), smoking (7), and systemic inflammation (65,66).
Fisher believed lacunes to be particularly associated with hypertension, defined as BP >
140/90mmHg, which he observed in almost 90% of cases (33). Contemporary research
however indicates that hypertension, though important, might not be responsible to the
extent that Fisher postulated. Jackson and Sudlow (59) performed a meta-analysis on
studies of risk factors for lacunar strokes and came to the conclusion that they have a
very similar vascular risk profile to other ischaemic stroke subtypes. Atrial fibrillation
and carotid stenosis were associated with non-lacunar infarcts to a greater extent than
with LI, but no other risk factor, including hypertension and diabetes, was significantly
different (59).
Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and
Leukoencephalopathy (CADASIL) is an inherited form of cSVD that results in
subcortical strokes and VaD (34). A number of other uncommon genetic causes of
cSVD are being recognised (67,68).
24
1.1.10 Management of cSVD
The evidence for the management of cSVD is limited by the lack of distinction of the
underlying aetiology of included patients, for example the failure to separate patients
with cSVD from those with multi-infarct dementia in the analysis of results.
The management of cSVD falls broadly into two areas: prevention and treatment.
cSVD has an insidious progress. Thus it is likely that many patients would not be aware
that they have the disease and so not present themselves to the health services until they
have a stroke, marked cognitive impairment, or changes in gait. This is compounded by
the difficulties in accurate clinical diagnosis which is often not straightforward,
especially without brain imaging.
However, many individuals will already have been receiving preventative management
in the form of modification of cardiovascular risk factors. This should involve lifestyle
advice on smoking cessation, diet and exercise (1). The evidence for drug treatment for
the prevention of cSVD itself is sparse. There are several studies of preventative
treatment of ischaemic stroke in general, however, from which conclusions must be
cautiously derived for cSVD, in lieu of specific evidence. The Systolic Hypertension in
Europe (Syst-Eur) trial found a reduced incidence of dementia with antihypertensive
treatment (69). Similarly, the large Perindopril Protection Against Recurrent Stroke
Study (PROGRESS) is a randomised controlled trial (RCT) that showed that active
treatment to lower BP reduces the risk of dementia and cognitive decline associated
with recurrent stroke (70). The risk of stroke was reduced in both hypertensive and
non-hypertensive patients with a history of previous stroke or transient ischaemic attack
(71). A separate report from the same trial found that reducing BP in those with a
history of stroke did not increase the risk of silent brain infarcts or atrophy (72).
Although evidently beneficial in the majority of patients with vascular risk factors,
concern remains that reducing BP might have a negative effect in patients with cSVD
(21). Furthermore, a Cochrane review came to the conclusion that lowering BP in
hypertensive patients without prior cerebrovascular disease had no benefit in preventing
cognitive decline, though many of the controls ultimately received antihypertensive
25
treatment too (73). One of the reviewed studies did, however, provide data on distinct
stroke sub-types and showed a reduced incidence of first-ever lacunar strokes (74).
Interestingly, a small study of 12 patients with a radiologically-confirmed lacunar stroke
found that VMR was significantly increased after treatment with Perindopril for two
weeks (75), which might suggest a potential benefit of antihypertensive therapy in
cSVD. Cerebral VMR is the process of provision of adequate cerebral perfusion in
response to a functional demand, by vasodilatation of cerebral arterioles leading to a
reduction in the resistance of the vascular bed and consequent increase in local CBF
(76). Impaired VMR indicates the reduced capability to carry out this process in
response to a vasodilatory stimulus, such as nitric oxide (NO) (76).
Despite its widespread use, a Cochrane review found no RCT providing evidence for
the use of aspirin in VaD (77). The CHS found that the regular use of aspirin was
associated with an increased risk of ischaemic stroke in women and haemorrhagic
stroke in both sexes (78). Two limitations to this observational study ought to be noted.
Firstly, aspirin users did not necessarily have cardiovascular risk factors and, secondly,
the authors suggest that the reasons for aspirin use might have confounded the results.
The evidence for the use of lipid lowering therapy is not convincing. Reports from the
CHS suggest that statin use slightly reduces the risk of cognitive decline (79) but not
dementia (80). The Pravastatin in elderly individuals at risk of vascular disease
(PROSPER) RCT found no improvement in stroke risk or cognitive function with statin
use (81). In a study of 94 patients with a recent lacunar stroke, no improvements were
seen in cerebral VMR or endothelial dysfunction of the carotid and brachial arteries
after three months (82). Conversely, other large studies have found a reduction in
dementia (83) and stroke (84) risk with statin use. Statins clearly have a benefit in other
cardiovascular disorders, which share common risk factors with cSVD (84).
Patients with cSVD may come to the attention of health services with a lacunar stroke
or cognitive dysfunction/ dementia. In the acute setting thrombolysis should be
considered in patients presenting with lacunar stroke if indicated clinically (85).
26
Aspirin and dipyridamole are also recommended (86). One study has also found benefit
in functional outcome, though not mortality, with the administration of intravenous
magnesium following an acute lacunar stroke (87). Large RCT are necessary to confirm
this benefit. Subsequent management involves multidisciplinary rehabilitation (6) and
secondary prevention as discussed above. A patient with an identified cause for the LI,
for example atrial fibrillation or carotid stenosis, ought to be treated appropriately.
Cochrane reviews have found no consistent evidence in VaD in support of cognitive
rehabilitation and training (88), memantine (89) or galantamine (90). However,
donepezil has been shown to improve cognitive function, clinical global impression
(which reflects symptom severity and treatment response) and activities of daily living
in patients with vascular cognitive impairment (91). Nimodipine has also shown some
benefit in the short-term in cognitive function and clinical global impression, in patients
with dementia due to cerebrovascular disease (92). A small study investigating the use
of rivastigmine in patients with probable VaD found it to be associated with
improvements in attention, executive function and activities of daily living (93).
1.1.11 Prognosis of cSVD
LI were previously considered to have a very benign course, with a low mortality rate, a
good prognosis for recovery, and a low recurrence rate (1,61). This belief was based on
short-term follow-up studies. Within the first year there is a much lower mortality rate,
ranging from 2%–15% (2), as well as less disability, than for other ischaemic stroke
subtypes, presumed to be due to the smaller infarct size of a LI (1,2). A rapid recovery,
resulting in a reduced risk of secondary complications, as well as fewer cardiac co-
morbidities in lacunar patients, are also thought to contribute to the low 1-year mortality
rate (2).
However, recent studies with a longer follow-up suggest that LI do not have such a
benign course (2,94,95). The 5-year mortality rate for LI ranges from 14%–38% (96-
99). The large difference can be explained by the average age of participants, with
studies recruiting younger patients portraying more favourable mortality statistics.
27
However, by 10 years after the stroke event between 54% (97) and 60% (98) of patients
had died.
The risk of a recurrent stroke increases from 7.7% at 1 year, to 22.4% at 4-5 years (2).
Factors that predict a subsequent stroke include vascular risk factors, the severity of
cSVD, a cardioembolic source, diabetes with concomitant hypertension, and diabetes
with silent lacunes (2). The progression of apparently asymptomatic cSVD is much
more likely than a new stroke, and is thought to insidiously impair brain function (1,2).
Ten years on from a symptomatic lacunar stroke, less than a third of patients are still
alive and free from recurrent stroke (2). Of the survivors, many are disabled or
cognitively impaired (2). There is good evidence then, that although the initial
prognosis for a LI is comparatively good, in the long term there is an increased risk of
death, stroke recurrence and cognitive impairment.
1.2 Cognitive disturbances in cSVD
O‟Sullivan et al. (9) have shown that disturbance in cognition can be used as an
objective clinical measure of discriminating patients with cSVD and suggest its use as
an outcome measure in clinical trials. The most commonly recognised feature of
impaired cognition associated with cSVD is executive dysfunction (5,9). Attention and
the speed of information processing are also impaired, whilst memory is relatively
spared (5,52,53). In contrast Alzheimer‟s disease, the predominant cause of dementia,
has a cognitive profile dominated by problems with memory (18).
In VaD pathological changes such as WML, impaired VMR and altered resting CBF are
most evident in the frontal lobes (49). Executive function is a component of the frontal
cortical areas, in particular the dorsolateral prefrontal cortex (52,100). It encompasses
higher cognitive processes, such as planning, attentional set shifting, the selection of
appropriate responses and working memory (9). Working memory refers to the ability
to temporarily store and manipulate information required for the task in hand, and is
28
separate from episodic or long term memory (9). Regardless of other deficits, executive
dysfunction results in morbidity, dependence, and an increased need for care (50).
As to be expected from their shared aetiology (cSVD), LI and vascular cognitive
impairment are closely linked (9,54). Cognition declines in the period immediately
following a LI, and there is an increased risk of cognitive decline leading to dementia,
in the long term (54). Lacunes are thought to contribute to neurological damage which
results in vascular cognitive impairment (54). Both pathologically and radiologically,
lacunes, as well as WML (LA), are observed in patients with VaD (52,53). Discrete
lacunes were reported in 93% of patients with VaD (18). Aharon-Peretz et al. (101)
found 83% of those with lacunes and LA to have cognitive impairment. However, they
also found the progression to frank dementia to be slow. This finding might be biased
by the use of dementia screening tools that are based upon an Alzheimer-type dementia
(9). The cognitive profile associated with VaD is quite distinct from that of
Alzheimer‟s disease. Thus an individual with cSVD may have accumulated very severe
cognitive impairment before being recognised as having dementia, when using such
screening tools.
Lacunes and VaD tend to be diseases of older people. Changes have been noted in the
cerebral vasculature of non-diseased older people. Schroeter et al. (102) noted a drop in
the haemodynamic response in the frontal lobes during a colour-word matching Stroop
Interference Task („Stroop Task‟) (103), in elderly versus young individuals.
Histological studies have shown alterations in the organisation of cerebral small vessels
associated with increasing age (104).
Older age is associated with a progressive decline in cognitive function (105).
O'Sullivan et al. (105) postulate that this might be a consequence of “structural
disconnection” due to disrupted white matter tracts, rather than dysfunctioning grey
matter regions, as evidenced by a post-mortem study (106) and by diffusion tensor
imaging (DTI), an MR technique used in their study. They propose that NO might lead
to the activation of microglia, which are involved in myelin phagocytosis, resulting in
29
the reduced number of myelinated white matter fibres apparent in normal ageing. The
hypothesis of structural disconnection has also been applied to cognitive impairment in
cSVD (10). cSVD might expedite these age-related changes (49).
The presence of cSVD-associated lesions on MRI predicts cognitive dysfunction (107).
However, structural markers of cSVD, such as T2 lesion volume (5,10), LA (51) or the
number of lacunes (5,108,109), do not actually correlate well with clinical or cognitive
impairment (5,10,109-113). They may represent the end-points of tissue damage (113).
Alternative MR techniques promise to offer a better measure of the problems associated
with cSVD. This study aims to identify better and more quantitative markers of cSVD.
In the ensuing chapters, the literature surrounding the use of imaging markers in cSVD
shall be systematically reviewed and basic MRI theory touched upon, before a
discussion of the functional MRI techniques used in this study. Chapters 4 and 5 shall
detail the methods and results and lead on to a discussion of the findings and their
implications.
30
Chapter 2: Review of imaging markers of cSVD
VaD is a heterogeneous disorder that is not consistently classified. Lopez et al. (114)
applied three different sets of VaD classification criteria to a set of patients from the
CHS, none of which selected the same group of patients. Like VaD, LI also have
multiple causes. However, cSVD is the predominant cause of both and is a much more
homogeneous disorder.
Small cerebral vessels 20–40µm in diameter cannot be directly visualised in vivo (8).
cSVD can be identified by meticulous and laborious pathological dissection, as
demonstrated by Fisher (33). However this is neither practical, nor does it allow pre-
mortem diagnosis. The advent of brain imaging with computed tomography (CT) and
MRI has brought a renewed interest to this disease process. These imaging modalities
are able to demonstrate the end-organ damage of cSVD. Two main lesions are visible,
LA and lacunes. There is much debate as to whether LA and lacunes correlate with
cSVD progression. Indeed LA (14) and lacunes (96) are commonly seen in apparently
healthy elderly subjects.
By charting the progression of cognitive decline it might be possible to delineate the
progression of cSVD. Cognitive decline might also act as a suitable outcome measure
for clinical trials. However, repeated cognitive tests are subject to a learning effect, are
influenced by mood and motivation, and are time consuming (10). Such tests are also
influenced by pre-morbid intelligence (115). Imaging markers may offer a more
objective and consistent measure of cSVD severity. Various markers have been
suggested, including both structural changes such as LA, and functional changes such as
CBF. However, there is considerable discrepancy in the available evidence. Thus, a
literature review was performed with the aim of attempting to clarify the evidence for
and against each imaging marker. A meta-analysis was not performed because of the
inconsistency in the type of outcome data, such that aggregation of results would not be
possible.
31
The databases Medline (via Ovid), Web of Knowledge, PsycINFO, SCOPUS, Science
Direct, AMED- Allied and Complementary Medicine, CINAHL, the Cochrane Library,
ebrary, TRIP database, Bandolier and Intute: medicine, were searched for relevant
clinical trials. Search terms included „cerebral small vessel disease‟, „cerebral
microangiopathy‟, „cognitive dysfunction‟, „imaging marker‟ and „NOT CADASIL‟.
Selected papers were limited to English clinical trials published before May 2009.
Further studies were identified from the reference lists of papers highlighted.
Only trials that selected a patient group with probable cSVD were included. The trials
also needed to have a neuropsychological test as an outcome measure to allow
comparison of results between papers.
Nineteen studies were identified that had reasonably well-defined patients with cSVD.
The main characteristics of the included studies are presented in appendix A. The
interventions investigated included standard structural MRI, magnetic resonance
spectroscopy (MRS), DTI, diffusion weighted imaging (DWI), CT, single photon
emission computed tomography (SPECT), positron emission tomography (PET),
functional near-infrared spectroscopy (fNIRS) and post mortem examination.
Surprisingly, no studies of microbleeds as an imaging marker were identified that met
the entry criteria.
2.1 WML
WML and their correlation with cognitive dysfunction have already been extensively
reviewed (13,15,20). There is a huge body of evidence both for and against any such
correlation. However, the majority of the identified studies with strict criteria for the
selection of patients with likely cSVD found no correlation between LA and cognitive
dysfunction (5,10,49,113).
32
Appelros et al. (96) found MMSE score at both baseline and 5 years to correlate with
WML score at both baseline and 5 years. However, the decline in MMSE score was not
associated with an increased WML score, nor indeed with the WML score at baseline or
5 years. These results suggest that cSVD is associated with both LA and with cognitive
dysfunction, as it is known to be, but that LA and cognitive dysfunction are not
themselves directly linked. A change in one does not directly induce a change in the
other.
Wen et al. (116) found LA to be independently associated with executive dysfunction,
as measured by the Mattis Dementia Rating Scale-Initiation/ Perseveration sub-scale
(MDRS-I/P). However, when the quartile with the most severe lesions was removed
from analysis, the association disappeared. The authors suggest a „threshold
hypothesis‟, whereby WML only exert an effect upon executive function when they are
sufficiently severe (116). This relationship does not lend itself to a suitable surrogate
marker of clinical severity, where a linear correlation would be best.
Finally, Mok et al. (117) found WML volume to predict MDRS-I/P score, amongst 45
patients with LA and a lacunar stroke. There was no healthy control group. Patients
had had a lacunar stroke, though whether this was a clinical event or a LI is not clear.
Although 111 imaging variables were compared to MMSE and MDRS-I/P tests, there
was no evidence of any statistical correction for the multiple comparisons.
2.2 Lacunes
Similarly to WML, the overwhelming majority of studies of well-selected patients with
cSVD found no evidence for an independent correlation between cognitive impairment
and site, number or volume of lacunes (5,49,113,116,118-122). However, three studies
have found a link between the presence of lacunes in either the basal ganglia (96,123) or
thalamus (121,123) and cognitive impairment. Only one of the three papers correlates
the site-specific lacune with a measure of executive function (MDRS-I/P) (121).
33
It is worth noting that a „strategic infarct dementia‟ is a well-recorded entity, resulting
as the name suggests, from an infarct in an area of the brain important for a particular
function (7,124). Although a strategic infarct might conceivably account for the
observed cognitive impairment, it is a rare event, with basal ganglia and thalamic
associated strategic infarct dementias accounting for 4.6% of VaD cases (125), and as
such might be considered unlikely to be the sole reason.
Appelros et al. (96) found patients with a high basal ganglia score to exhibit lower
MMSE scores, but there was no correlation on longitudinal analysis. The study by Gold
et al. (123) was an autopsy-based study of 72 patients. However, 29 of those patients
had cortical infarcts and were included in the analysis. Although 16 of the 72 patients
had likely cSVD, these were not analysed separately, and thus the findings of this paper
cannot be assumed to be related to patients with likely cSVD. Mok et al. (121) found
the presence of thalamic lacunes to be independently associated with both the MMSE
and MDRS-I/P, in an uncontrolled sample. Likewise with the relationship between LA
and cognitive dysfunction, it is not surprising to find the presence of lacunes to be
associated with cognitive dysfunction, as both are associated with underlying cSVD.
However, the presence of lacunes provides just two groups, those with and those
without. It provides nominal data, which is a poor marker of changing severity, when
compared to the ordinal data of a scale such as the MDRS-I/P.
In summary, there is no convincing evidence for an association between cognitive
impairment and the number or volume of lacunes. There is limited evidence to suggest
that the site of a lesion might affect the severity of impairment, but this could not be
used to chart the progression of disease in an individual.
2.3 Atrophy
The evidence for an association between brain atrophy and cognitive impairment is
compelling (5,10,96,113,117,121,126). In two cases the link disappeared on
multivariate analysis (10,121) and one paper did not find a significant association, but
34
did report a trend towards such (p > 0.05) (120). Cerebral atrophy correlates well with
neuropsychological impairment, but it is not specific for cSVD (127) and might
represent a point at which intervention is too late to effect much change.
2.4 DTI measures
Three studies have investigated DTI in patients with cSVD that fit the entry criteria, and
all showed very promising results, though all are from the same study institution
(5,10,128). Exemplary selection criteria and neuropsychological tests are used. DTI is
used to detect white matter tract integrity (10). The studies lend evidence to the
“disconnection syndrome” theory, which describes cognitive impairment in cSVD as
resulting from disruption of white matter tracts that form cortical-subcortical and
cortical-cortical connections (10).
Figure 4: Axial view of T1-weighted images from two individuals, one with (right) and
one without (left) evidence of cerebral atrophy.
35
Larger studies by other institutions are necessary to confirm the correlation between
DTI measures and cognitive impairment. It is not clear whether there is an overlap of
study participants amongst the three studies.
2.5 MRS
MRS is used to non-invasively estimate brain metabolite concentrations (129). Van
Zandvoort et al. (130) found relative concentrations of N-acetylaspartate (NAA) (as a
ratio of Creatine (Cr)) to be reduced in the normal-appearing white matter of the
centrum semiovale. Hund-Georgiadis et al. (119) found relative concentrations of NAA
(as a ratio of both Cr and Choline (Cho)) to be reduced in the parietal intact white
matter. This reduction in relative NAA concentration was correlated in both cases with
impaired cognitive functioning. Reduced NAA suggests axonal damage or neuronal
loss (119). In contrast, Nitkunan et al. (129) found no correlation between any
metabolite measured in the white matter of the centrum semiovale and cognitive
impairment, although absolute NAA concentration was reduced in patients within
regions of LA. This is consistent with tissue damage occurring in the visible WML
rather than in normal-appearing white matter (129). However, it does not appear to be
consistent with DTI findings of white matter tract damage occurring in the normal
appearing white matter (128).
2.6 Cerebral perfusion and glucose metabolism
Yang et al. (131) showed a reduction in CBF in several regions in patients with
subcortical VaD versus controls, though this did not appear to correlate with dementia
severity as gauged by the Clinical Dementia Rating (CDR) scale. The sample size was
small and though a thorough neuropsychological battery was said to have been done, the
results were not reported.
Sabri et al. (113) found that both reduced regional CBF and glucose metabolism were
associated with neuropsychological deficits. The included patients were well selected
36
for cSVD, the sample size was comparatively large (57 patients), and a large battery of
neuropsychological tests were used. Further studies in patients well selected for cSVD
are needed to confirm these results.
2.7 Amplitude of vascular response to Stroop Task
A recent fNIRS study by Schroeter et al. (49) found cognitive dysfunction to be
associated with a reduced and delayed haemodynamic response in patients with cSVD.
Neuronal activation stimulated by a given task induces a haemodynamic response,
characterised by a relative increase in blood oxygenation, which can be measured (49).
This exciting development might be used to measure earlier changes in the cSVD
process, before end-organ damage such as WML are apparent, increasing the time for
intervention. However, the study only included 12 patients, 5 of whom had evidence of
macroangiopathy.
Functional MRI (fMRI) using the blood-oxygen-level-dependent (BOLD) signal also
measures the haemodynamic response. As yet, there are no studies that attempt to
correlate fMRI findings with cognitive impairment in patients with cSVD. However,
one study has noted a reduced rate of rise and maximum rise in BOLD signal in patients
with a first-ever lacunar stroke, in response to a finger-tapping task (132).
2.8 Discussion
The diagnosis of cSVD is dependent upon the patient having a lacunar stroke, a lacune,
WML, or a combination of the three. As the structural markers of cSVD account for
part of the definition, it is inevitable that these markers will correlate with the disease
group, when compared with healthy controls without such pathology. cSVD is
associated with cognitive impairment, so a patient group is likely to have a greater level
of such impairment when compared with controls. However, it does not follow that
these pathological lesions closely correlate with cognitive impairment, merely that they
37
share a common aetiology. Thus, an apparent correlation between an imaging marker
and cSVD might be due to selection bias rather than reflect a true correlation.
It is important to select a well-defined group of patients with cSVD. It allows a
statement of any differences noted to be made with confidence, for example that a
particular marker is a good surrogate marker for cognitive decline in cSVD. Having
made this first step with relative certainty, the knowledge gained can be used to
investigate more diverse clinical groups, such as the elderly or patients with VaD.
There are several reasons for the discrepancies in the evidence and confusion
surrounding imaging markers for cSVD, summarised in Box 2.
In conclusion, functional markers and DTI appear to show a better correlation with
cognitive dysfunction than the traditional structural markers of cSVD, LI and WML.
However, the evidence for these newer markers needs to be corroborated by larger
studies with more uniform tests of executive dysfunction and strict criteria for cSVD.
38
Box 2: Limitations of research into potential imaging markers for cSVD.
1) Inconsistencies and uncertainties in terminology, definitions and criteria
a) Inconsistent use of terminology in the literature for cSVD itself. Other terms
used for cSVD include CMA and subcortical ischaemic VaD. However, these
are not always synonymous.
b) Inconsistent terminology for lacunes, LI and lacunar strokes (24).
c) Variable definitions/ criteria are used for VaD and cognitive impairment. This
might result in different groups of patients being selected and compared (94).
d) Variable definitions are used for observed lesions, e.g. a lacune is defined as
<15mm in some and <20mm in other papers.
e) Confusion over what constitutes cSVD, for example many authors consider
WML to be synonymous with cSVD.
f) Lack of pathological studies.
g) Other dementing illnesses, particularly Alzheimer‟s disease, may not be
sufficiently excluded.
2) Variable measures
a) Variable scales are used by different researchers. Some use qualitative
methods, others semi-quantitative for measuring WML.
b) Variable measures of cognitive function. E.g. some trials don‟t use tests of
executive function, but tests such as MMSE as an outcome measure.
3) Design limitations
a) Small sample size.
b) Lack of a matched control group.
c) Different populations are used, e.g. general population, dementia, VaD, cSVD.
d) Often, the independent contributions of different measures are not assessed,
i.e. by multivariate analysis (114).
4) Variable time from onset of LI to study participation.
39
Chapter 3: MRI
3.1 MRI Theory
It is beyond the scope of this thesis to provide detailed information on MRI theory, but
further information can be found in many of the standard textbooks (133,134).
However, a brief synopsis of basic MRI theory is provided. MRI is based upon the
behaviour of the nuclei of certain atoms, particularly hydrogen atoms (henceforth
termed protons), when exposed to an external magnetic field. Protons can be thought of
as miniature bar magnets. A spinning proton is a spinning positive charge, which is an
electrical current, which produces a magnetic field. As such, protons have intrinsic
magnetisation. Thus, when placed in an external magnetic field (as provided in an MRI
scanner), protons will interact with this field. Protons spin about their axis as described,
but also precess, that is they also turn in a “cone-shaped fashion” (135) around the axis
of an applied magnetic field (Figure 5).
a b
Figure 5: Diagram to show the movements of a proton (shown as a thin arrow): spin
about its axis (a); and precession (b) with the proton shown at three different time-
points. The vertical block arrow represents the direction of the applied external
magnetic field.
40
MRI scanners contain a large superconducting magnet. Protons placed within this
magnet align themselves along its magnetic field, resulting in a longitudinal
magnetisation along the z-axis. A transmitted radio-frequency (RF) pulse transfers
energy to protons that are precessing at the same frequency and causes the protons to be
in phase, i.e. to all be in the same position around the precessing cone (Figure 6). The
protons also begin to precess (or „tip‟) around the axis of the applied RF pulse (the RF
pulse is in effect another applied magnetic field). This causes a reduction in the
longitudinal magnetisation and the induction of a transverse magnetisation along a
perpendicular axis. When the RF pulse is switched off the protons return to their pre-
RF pulse state. There is longitudinal relaxation as the energy provided by the RF pulse
is transferred from the protons to their surroundings. This is known as T1 relaxation.
T1 is the time taken for a tissue to recover 63% of its longitudinal magnetisation.
Following the RF pulse, there is also transverse, or T2, relaxation, as the precessing
protons lose phase coherence. T2 is the time taken for 63% of the induced transverse
magnetisation to have disappeared.
a b
Figure 6: Diagram showing protons (each represented by a single arrow) in (a) and
out of (b) phase.
41
Different tissues have different characteristics. Water has a long T1 and T2. In
comparison to water, fat has a short T1 and T2. By varying scanning parameters, such
as the time to RF pulse repetition (TR) or the time to echo (TE), which is the time
between RF signal transmission and acquisition, different tissue characteristics can be
highlighted. A short TR and TE result in T1-weighted images that highlight
longitudinal relaxation time. Fluid appears dark on the images. In contrast, fluid
appears bright on T2-weighted images, formed using a long TR and TE. FLAIR (Fluid
attenuated inversion recovery) sequences enhance the visibility of lesions by
suppressing the signal from background CSF (136). CSF signal and the contrast
between white and grey matter are reduced (136). This sequence is particularly useful
for separating WML from CSF-like lesions (137).
Other tissue properties can be exploited to provide different images: the flow of blood
can be measured using MR-angiography (MRA) or arterial spin labelling (ASL); the
BOLD signal utilises the deoxyhaemoglobin content of blood; and signal loss due to the
presence of tissue with high susceptibility such as veins can be detected by
susceptibility-weighted imaging (SWI). ASL and BOLD shall be discussed further in
the next section.
Figure 7: Examples of, from left to right, T1-, T2-, and FLAIR-weighted images.
42
3.2 Functional MRI
Two main functional techniques will be focused upon: the BOLD signal and CBF
measurement using ASL. A reasonable body of literature has been compiled on each
individually. However, they are both techniques still in their comparative infancy in
terms of development, as borne out by their limited current application clinically. Used
concurrently, the BOLD and CBF measurement by ASL might enable a more accurate
measure of neural activity than either alone.
3.2.1 CBF measurement using ASL
In healthy individuals CBF changes with varying neural activity levels (138). Local
CBF changes induced specifically by a neural task can be used as a marker of changing
neural activity (138). There are many established imaging techniques for measuring
CBF, reviewed recently by Wintermark et al. (139): SPECT, PET, and MRI with
contrast, such as gadolinium, are three well known examples. ASL is a developing non-
invasive functional MR technique that can provide a quantitative measure of resting and
dynamic CBF (140-149). All the techniques are based upon a similar principle in that
they measure the concentration of a contrast material that is delivered in the blood and
cleared from the tissue of interest (138,150).
ASL uses water as an endogenous, non-radioactive tracer (138). A magnetic label is
applied to the water molecules of flowing blood (138). The rate at which these labelled
molecules are delivered to an imaging slice is determined by local CBF (140). The
magnetic label decays within approximately 1-2 seconds (150). The short turnover time
allows rapid repetition and hence dynamic measurements of CBF (150). Repeated
scans are also possible with ASL because there is no risk of radiation over-exposure, as
there is with radioactive tracer techniques (138).
To acquire the labelled image a magnetic label is applied to water molecules by
inverting or saturating the longitudinal (Z-axis) component of the magnetisation of the
hydrogen nuclei (or „spins‟) (138). These labelled molecules enter the capillaries,
carrying reduced longitudinal magnetisation, so decreasing the average magnetisation in
43
the voxel (a three-dimensional measure of volume). This manifests as a reduction in the
T1-weighted signal compared to the control images, in which the blood is already fully
relaxed (138,151). The more labelled blood that enters the image slice, the greater the
signal changes versus the control image (138). After a delay TI (the inversion time)
following the labelling of arterial blood to allow for it to have entered the region of
interest, an image is obtained (138). A control image, with no labelled blood, is taken
of the same region to account for the contribution of the static image (138).
Labelled and unlabelled images are alternated (150). The difference of these paired
images is used to measure CBF (138). A series of paired images are usually taken, and
averaged, to account for the inherently low signal-to-noise ratio (SNR) and to produce a
map of CBF (152).
One reason for the inherently low SNR is the necessary short inversion time
(approximately 1 second), due to the short lifespan of the label T1, allowed for labelled
blood to arrive in the region of interest (138). In this time, about 1ml of blood is
supplied for every 100ml of tissue (138). Thus the tagged blood only makes up roughly
1% of the total MR signal from the tissue, resulting in a low SNR (138).
There are different types of ASL, which label the arterial blood based on location
(Pulsed ASL), velocity (Velocity-selective ASL) or both (Continuous ASL) (138).
ASL has been validated in healthy subjects using dynamic susceptibility contrast (153)
and with 15
O-PET at rest (154) and with activation (155). It has also been validated for
measuring CBF in patients with cerebrovascular disease (156-158).
44
3.2.2 BOLD signal
The BOLD signal is an indirect marker of neural activity (159). As the name suggests,
it is a signal produced according to differing levels of oxygen in the blood. In fact it
depends upon the deoxyhaemoglobin concentration in venous blood (160).
Oxyhaemoglobin is diamagnetic, whereas deoxyhaemoglobin is slightly paramagnetic
(159,161). This property of deoxyhaemoglobin means that it induces microscopic
magnetic field inhomogeneities around it, which can be picked up using MRI as a drop
in the T2* relaxation time, and hence a reduction in the MRI signal (141,161).
The changes in deoxyhaemoglobin concentration result from the „haemodynamic
response‟ to neural activity (141). The haemodynamic response has three main
components: changes in CBF, cerebral blood volume (CBV), and the metabolic rate of
oxygen consumption (CMRO2) (141,159). Increased neural activity is associated with a
seemingly paradoxical drop in deoxyhaemoglobin, which in turn leads to reduced
magnetic field distortion and hence to an increased BOLD signal (141,159).
Figure 8: Diagram of BOLD signal, for a stimulus of duration 30 seconds.
45
Matthews and Jezzard (161) succinctly describe the relationship between neural activity
and a measurable BOLD signal, expanded upon here. Neural activity uses energy,
which is thought to be mostly needed for postsynaptic neuronal depolarisation. This
results in the uptake of oxygen and hence the increase in deoxyhaemoglobin, which is
assumed to cause a contentious early dip in the BOLD signal, contentious as the dip is
not always seen (162). Blood flow to the area increases. There has been great debate as
to whether the drop in oxygen or glucose drives the increased blood flow. Current
opinion holds that it is neither, but rather a result of local signalling (142,161).
Nevertheless, the increased blood flow provides the substrates needed for energy
metabolism. However, the blood flow increases to a greater extent than is needed
merely for substrate provision. Due to the metabolic demands of neural activity, more
oxygen is taken from the blood than in the resting state, but the proportion of oxygen
taken from each unit of blood decreases because of the increased flow. In other words,
there is a drop in the oxygen extraction fraction, which manifests as a smaller
concentration of deoxyhaemoglobin relative to oxyhaemoglobin in the area of neuronal
activation. This, in turn, is detected as an increase in T2* relaxation time and hence the
production of an increased BOLD signal.
Figure 6: Summary flow chart of events leading to BOLD signal.
Figure 9: Summary flow chart of events leading to BOLD signal.
Stroop task
Neural activity
↑ Oxygen uptake
↑↑CBF
↑Oxy-Hb:Deoxy-Hb
↑BOLD
signal
↑ Glucose uptake
↓ OEF
Initial ↑ in Deoxy-Hb
46
The BOLD signal is complex. It „measures‟, albeit indirectly, the averaged neural
activity of an area and thus could represent a small change in a lot of neurons or a large
change in a few (162). Sub-threshold activity, excitation, inhibition, or modulatory
inputs might all contribute to the BOLD signal (76,162). Logothetis et al. (163)
maintain that the BOLD signal reflects the “input and intracortical processing” of an
area. Regardless of what exactly constitutes the BOLD signal, the magnitude of the
BOLD response can be altered by MRI specific parameters, for example magnetic field
strength (164).
Although a complex signal, it is still a useful marker of neural activity, with good
agreement observed between amplitudes of fMRI and electroencephalography (EEG)
(161,165,166). Logothetis et al. (163) show a proportional link between the BOLD
signal and the average level of neuronal activation.
BOLD signal is a measure of the haemodynamic response, rather than a direct measure
of neural activity. The use of the BOLD signal as an indirect marker of neural activity
is dependent upon neurovascular coupling (NVC), the concept that a predictable
haemodynamic response follows increased neural activity (56). In clinical studies,
particularly those investigating cerebrovascular disease, this response might be altered
(56). Thus any difference seen in the BOLD amplitude seen for example between a
normal and a clinical group might not reflect neural activity, but impaired NVC. Apart
from cerebrovascular disease such as stroke, increasing age and medication may affect
NVC. By altering the haemodynamic response, causing changes in CBF, baseline
CBV, or the baseline oxygen extraction fraction, there might be changes in BOLD
signal measurements that are not due to neural activity (167).
Neural activity can be induced by giving a subject a task. By matching the time course
of the BOLD signal with that of the task, areas that show changes in activity (changes in
BOLD signal) as a result of the function can be observed (161). However, BOLD fMRI
can only detect relative (i.e. qualitative) signal intensity changes (161). It is thus mainly
used as a mapping tool (159). ASL, on the other hand, can provide quantitative
47
measures of CBF (161). ASL can be used to simultaneously measure BOLD and CBF
(159), and the BOLD signal can then be calibrated against the absolute CBF
measurements (161). With the inclusion of an additional isometabolic calibration scan,
some of the neural and vascular components of the BOLD response can be separated,
allowing the estimation of the change in CMRO2 (138), which is thought to be closely
linked to neural activity (167). In this way an insight into the relative vascular,
measured by CBF change, and neural, by CMRO2 change, contributions might be
achieved (141). Thus, Buxton et al. (159) suggest a „measure‟ of neurovascular
coupling („n‟) can be made by the ratio ∆CBF/∆CMRO2.
3.2.3 Oxygen calibration
Carbon dioxide (CO2) has been used to calibrate the BOLD signal (168,169). The
patient inspires CO2, resulting in an increase in CBF with little change in CMRO2 (159).
However, there are drawbacks to this widely used method (170). It may in fact affect
CMRO2 (171), it induces breathlessness (172), it is intolerable to infirm or distressed
patients (170), and it has a regionally varying effect on CBF (172). In light of these
problems, Chiarelli et al. (170) have suggested an alternative calibration method for
quantitative BOLD fMRI, using hyperoxia. Advantages of the hyperoxia over the
hypercapnia technique are that hyperoxia is much better tolerated, and for longer
periods; oxygen is cheap and available; the CMRO2 appears to remain constant in adults
in response to hyperoxia; and the hyperoxia technique has lower inter-subject and inter-
session variability (170). However, caution is needed when increasing oxygen delivery
to patients with COPD (chronic obstructive pulmonary disease) (170).
48
BOLD advantages:
High contrast: noise ratio (138).
Better temporal resolution (2-4 seconds) than ASL, which relies on pair-wise
subtraction of temporally adjacent images (173). However, poor temporal
resolution compared to that of MEG or EEG (milliseconds) (174).
BOLD limitations:
BOLD signal change can vary substantially between subjects and across
sessions (175,176) .
Complex signal, encompassing CBF, CBV, CMRO2, and magnetic field
strength. Any variation in these due to age or disease can confound
interpretation of BOLD changes (138).
Exact location of signal change may not define exactly where the presynaptic
neural activity is found (163).
Advantages of ASL over BOLD:
Smaller inter-subject variability (177-179).
Reduced susceptibility artefact in regions of high static inhomogeneity
(179,180).
More specific functional localisation (138,181,182). Tjandra et al. (179), in a
study of 6 patients, suggest that CBF may localise activation more accurately
as it is more sensitive to changes in the capillary bed local to the activated
neuronal population than BOLD.
Can provide quantitative measures of baseline and functional changes in CBF
(138).
Limitations of ASL
A fewer number of slices (173).
Low temporal resolution (173).
Relatively low SNR (173).
Less sensitive than BOLD (66,72).
Box 3: Advantages and disadvantages of BOLD and ASL techniques for fMRI.
49
3.3 Aims
cSVD, leading to stroke and cognitive impairment, is an important pathology, more-so
in light of an ageing population. Early and accurate diagnosis might allow development
of prevention and treatment modalities. Structural MRI markers have proved
ineffective in delineating cSVD associated with cognitive decline and might offer little
more than endpoint markers of disease. However, alternative developing MR
techniques offer a promising approach to diagnosis. This study shall look at two such
techniques, BOLD and CBF measurement by ASL. The primary aim is to determine
whether the neurovascular coupling constant, confusingly denoted „n‟, is reduced in
patients with cSVD when compared to healthy controls. Thus the null hypothesis is that
there is no difference in „n‟ between cSVD and control groups. If „n‟ does prove to be
reduced in the cSVD group, a secondary hypothesis to be tested is whether „n‟
correlates with tests of cognitive, and especially executive, function.
50
Chapter four: Methods
Approval was obtained from the Cheshire Local Research ethics committee. All
participants gave written informed consent. The study was adopted by the Stroke
Research Network (SRN), as the Liverpool Lacunar Stroke Study (LILACS).
4.1 Subjects
The „lacunar hypothesis‟ suggests that cSVD, the pathology of which being in situ
microatheroma or lipohyalinosis but not embolism (183), leads to LI that cause
particular syndromes, namely those detailed in Box 1 (184). This hypothesis indicates
that LI are important markers of cSVD (2). However, the presence of a lacunar
syndrome in itself does not necessarily indicate underlying cSVD, as the aetiology in up
to a third of these patients is embolism or, more rarely, a cause such as arterial
dissection (1,33,184). Yet, for the purposes of clinical research, careful selection
criteria can identify patients with a high likelihood of cSVD, particularly by ensuring
radiological confirmation of a clinically relevant lacunar stroke and exclusion of other
stroke aetiologies. In line with this, criteria based on those used by Markus et al. (185)
in their online protocol for creating a database of cSVD patients, were formulated.
Patients within one year of diagnosis with a lacunar syndrome were recruited from The
Walton Centre for Neurology and Neurosurgery, Royal Liverpool University Hospital,
Broadgreen Hospital, Countess of Chester Hospital, Royal Preston Hospital and
University Hospital Aintree (UHA). Dr Emsley, a consultant neurologist1, confirmed a
diagnosis of likely cSVD based upon the notes of potential participants. Those
considered appropriate were invited to enrol in the study. Age-matched healthy controls
with no history of neurological or vascular disease were recruited from the local
university population and previous healthy volunteers at Magnetic Resonance and
Image Analysis Research Centre (MARIARC).
1 Consultant appointment made during the course of the study.
51
Exclusion criteria Reason
Known significant (≥ 50%) extra- or intra-
cranial carotid or cerebral vessel stenosis
To exclude subcortical infarcts that were
not likely to have been caused by cSVD.
Cardio-embolic source of stroke
Cortical infarct
Subcortical infarct > 1.5cm
Any other specific cause of stroke
Significant pre-stroke cognitive impairment
(sufficient to interfere with activities of
daily living)
To ensure participants were able to
complete the study protocol and to
provide written informed consent.
Significant pre-stroke physical impairment
(modified Rankin score > 1, or Barthel
Index < 95)
To ensure participants were able to
complete protocol.
Contraindication to MRI Safety of participants
Table 2: Exclusion criteria
Inclusion criteria Reason
Written informed consent To ensure patients understand and agree
to taking part in the research.
Clinical lacunar syndrome These criteria represent a group of
patients likely to have cSVD. Appropriate MRI lesion(s) (defined as
subcortical infarct ≤1.5cm in diameter)
Aged 18-85
≥1 month from stroke onset To minimise the influence of short-term
haemodynamic changes in the acute post-
stroke phase and to avoid acute effects on
MRI or neuropsychology.
≤12 months from stroke onset To avoid cumulative effects of cSVD,
and to ensure that the study findings
might be useful for prognosis.
Table 1: Inclusion criteria.
52
4.1.1 Sample size calculation
Quantitative fMRI has not been used in patients with cSVD before. Indeed the reality
of trying to quantitatively measure the BOLD response is a recent one. Goodwin et al.
(186) recently used hyperoxia-calibrated fMRI with a modified Stroop Task, reporting a
reasonable reproducibility in the measurements. They provide data on the change in
„n‟, a measure of NVC, across six regions in ten young healthy adults. The value varies
across the regions from 2.0 in the primary motor cortex to 3.14 in the right parietal lobe.
An average change in „n‟ is given, but without a standard deviation needed for sample
size calculation. „n‟ was found to be significantly greater in the middle frontal gyrus
than in the motor cortex, which the authors tentatively attribute to a greater vascular
responsiveness in the frontal region. Left and right middle frontal gyri values for „n‟
were not significantly different. In the left middle frontal gyrus „n‟ was equal to 2.83
with a standard deviation of 1.0. These data were used to calculate the sample size for
the current study.
To detect a 25% reduction in „n‟ in patients with cSVD versus healthy controls, it was
calculated that a minimum of 16 patients (and 16 controls) would be needed, using 80%
power and 5% significance:
Where,
N = number of patients needed
U = Power; U = 0.84 for 80% power
V = Significance; V = 1.96 for 5% significance
σ = Standard deviation
53
µ = Expected mean value of „n‟
µ0 = Null hypothesis mean value of „n‟
Sixteen patients seems a small sample. However, due to the expensive and time-
consuming nature of MRI studies it does not represent a comparatively
disproportionately small cohort. Indeed the calculation of „n‟ from the Goodwin paper
(186) was based upon a sample of just 10 participants.
4.2 Neuropsychological tests
On the day of the scan, subjects had clinical assessments of cognition. Executive
dysfunction is the dominant feature of cognitive impairment associated with vascular
disease. It is also thought that there might be an age-related decline in executive
relative to other cognitive functions (9). O‟Sullivan et al. (9) show that executive
assessment provides better discrimination of patients with ischaemic leukoaraiosis
(presumed cSVD) from healthy elderly controls than other established brief assessment
tools, such as the MMSE. They also conclude that this form of assessment
distinguishes age- from disease-related changes, in those aged 50–84 years. The battery
of clinical cognitive tests used comprised those tests recommended by O‟Sullivan et al.
(9) and are detailed in Box 4.
These data will be analysed using a two-tailed Student‟s t-test. This would be expected
to show a difference between the two groups: that executive function is diminished in
the cSVD group. The difference will then be assessed against the scans to suggest
which, if any, correlate best with executive impairment.
Patients were also functionally assessed using the National Institute of Health Stroke
Scale (NIHSS) on the day of the scan.
54
Box 4: Neuropsychological test battery.
Trail making test
o Assessment of cognitive set shifting and mental flexibility
o This has two parts, both of which are timed. The first part, A,
involves joining randomly positioned numbered points in the correct
numerical order. Part B is similar but has lettered as well as
numbered points and requires a sequence of 1-A-2-B-3-C etc. Part B
contains the executive component of shifting between cognitive sets,
i.e. between numerical and alphabetical rules. However, parts A and
B share non-executive aspects, such as visual scanning and motor
function. Thus, subtracting the time for part A from that of B,
theoretically accounts for non-executive effects.
Verbal fluency
o Assessment of the ability to generate words
o The subject has one minute for each letter to name as many words as
possible beginning with F, A and then S.
Digit span
o Assessment of working memory (168,169).
o Subjects are required to repeat back progressively lengthening lists of
numbers, first forwards and then in reverse. The backwards
component is considered a better test of working memory and hence,
executive function.
DSST
o Assessment of executive functioning and performance IQ.
o Digits 1-9 are assigned a unique symbol. The subject has 90 seconds
to enter the corresponding symbols for a series of digits, requiring
shifts between the rules for each digit.
55
4.3 Imaging
4.3.1 Image acquisition
A 3 Tesla whole-body MRI scanner
(Siemens Trio, Erlangen, Germany) was
used, with an 8-channel head
radiofrequency coil and a body coil for
signal collection and transmission,
respectively.
BOLD and CBF signal were measured
simultaneously using a QUIPSSII (187)
ASL sequence. The MRI acquisition
parameters for the fMRI scans were: TR (repetition time) 2.13s, TI2 1.4s, TE (echo
time) 20ms. Twelve slices of 3.5mm thickness and 0.35mm gap covering frontal and
motor cortices were taken.
Scan Purpose
T2-weighted Exclude cortical lesions and measure LA (WML).
Angiography of carotids and
circle of Willis
Exclude intra- and extra-cranial vessel stenosis.
Calibration – ASL and BOLD To allow calculation of calibration constant „M‟
according to Chiarelli et al.(170)
Stroop Task – ASL and
BOLD fMRI
To measure CBF and BOLD, and to allow
calculation of CMRO2, using the constant „M‟ from
calibration scan.
SWI To identify microbleeds.
T1-weighted Reviewed by a radiologist to exclude cranial
abnormalities unrelated to cSVD.
Estimate brain structure volume.
FLAIR Exclude cortical lesions and measure LA (WML).
Table 3: List of MRI scans and their purpose.
Figure 10: Slice coverage.
56
4.3.2 Stimulus paradigm
A colour-word Stroop Task was used (103).
Subjects were presented with a black screen
with two words, positioned above and below
a white fixation cross (Figure 11). They were
required to determine if the meaning of the
bottom word, written in white ink, matched
the ink colour of the top word. Using a button
box, they answered with the index finger of
the right hand for „yes‟ and middle finger for
„no‟. They were asked to respond as quickly
and accurately as possible. The task was self-paced, meaning that the screen would not
change until a response was registered. The minimum time between stimuli was 2
seconds. There were 8 active blocks of 30 seconds, each followed by 30 seconds of a
fixation cross on a black screen, resulting in a run time of minimum 8 minutes. Each
subject practised 4 blocks outside of the scanner to ensure the task was understood.
4.3.3 Hyperoxia paradigm
The hyperoxia paradigm was used to calibrate the BOLD signal. The subject had 2
periods each lasting 3 minutes of breathing high-flow oxygen through an open mask, at
a rate of 15L/min. Each of these periods was preceded by breathing normal air (Figure
12). End-tidal oxygen was sampled continuously via a nasal cannula. A vacuum pump
(Flowcontrol R-2 vacuum pump) connected the nasal cannula to an oxygen analyser (S-
3A/IO2 oxygen analyser; provided by Applied Electro-chemistry Inc, Pittsburgh, USA).
The oxygen analyser was calibrated against the oxygen concentration of room air
(GasMonitor4) before each scan. Respiratory data was logged at 1ms intervals using
Powerlab software (ADInstruments, Colorado Springs, USA).
Figure 11: Example slide from Stroop
task.
57
4.4 Data analysis
BOLD and perfusion image time series were created in MATLAB (The MathWorks
Inc, Massachussets) from the ASL images collected during both the hyperoxia and
Stroop Task runs, using „in-house‟ software (186).
4.4.1 Identifying the active regions of interest
The CBF and BOLD image time series for each subject were pre-processed using Brain
Voyager (Brain Innovation B.V., Maastricht, Netherlands). This involved spatial (full
width at half maximum (FWHM) 6mm) and temporal (FWHM 10s) smoothing to
reduce the noise in the data (161), and linear trend removal to remove any linear drift in
the signal. The CBF and BOLD images were co-registered to a structural T1–weighted
image that had been transformed into Talairach space. The aggregated Stroop Task runs
of all patients and controls were analysed using a general linear model (GLM).
Significantly active voxels were highlighted (Figure 13).
The areas of highlighted voxels were identified. In these regions of interest (ROI) the
Stroop Task activity resulted in significant (p<0.05; corrected for false-discovery rate)
differences in the BOLD and CBF time-courses. In each ROI, a volume approximately
1cm3 was chosen encompassing the site of highest activity. In each of these, the
average signal time-course for both CBF and BOLD were recorded for both the Stroop
Task and hyperoxia runs.
2 min
3 min
3 min
O2 15L/min
3 min
O2 15L/min
Figure 12: Hyperoxia paradigm.
58
4.4.2 Quantification of neural and vascular parameters
The neural and vascular parameters of interest in this study were the change in (Δ)
BOLD, ΔCBF, calibration constant „M‟, ΔCMRO2 and the NVC parameter „n‟
(=ΔCBF/ΔCMRO2). The first step, then, was to calculate „M‟ with which the other
parameters might be computed. MATLAB was used to extract the end-tidal oxygen
values for 4 time periods, each averaged over 1 minute: the final minute of each
hyperoxia and each normoxia period. The corresponding BOLD signal was used from
the same periods. The model provided by Chiarelli et al. (170) for calculating an
estimate of „M‟ was used with experimental data substituted in:
Figure 13: Example showing highlighted regions
of interest in the left and right precentral gyri
(circled), coronal view.
59
1
][
][1 0
000 HO
HOHOHO
CBF
CBF
dHb
dHb
CBF
CBFM
BOLD
BOLD [1]
A B C D E
The subscript „HO‟ refers to the corresponding parameter during hyperoxia, and
„0‟ at baseline. Therefore, ΔBOLDHO is the change in the BOLD signal during
hyperoxia and BOLD0, the baseline value for BOLD signal.
A reduction in CBF of 5% during hyperoxia is assumed, based on the work of
Chiarelli et al. (170). The assumption is used in preference to measured values
which have a low SNR and need to be corrected to take into account the altered
T1 of arterial blood during hyperoxia.
[dHb] = the concentration of deoxygenated haemoglobin in the venous
vasculature. The [dHb] values were calculated from the end-tidal oxygen
measurements as described by Chiarelli et al. (170) A baseline oxygen extraction
fraction (OEF) of 0.4 was assumed (170). Furthermore, it was assumed that the
OEF did not change with hyperoxia (170).
α = 0.38. This value, the Grubb relation, was obtained in previous work by Grubb
et al. (188) by calculating CBV changes from CBF changes, based on the
assumption inherent in this model that the system is in equilibrium.
β = 1.5. This value is assumed as in previous work (159,170).
M = Calibration constant; it represents the maximum theoretical BOLD signal
change (170).
Values were calculated for „M‟ in each ROI over both individual hyperoxia periods and
the mean value of the two combined.
The calibration of the BOLD signal is described in detail by Chiarelli et al.(170). In
brief, parts A, C, D and E of equation [1] are calculated from experimental data, with
which part B can be calculated. Part D is derived from the end-tidal oxygen data, from
60
which the partial pressure of oxygen is inferred. Three assumptions are involved in this
calculation. Firstly CBV is assumed to not be influenced by a change in pure oxygen.
Secondly, a baseline value of OEF is assumed to be consistent throughout the brain and
constant during hyperoxia. Finally, haemoglobin (Hb) is assumed to be 15g/dL. The
saturation of arterial Hb is calculated using the partial pressure of oxygen, and is then
used to calculate the arterial oxygen content using the species-dependent oxygen-
carrying capacity of Hb. From this the concentration of deoxygenated haemoglobin in
the venous vasculature can be derived.
The change in CBF and BOLD were calculated by averaging the time-courses for each
ROI over the eight stimulation blocks and subtracting the corresponding averaged value
of the „rest‟ blocks. These values were calculated on an individual basis in each ROI.
The model described by Buxton et al. (159) was used to calculate the ΔCMRO2:
)1(02
2
00 CMRO
CMRO
CBF
CBFM
BOLD
BOLD [2]
„n‟ is a measure of the relationship between changes in CBF and CMRO2:
022
0
/
/
CMROCMRO
CBFCBFn
[3]
The models for oxygen-calibration have been developed in healthy subjects. This is the
first study in which the technique has been applied to a population with cSVD. It is
thus entirely plausible that the constants and parameters used might not apply to a
patient population. These limitations shall be addressed further in the discussion.
61
4.4.3 Assessment of cerebral atrophy and ventricular enlargement
Coronal T1-weighted images, in the
anterior commissure- posterior
commissure (ACPC) plane, were
exported from BrainVoyager into
EasyMeasure (©2005 Mike
Puddephat). This programme slices the
brain into regular sections and
randomly overlays a grid of points on
to each section surface. The number of
points over a structure of interest, e.g.
the ventricles, are totalled for each
section and used to calculate the
volume for that structure in mm3. Total
brain matter (excluding the
cerebellum), subcortical matter and grey matter volumes were measured using a grid
size of 8mm and interval between slices of 15mm. The volume of the lateral ventricles
was measured using a grid size of 3mm and interval between slices of 6mm. In the
event of a cross seeming to lay over two structures, e.g. the border of white and grey
matter, the structure in the lower right quadrant of the cross was used. All measured
volumes were given as a percentage of whole brain volume, to correct for any general
difference in brain size which is not relevant.
Figure 14: Example grid used to measure
volumes of brain structures with
EasyMeasure.
62
4.4.4 Assessment of WML
WML were assessed blind using the rating scale of Wahlund et al. (189). WML
visualised on MRI were defined as “ill-defined hyperintensities ≥ 5mm on both T2 and
... FLAIR images”. Five regions were rated on a scale of 0-3 for each hemisphere:
frontal, parieto-occipital, temporal, infratentorial and basal ganglia regions. A score of
0 represented no visible lesions, and a score of 3 represented diffuse involvement of the
region. The score for each region was summed across right and left hemispheres,
resulting in a score of 0-6 for each region.
4.4.5 Assessment of microbleeds
Microbleeds were assessed using the Microbleed Anatomical Rating Scale of Gregoire
et al.(190). Microbleeds were defined as “small, round, well-defined, hypointense on
Gradient Echo T2*; 2-10mm; not well seen on T2.” Susceptibility-weighted images
were used rather than gradient echo T2* as they have been shown to have a better rate
of detection of cerebral microbleeds (191,192). The number of microbleeds was
counted for each of 13 regions in both right and left hemispheres:
Infratentorial: Brainstem; and cerebellum.
Deep: Basal ganglia (caudate and lentiform); thalamus; internal capsule; external
capsules; corpus callosum; and deep and periventricular white matter.
Lobar (including cortex and subcortical white matter): Frontal; parietal;
temporal; occipital; and insula.
The total number of microbleeds was calculated for infratentorial, deep, lobar and total
regions.
4.4.6 Angiography
Stenoses greater than 50% were excluded by an experienced neuroradiologist, using the
MRA images of the carotid arteries and circle of Willis.
63
4.5 Statistical analysis
Microsoft Excel and SPSS were used for statistical analysis. The mean, standard
deviation and standard error were calculated for all variables. Patient and control
groups were compared using a two-tailed Student‟s T-test for all variables. P-values
less than 0.05 were deemed significant. One way analysis of variance (ANOVA) with
repeated measures was used in SPSS to identify difference between regions of interest,
where t-test showed a significant difference between patients and controls. Negative
values for „n‟, CMRO2 or M were deemed nonsense values attributed to noise and were
excluded. Values greater than three standard deviations from the mean were considered
outliers and also excluded. Image measures were individually plotted against
statistically significant cognitive function measures. Regression analyses were
performed using the LINEST function in Excel and the probability value of any
correlation was calculated using the Fisher F-statistic.
64
Chapter five: Results
5.1 Recruitment
Fifty-two suitable patients with likely cSVD were identified and contacted by letter and
followed up with a telephone call. Fourteen individuals agreed to participate in the
study. Two were unable to complete the study protocol, one due to arthritis and the
other due to claustrophobia. A further participant was excluded because of a space
occupying lesion observed on MRI. Thus the data were derived from the eleven
remaining patients. One of these did not have the hyperoxia calibration scan due to
COPD. As a result ∆CMRO2, M and „n‟ could not be calculated for this subject.
Sixteen age-, sex- and education-matched healthy controls were also recruited.
Figure 15: Chart to
demonstrate the difficulties in
achieving the desired sample
size; an example taken from
individuals presenting to UHA
between 01/02/2008 and
01/02/2009.
Substantially more patients were identified and recruited from UHA than from the other
hospitals (Table 4). I obtained an honorary contract at UHA and personally searched
the patient database of all stroke admissions. Thus it is feasible that a more thorough
search of patient records, including discharge and clinic letters, permitted greater
identification. Such a search was not possible at other hospital sites.
540 Stroke admissions
55 Coded as lacunar stroke
25 met selection criteria
6 participated
65
Hospital Eligible Participated
UHA 25 6
Royal Liverpool University Hospital and
Broadgreen Hospital
9 3
Walton Centre for Neurology and
Neurosurgery
5 2
Royal Preston Hospital 4 0
Countess of Chester Hospital 9 3
Total: 52 14
Table 4: Derivation of included patients.
5.2 Participant demographics
Controls Patients p-values
Number 16 11 -
Mean age (years) ±
SD
60.9 ± 6.9 59.5 ± 6.7 0.60
Female, n (%) 8 (50) 2 (18) 0.086
Mean education
(years) ± SD
16.2 ± 2.6 14.5 ± 3.3 0.16
Mean BMI (kg/m2)
± SD
27.5 ± 4.4 28.7 ±6.8 0.61
Mean pulse ± SD 71.0 ± 7.65 72.7 ± 10.0 0.63
Mean systolic BP
± SD
130.6 ± 15.4 136.2 ± 19.5 0.44
Mean diastolic BP
± SD
79.3 ± 11.1 77.5 ± 10.8 0.69
Table 5: Participant demographics; SD = Standard deviation, BMI = Body Mass Index.
66
There were no significant differences in age, sex, years of education, BMI, pulse or BP
between patient and control groups (Table 5). Although not statistically significant (p =
0.086) there were substantially fewer women in the patient (18%) compared to the
control (50%) group.
5.3 Neuropsychological test results
Results from two out of the four cognitive tests differed significantly between patient
and control groups (Table 6). However, after correction for multiple tests, only the
DSST score remained significant (p < 0.0125). The DSST score was thus used as a
standard surrogate measure of clinical severity against which to compare any
differences in „n‟ and other possible markers that showed a significant difference
between control and patient groups.
Cognitive tests Controls Patients p
Trail making B – A
(seconds)
22.7 ± 16.6 32.3 ± 19.9 0.21
Digit span
backward
7.7 ± 1.9 6.4 ± 3.2 0.24
Digit symbol
substitution test
56.8 ± 10.0 38.1 ± 9.3 0.000052
Verbal fluency
FAS total
54.9 ± 14.7 41.4 ± 16.3 0.040
Table 6: Mean cognitive scores ± standard deviation for patients and controls.
Significant differences (p < 0.05) in test results are highlighted.
67
5.4 Stroop Task
Patients were significantly slower (p = 0.049) than the controls at the Stroop Task
(Figure 16). There was no significant difference in accuracy.
Figure 16: Bar charts with standard error bars to demonstrate the difference in
accuracy (left) and response time (right) to the Stroop Task between controls and
patients.
5.5 Neurovascular coupling parameter ‘n’
BOLD activation was extremely robust and covered the whole brain at p = 0.05
corrected for false discovery rate (193,194). Therefore, ROI were restricted by cluster
size rather than by statistical threshold. Each ROI was limited to a cluster size of 1000
voxels. Eleven activated ROI were identified (Figure 17). The Talairach daemon (195)
was used to confirm the central Talairach coordinates (x, y, z) and anatomical label for
each region:
Left medial frontal gyrus (LMeFG; -9, -1, 48)
Right medial frontal gyrus (RMeFG; 6, 8, 43)
Left middle frontal gyrus (LMiFG; -40, 34, 24)
Right middle frontal gyrus (RMiFG; 37, 25, 33)
Left insula (LI; -32, 17, 14)
Right insula (RI; 33, 18, 10)
Left precentral gyrus (LPCG; -35, -23, 54). This is the primary motor area.
Right precentral gyrus (RPCG; 43, -1, 36)
Left parietal lobe (LPL; -30, -63, 42)
Right parietal lobe (RPL; 26, -64, 41)
Left inferior parietal lobule (LIPL; -38, -50, 39)
68
Figure 17: ROI
selected for
analysis are
circled.
I PCG MiFG
MeFG LIPL PL
The results were averaged over the following larger regions for analysis: frontal
(incorporating LMeFG, RMeFG, LMiFG, RMiFG and RPCG), parietal (LPL, RPL and
LIPL), insular (LI and RI), motor (LPCG) and overall (all ROI) regions.
Negative values of „n‟, „M‟ or „CMRO2‟ were excluded from the raw data, as these
were deemed to be physiologically implausible. Negative values were attributed to
noise in the data, or experimental error, for example a poor connection leading to
incorrect end-tidal oxygen values. Outliers, defined as values more than three standard
deviations greater than the mean, were also discarded. There were no values of „n‟, „M‟
or „CMRO2‟ for the patient in whom there was no oxygen calibration scan. In total 9%
of the raw data were excluded. Of the excluded data 72.7% were due to negative
values, 22.7% due to SVD1 (patient without oxygen calibration scan) and 4.5% due to
outliers.
The neurovascular coupling parameter „n‟ was reduced in the patient group compared to
the control group (p = 0.021) (Figure 18).
69
Figure 18: Bar chart with standard error bars showing a reduction in ‘n’ in patients
with cSVD compared with controls.
Figure 19: Bar chart to show differences in 'n' between patient and control groups
across regions, with standard error bars.
„n‟ appears to be consistently reduced in patients in all individual regions, but with no
clear regional variation graphically (Figure 19). A one way ANOVA with repeated
measures and Greenhouse-Geisser correction showed no significant difference between
regions of interest (Box 5).
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Control Patients
n
Neurovascular coupling parameter 'n' between groups
Control
Patients
70
(I) ROI (J) ROI
Mean
Difference (I-J) Std. Error Sig.a
95% Confidence Interval for
Differencea
Lower Bound Upper Bound
Frontal Parietal -.162 .199 1.000 -.757 .432
Insular -.422 .243 .603 -1.146 .303
Motor -1.041 .527 .386 -2.613 .530
Parietal Frontal .162 .199 1.000 -.432 .757
Insular -.259 .298 1.000 -1.150 .631
Motor -.879 .504 .597 -2.384 .626
Insular Frontal .422 .243 .603 -.303 1.146
Parietal .259 .298 1.000 -.631 1.150
Motor -.620 .448 1.000 -1.956 .716
Motor Frontal 1.041 .527 .386 -.530 2.613
Parietal .879 .504 .597 -.626 2.384
Insular .620 .448 1.000 -.716 1.956
Box 5: One-way ANOVA with repeated measures comparing n in ROI. Adjustment for
multiple comparisons: Bonferroni.
„n‟ was compared to DSST separately for patient and control groups (Figure 20). The
patient group showed no association between DSST and „n‟ (p = 0.74, r2 = 0.014). The
control group showed no association between DSST and „n‟ (p = 0.16, r2 = 0.14). The
relationship between DSST and „n‟ was reassessed with a possible outlier removed
apparent for the purposes of discussion (Figure 21).
71
Figure 20: Comparison of DSST and ‘n’ in patient and control groups.
Figure 21: Graph showing a correlation between DSST and ‘n’ in the control group,
once an apparently anomalous result had been removed.
72
5.6 Additional functional measures
5.6.1 BOLD
The average BOLD signal was not significantly reduced over all regions in the patient
group compared to the control group (Figure 22). The BOLD curves were similar for
both patients and controls, except in the motor region where the amplitude of the BOLD
signal was reduced in patients compared to controls (Figures 23 and 24).
BOLD curves averaged over all regions
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
-20 0 20 40 60
Time (seconds)
BO
LD
sig
nal
ch
an
ge (
%)
Controls
Patients
Frontal BOLD curves
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
-20 0 20 40 60
Time (seconds)
BO
LD
sig
nal
ch
an
ge (
%)
Controls
Patients
Figure 22: Bar chart to show the
difference in BOLD signal change
between patient and control
groups across selected regions,
with standard error bars.
Figure 23: Averaged BOLD
curves during the Stroop task
across total (top) and frontal
(bottom) regions, with standard
error bars.
73
Motor BOLD curves
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
-20 0 20 40 60
Time (seconds)
BO
LD
sig
nal
ch
an
ge (
%)
Controls
Patients
Parietal BOLD curves
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
-20 0 20 40 60
Time (seconds)
BO
LD
sig
nal
ch
an
ge (
%)
Controls
Patients
Insular BOLD curves
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
-20 0 20 40 60
Time (seconds)
BO
LD
sig
nal
ch
an
ge (
%)
Controls
Patients
Figure 24: Averaged BOLD curves during the
Stroop task across (from top to bottom) motor,
parietal and insular regions.
74
5.6.2 CMRO2
CMRO2 was not significantly different between the patient and control groups (Figure
25).
5.6.3 CBF
Resting CBF was not significantly different between the patient and control groups
(Figure 26).
Figure 25: Bar chart to
show the difference in
CMRO2 between patient
and control groups across
selected regions, with
standard error bars.
Figure 26: Bar chart with
standard error bars to show
differences in CBF in response
to the Stroop Task between
patient and control groups
75
5.6.4 CBF in response to Stroop Task
There were no significant differences in CBF response to the Stroop Task between
patient and control groups (Figure 27).
CBF was not significantly different between the patient and control groups. There was
no correlation between „n‟ and CBF in either patient (p = 0.51, r2 = 0.055; Figure 28) or
control (p = 0.50, r2 = 0.033; Figure 29) groups.
0
0.5
1
1.5
2
2.5
3
3.5
4
0 5 10 15 20 25 30 35
'n'
CBF
CBF Vs 'n' in Patient group
Figure 27: Bar chart with
standard error bars to show
differences in CBF in
response to the Stroop Task
between patient and control
groups.
Figure 28: Graph showing no significant correlation between CBF and
‘n’ in the patient group.
76
Figure 29: Graph showing no significant correlation between CBF and ‘n’ in the
control group.
5.6.5 ‘M’
The averaged value of „M‟ was significantly increased in patients in all regions (p =
0.05) (Figure 30).
0
1
2
3
4
5
6
7
0 10 20 30 40 50 60
'n'
CBF
CBF Vs 'n' in Control group
0
2
4
6
8
10
12
Control Patients
M
Mean M between groups
Control
Patients
Figure 30: Bar chart with standard error bars to show
the differences in M between control and patient groups.
77
A one way ANOVA with repeated measures and Greenhouse-Geisser correction
showed „M‟ to be significantly increased in frontal regions compared with motor
regions (p = 0.012) (Box 6). However, there was no significant difference between M
in frontal regions compared with parietal and insular regions.
(I) ROI (J) ROI
Mean
Difference (I-J) Std. Error Sig.a
95% Confidence Interval for
Differencea
Lower Bound Upper Bound
Frontal Parietal .596 .458 1.000 -.760 1.953
Insular .651 .411 .785 -.568 1.870
Motor 1.544* .428 .012 .276 2.812
Parietal Frontal -.596 .458 1.000 -1.953 .760
Insular .055 .505 1.000 -1.442 1.551
Motor .947 .370 .117 -.148 2.043
Insular Frontal -.651 .411 .785 -1.870 .568
Parietal -.055 .505 1.000 -1.551 1.442
Motor .893 .425 .300 -.367 2.152
Motor Frontal -1.544* .428 .012 -2.812 -.276
Parietal -.947 .370 .117 -2.043 .148
Insular -.893 .425 .300 -2.152 .367
Box 6: One-way ANOVA with repeated measures comparing M in ROI. Adjustment for
multiple comparisons: Bonferroni.
Figure 31: Bar chart with
standard error bars to show
regional variation in M.
78
Figure 32 demonstrates an inverse correlation between n and M (r2 = 0.21, p = 0.017).
Figure 32: Scatter graph to show the correlation between ‘M’ and ‘n’.
0
1
2
3
4
5
6
7
0 2 4 6 8 10 12 14
n
M
Correlation between M and n
79
Ventricular enlargement
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
3.00%
Controls Patients
No
rmali
sed
ven
tric
ula
r vo
lum
e
5.7 Structural measures
5.7.1 Cerebral volumes
Mean of measured
volumes as a percentage
of whole brain volume
Control Patient p
Subcortical matter 52.1 ± 3.0 50.5 ± 4.0 0.28
Grey matter 46.8 ± 3.2 47.3 ± 4.7 0.73
Ventricular 1.1 ± 0.8 2.2 ± 1.5 0.058
Table 7: Mean normalised volumes (%) ± standard deviation.
None of the volume measurements were significantly different between patient and
control groups (Table 7 and Figure 33). Although the patients‟ ventricles were double
the size of the controls‟ ventricles, the difference did not reach significance (p = 0.058).
Grey matter atophy
44.50%
45.00%
45.50%
46.00%
46.50%
47.00%
47.50%
48.00%
48.50%
49.00%
Controls Patients
No
rmali
sed
gre
y m
att
er
vo
lum
e
Subcortical atrophy
47.00%
48.00%
49.00%
50.00%
51.00%
52.00%
53.00%
54.00%
Controls Patients
No
rmali
sed
su
bco
rtex v
olu
me
Figure 33: Bar charts with standard error
bars demonstrating the differences in
normalised brain structure volumes
between patient and control groups.
80
The volume scores of individual participants were charted against their score on the
DSST. The cognitive test score showed a trend towards correlation with all three
measures, ventricular enlargement (r2 = 0.051, p = 0.03) and subcortical atrophy (r
2 =
0.19, p = 0.02), but not grey matter atrophy (r2 = 0.11, p = 0.09) (Figure 34), but was not
significant after Bonferroni correction.
Graph showing normalised grey matter volume
against DSST score
30.00%
35.00%
40.00%
45.00%
50.00%
55.00%
60.00%
0 20 40 60 80
DSST score
No
rmali
sed
gre
y m
att
er
vo
lum
e (
%)
Controls
Patients
Graph showing normalised subcortical matter
volume against DSST score
40.00%
45.00%
50.00%
55.00%
60.00%
0 20 40 60 80
DSST score
No
rmali
sed
su
bco
rtic
al
matt
er
vo
lum
e (
%)
Controls
Patients
Graph showing normalised ventricular volume
against DSST score
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
0 20 40 60 80
DSST score
No
rmali
sed
ven
tric
ula
r
vo
lum
e (
%)
Controls
Patients
Figure 34: Scatter graphs to
show individual normalized
volume measurements plotted
against DSST score. From top
to bottom: grey matter,
subcortical matter and
ventricular volumes.
81
3.7.2 WML
WML score was greater in patients than controls (p = 0.012) (Table 8). Frontal (p =
0.037) and basal ganglia (p = 0.041) regions were particularly affected by WML in the
patient group. WML lesion scores for frontal, basal ganglia and total regions were
individually plotted against DSST score. There was no correlation between „n‟ and
WML score in any region.
WML Controls Patients P value
Frontal 13 29 0.037
Parieto-occipital 9 21 0.082
Temporal 2 5 0.23
Infratentorial 0 3 0.19
Basal ganglia 7 16 0.041
Total 31 74 0.012
Table 8: Sum WML by region for patient and control groups. There were no
statistically significant values after Bonferroni correction.
3.7.3 Microbleeds
Microbleeds Controls Patients P value
Infratentorial 2 1 0.85
Deep 4 2 0.80
Lobar 4 4 0.50
Total 10 7 0.80
Table 9: Total number of microbleeds by region for patient and control groups.
There was no microbleed result for one patient because the images were unreadable due
to movement artefact. There were no significant differences in the number of
microbleeds between patient and control groups (Table 9).
82
Chapter six: Discussion
6.1 Discussion of findings
This study used a modified Stroop Task with hyperoxia-calibrated fMRI to investigate
whether patients with cSVD exhibited differences in NVC when compared to an age-,
sex- and education-matched healthy control group. „n‟, a measure of NVC, was
significantly reduced in the patient group. „M‟, which represents the theoretical
maximum BOLD signal change, was significantly increased in patients compared to
controls. Furthermore „M‟ was raised in frontal regions compared to motor regions. Of
the structural measures, only WML score was found to be more common in the patient
group compared to control group. WML score did not correlate with „n‟.
Although neither group showed a significant correlation between „n‟ and DSST, the
graphical demonstration of the data suggests a trend in the control group. There is one
clear anomaly (circled in Figure 20). If the circled anomaly were to be momentarily
removed from analysis then there is a clearly significant correlation between DSST and
„n‟ in the control group (p = 0.0057, r2 = 0.46) (Figure 21). The study has shown no
true correlation, but it seems plausible that the study might be under-powered to show
such a difference.
The haemodynamic response correlates well with neuronal activity in healthy
individuals, but the correlation is less clear in certain disease states, i.e. there is an
uncoupling of the neurovascular response (56). Differences in BOLD signal between
groups can only be assumed to relate to neural activity if NVC is the same in both
groups (104). The findings of the current study show a difference in NVC between
patients with cSVD and healthy controls. The results suggest an un-coupling of the
haemodynamic response to neural demand in patients with cSVD.
In theory, „n‟ suggests itself as a suitable marker for cSVD. It is a measure of the
coupling of a vascular response with neural activity. cSVD affects the small vessels
that are in close proximity to neural tissue. The neurovascular unit (NVU) consists of
83
neurones, glia and vascular cells, which are developmentally, structurally and
functionally closely related (56). Mediators of NVC are thought to possibly be released
by cerebral endothelial cells (56). One might hypothesise that cSVD leads to damage of
small cerebral vessels which make up the vascular contribution of the NVU, perhaps as
a result of impaired VMR. This might prevent the supply of oxygen and glucose and
the removal of waste products from the neuronal tissue leading to damage of the
neuronal tissue and eventually to cognitive decline. If this were the case, that neural
activity resulted in an imperfect vascular response, or impaired NVC, then „n‟ would be
a very early marker of cSVD, should it prove to be a sensitive measure of NVC. The
current study has shown „n‟ to be reduced in patients with cSVD. It does not follow that
„n‟ is an early and sensitive marker of cSVD, nor even of NVC, but it does suggest the
possibility that it might be.
Impairment of NVC in patients with cerebrovascular disease has been reported in other
small studies. Schroeter et al. (49) reported a delay of the haemodynamic response in
patients with CMA and also an early deoxygenation that was not seen in the control
group. These findings were attributed to impaired NVC. CMA is a term often used
interchangeably with cSVD. However, five of the twelve patients with CMA in the
study by Schroeter et al. (49) had evidence of macroangiopathy.
Rossini et al. (76) also suggest that BOLD fMRI might not accurately define the sites
and amount of neural activation in response to a task in patients with cerebrovascular
disease. In a study of ten patients with a history of stroke or transient ischaemic attack,
BOLD fMRI activation and neurophysiological events, quantified by
magnetoencephalography (MEG), were frequently not correlated. In other words, NVC
was impaired. Moreover, impaired NVC correlated with impaired cerebral VMR.
Rossini et al. (76) propose that maximal dilatation of the vascular tree might limit the
haemodynamic response despite normal neural activity and the normal release of factors
implicated in initiating the haemodynamic response, such as NO (56). In other words,
impaired VMR is not just correlated with, but in fact results in, impaired NVC.
84
Impaired VMR has been shown in patients with cSVD using PET (196), transcranial
Doppler sonography (26), fNIRS (26,197) and fMRI (198,199). Terborg et al. (26)
suggest that impaired VMR in cSVD might explain the pathophysiology of WML:
episodes of hypotension might result in periods of „critical hypoperfusion‟ in the deep
white matter due to a failure of autoregulation because of impaired VMR, leading to
ischaemia.
Impaired VMR has been shown to correlate with impaired NVC (76), and has been
suggested as a causative factor in WML (26). One might thus consider a correlation
between impaired NVC and WML. However, the current study found no association
between WML score and „n‟.
Krainik et al. (199) suggest damage to neuronal tissue innervating the vasculature in
regions adjacent to an infarct as an alternative to impaired VMR as a cause of impaired
NVC. A third hypothesis of altered NVC is that release of NO might be limited by
chronic hypertension (132).
Differences in NVC between diseased and non-diseased groups imply that the BOLD
signal itself is not a reliable marker of neural activity in diseased groups. The findings
of this study lend credence to such a supposition. There were no significant differences
in BOLD signal change between patient and control groups. Reductions in BOLD
signal have been reported in the literature in patients with stroke (199) and cSVD
(132,198). Reductions in the haemodynamic response have also been noted in the
healthy elderly compared to younger persons using fMRI (200) and fNIRS (102).
However, reduced haemodynamic response might reflect impaired NVC, rather than a
reduced neural response to functional activation.
The BOLD curves of the patient group in the current study took slightly longer to return
to pre-stimulus levels in all regions. This might reflect the longer response time to the
Stroop Task observed in the patient group. The prolonged response time in patients to
85
the Stroop Task is in concordance with previous studies of elderly versus young healthy
individuals (102) and cSVD (49).
A reduction in CBF has been commonly observed in cSVD (55,113,131,167,201-
204,204-206) and has been shown to correlate with cognitive decline (113,205). The
results of the current study did not corroborate these findings, however. CBF in
response to the Stroop Task was not significantly different between control and patient
groups. Resting CBF was also not significantly different. However, the current study
showed a tendency to a rise in CBF. Many studies compared a dementia group to a non-
dementia group (55,131,201,203,206). It might be argued that cerebral atrophy leads to
dementia and a reduction in CBF, i.e. that observable reductions in CBF are only
apparent with dementia. Demented patients were excluded from the current study,
perhaps explaining why CBF reductions were not observed. However, the
comparatively large study of 57 well-selected patients with cSVD by Sabri et al. (113)
provides strong evidence against such an argument. They found reduced CBF in
patients with cSVD in the absence of atrophy and also found CBF to correlate with
measures of executive function. However, the reductions in CBF might have been
related to hypertension rather than cSVD per se. Of the 57 patients with cSVD 91%
had chronic hypertension. There were 19 age-matched controls, but the number of
hypertensive subjects was not reported nor accounted for. The significance of no
difference in resting CBF between groups in the current study is not clear. This might
represent a chance anomaly or reflect the small sample size.
The methodology assumes a fixed 5% reduction in CBF during the periods of hyperoxia
(186). This assumption is based on the figure used by Chiarelli et al. (170) and
Goodwin et al. (186). However, the reduction in CBF during hyperoxia reported in the
literature is variable, ranging from 5% to 26.8% (170,186,207-209). Goodwin et al.
(186) suggest that the reduction in CBF is less with short periods of hyperoxia, though
the original evidence for this statement is unclear.
86
There is evidence that CBF reduction during hyperoxia changes with age (209). In
many respects, cSVD appears to exaggerate age-related changes (7,49). It is thus
plausible that the change in CBF during hyperoxia might be different in patients with
cSVD compared to healthy age-matched controls. Watson et al. (209) found a smaller
reduction in CBF in response to hyperoxia in older age groups. Hypertension is more
common in older age groups. One might conjecture that should this finding be
extrapolated to cSVD, patients with cSVD might have a smaller drop in CBF during
hyperoxia than age-matched controls. Thus, if resting CBF is measured during
hyperoxia, as in the current study, then resting CBF values might appear higher in
patients with cSVD than in the control group, not due to actual resting CBF but to the
effect of hyperoxia. This might account for the (non-significant) finding of a tendency
towards elevated CBF in frontal regions in the patient group compared to the control
group, reported in the current study and at odds with the prevailing literature.
Investigation into the change in resting CBF with hyperoxia in cSVD compared to age-
matched controls is necessary to confirm such a supposition.
Although „n‟ was surprisingly not found to be correlated with CBF in the current study,
the differences in CBF might have accounted for some of the observed difference in „n‟.
The results of the current study must be viewed with caution until the effect of
hyperoxia on CBF in patients with cSVD is elucidated by further research.
Mark et al. (210) have identified a further possible limitation with the oxygen
calibration technique designed by Chiarelli et al. (170), and used in this study.
Hyperoxia was fixed rather than altered in response to end-tidal CO2. Thus, the
reduction in CBF ought not to be assumed to be constant. This might introduce bias in
to the calculation of „M‟.
„M‟ was found to be increased in patients compared to controls. „M‟ is a local
parameter used to calculate CMRO2, simplistically described as the theoretical
maximum BOLD signal change (159). It is related to baseline CBV and OEF (159): a
reduction in OEF or CBV correlates with a reduction in „M‟. This study found„M‟ to be
87
inversely correlated with „n‟ (p = 0.017; Figure 32). Low values of „n‟ were associated
with higher values of „M‟. A similar study by Tak et al. (211) supports these findings:
in a study of six patients with subcortical vascular dementia compared with six controls,
using both fMRI (not calibrated) and NIRS, BOLD, CBF and CMRO2 were all found to
be reduced whilst OEF was increased.
The study by Tak et al. (211) had fewer patients than the current study and took cortical
metabolic measurements in a disease group with predominantly subcortical disease.
The fMRI was based on a motor rather than executive task. Furthermore, one patient
had CT rather than fMRI. Despite these limitations, the results of the two studies are in
concordance with one another. Tak et al. (211) postulate that in patients with cSVD
there is a rise in OEF to compensate for reductions in CBF secondary to microvascular
changes: neurovascular un-coupling is associated with a rise in OEF. Although the
current study did not find a drop in CBF or CMRO2, it did find an association between a
reduction in „n‟ and an increase in „M‟.
Ischaemic damage of the white matter associated with cSVD is thought to possibly lead
to disconnection of the cerebral cortex from subcortical structures (10,105,201,203).
The subcortical damage might result in reduced CBF and CMRO2 in the cerebral cortex,
which in turn might lead to cognitive impairment and dementia (203). Neither CBF nor
CMRO2 were significantly reduced in the patient group in the current study, however.
The sample size was too small to detect a significant difference in cerebral volumes
between patient and control groups, but did show a tendency towards a correlation with
DSST score for both ventricular enlargement and subcortical matter volume. Cerebral
atrophy and ventricular enlargement are strongly correlated with cognitive dysfunction
(96,109,117,121,126,212-215), but are not specific to cSVD (216-218). Thus, the
findings in the literature are mirrored by the findings in the current study: normalised
brain structure volumes were unable to differentiate participants with cSVD from those
without, but did tend to correlate with a measure of cognitive dysfunction (DSST score).
A large long term longitudinal study with strict entry criteria would be necessary to
88
determine a temporal relationship between cerebral atrophy and cognition. Due to its
cross-sectional nature, the current study cannot provide evidence as to whether atrophy
leads to cognitive decline or cognitive decline occurs before atrophy.
WML score was significantly higher in the patient group in frontal, basal ganglia and
total regions and was shown to correlate with DSST score in frontal and total regions.
The evidence in the literature for the correlation between WML load and cognitive
dysfunction is mixed as observed in the systematic review (chapter 2). The result of the
current study might be interpreted as a direct link between WML and cognitive
dysfunction. However, the evidence also fits the „threshold hypothesis‟ (116), which
states that WML only affect cognition when they are sufficiently severe. If the two
most severe WML scores were to be removed from the analysis of the current study of
total WML score, the correlation between total WML score and DSST score disappears
(r2 = 0.055, p = 0.26).
WML score and „n‟ were both significantly different between patient and control
groups. Regression analyses were employed to test for a relationship between WML
score and „n‟. No correlation was observed, which might suggest that WML are not
directly linked to altered NVC.
6.2 Strengths of the study
This study had four particular positive features. The main strength of this study of
cSVD was the application of strict entry criteria. A homogeneous group was selected
that allows the findings to be confidently attributed to cSVD. Despite the strict criteria,
it is possible that included patients might have had LACS attributable to an embolic
source. However, a recent study of 2875 persons with a first-ever anterior circulation
stroke showed an athero-embolic cause to be far less likely in LACS than cortical
strokes (219). Clinical diagnosis alone is unreliable: 16% of 44 subjects with a clinical
LACS were found to have a cortical stroke on imaging and 23% of 93 patients with a
cortical syndrome had an acute lacunar infarct (220). The combination of a clinical
89
diagnosis of LACS coupled with a radiological diagnosis further reduces the likelihood
of an embolic cause and is an approach used by leading researchers in the field
(44,185). Brain imaging shows evidence of other markers associated with cSVD,
including EPVS, microbleeds and LA, but have yet been shown to be specific to the
condition, rather markers of cerebral damage (44).
Secondly, neuropsychological tests were used that were known to be sensitive to the
changes in executive function associated with the disease. Thirdly, the patients were
compared to age- and education-matched controls with similar rates of hypertension.
Finally, both a novel imaging marker and more traditional structural imaging markers
were compared in the same groups against the same neuropsychological tests.
6.3 Limitations of the study
There were many limitations of this study, some particular to the methodology and
some associated with the difficulties in the investigation of cSVD. The overwhelming
weakness was the inability to obtain the desired number of patients. Fourteen were
recruited, of whom ten provided data for the calculation of „n‟. Furthermore, just less
than 10% of the data had to be excluded due to noise. It was calculated that a minimum
of 16 patients would be needed to detect a 25% change in „n‟. It is clear that
associations with „n‟ might have been missed due to under-powering of the study.
Considerable effort was made to encourage recruitment. The study was adopted by the
SRN which facilitated access to a larger pool of potential patients. However, the
primary strength of the study was the predominant cause of the primary weakness.
Stringent entry criteria, necessary to establish a meaningful study group, greatly limited
the number of suitable patients. The study offered no direct benefits to the individual
patients. Moreover, it involved a long time in the confined space of the MRI scanner,
which many individuals would find disquieting.
There were also limitations in the selection of controls. Patient and control groups were
not well-matched for gender, although the difference was not statistically significant (p
90
= 0.086). That was a result of insufficient patient recruitment. Fewer females with
cSVD participated in the study. There is some evidence to suggest that gender may
affect the components of the haemodynamic response (221-225), albeit inconclusive. In
response to a visual task, BOLD amplitudes have been found to be significantly higher
(222) and also lower (223) in women compared to men.
Many of the controls were selected from a university population. They might be better
educated and thus likely to perform better on cognitive tests. The number of years of
education was recorded for each participant and overall did not differ significantly
between control and patient groups. However, the National Adult Reading Test
(NART) might have offered a better measurement of pre-morbid intelligence to account
for any bias in the cognitive tests arising from differences in pre-morbid intelligence
(226).
Furthermore, subjects with cortical stroke might have provided a better control group
for radiologically-confirmed LACS (44). Cardiovascular risk factors such as
hypertension and diabetes, as well as, for example, antihypertensive medications would
be better balanced between the groups.
The measurement of cerebral volumes was not performed blind to the participant‟s
group (patient or control). Differences noted in these measures may have been subject
to an observer bias. Measures of WML and microbleeds were performed blind. It is
possible that the appearance of morphological features associated with cSVD (WML
lacunes and microbleeds) may have inadvertently influenced measurements. However,
those morphological features associated with cSVD have also been noted in the elderly
and so might be expected in the controls (1,14,32).
Venous oxygen concentration was calculated based upon end-tidal oxygen
concentration, with an assumption that oxygen exchange across the vessel wall is
similar in both groups and that the resting state OEF is the same in both groups.
However, there is evidence that the OEF might in fact be increased in VaD (201,203)
91
and silent brain infarction (227), which might confound the findings. The significant
rise in „M‟ suggests that OEF might be increased in patients with cSVD in this study
and thus be a serious confounding factor, although the findings are supported by Tak et
al. (211). A method to directly sample venous oxygen concentration might remove this
potential bias.
This is the first study to try to investigate „n‟ using oxygen-calibrated fMRI in patients
with cSVD. Confounding factors such as CBF response to hyperoxia, as discussed
above, might introduce error into the results.
Ideally, the study would have been able to directly compare DTI measures to fMRI and
structural findings in this well-selected cSVD population. However, inclusion of DTI
would have increased scan time to an unreasonable length.
The quality of the MRI images was also limited by the time spent in the scanner. Had
participants been scanned over two sessions of an hour each, better quality images
might have been made. However, this was considered too expensive and time-
consuming for the current study.
fMRI is particularly sensitive to motion, caused by participants moving their head and
also internal pulsations associated with respiratory and cardiac cycles (161). The
interference from motion is limited by keeping the head still with pads and by motion
correction, which uses automated algorithms to realign brain volumes (161). Motion
prevented the reading of a SWI of one patient to determine their microbleed score. It is
feasible that motion might have had an effect on the fMRI data.
6.3.1 Limitations of studying cSVD
The necessity of strict entry criteria is a result of the difficulties associated with
identifying a group of patients with cSVD. These inclusion and exclusion criteria allow
selection of patients with a high likelihood of cSVD, but there is no practical „gold
standard‟ against which to confirm the diagnosis of cSVD. Routine post mortem
92
examination is insufficient as it does not include the exacting process of sectioning the
small penetrating cerebral vessels. Furthermore, the interval from the onset of lesions to
the time of autopsy, coupled with the insidious onset of cognitive symptoms creates
difficulties in associating structural lesions to the clinical picture (8).
Related to the difficulty in diagnosing cSVD is the charting of its progression, which
can be estimated by clinical measures of severity or morphological changes such as
WML progression. However, as previously demonstrated, there is no consistent
correlation between these two methods of assessing progression, that is clinical and
observable lesion changes. This study used neuropsychological test scores as a standard
against which to assess the severity of cSVD, which have been shown to discriminate
patients with cSVD from healthy age- and education- matched controls, with a
sensitivity and specificity both of 88% (9).
A further limitation to the well-selected cSVD study population used is that the findings
cannot be transferred to other populations, including individuals with evidence of cSVD
but who have not had a stroke, in whom preventative action might be particularly
beneficial. However, the findings of this and similar studies might be used to formulate
hypotheses in broader populations, such as the general elderly.
Despite the difficulties involved in the study of cSVD, it represents an exciting area of
research, with the potential to influence clinical care. A few ideas for further study are
suggested in Box 7.
93
Box 7: Ideas for further study.
Validation of results in studies with a larger sample size.
Validate the oxygen-calibration technique in patient groups.
Cross-sectional studies, such as the current study, do not elucidate causation.
More robust conclusions might be made if the patients were followed up over
several years in a longitudinal study.
A longitudinal study with eventual pathological confirmation of cSVD.
Direct DTI comparison with „n‟ in a cSVD population.
Small vessel disease of the brain might be related to small vessel disease of
other organs, e.g. the kidneys, and as such might represent a systemic
disorder (206).
Apply the study methodology to assess changes in NVC in different
populations:
o VaD
o Community-dwelling elderly
o Individuals with evidence of cSVD, but no clinical stroke
A large series of RCTs involving patients with cSVD to elicit appropriate
management, including the benefits of:
o Aspirin
o Anti-hypertensives
o Statins
o Intravenous magnesium in the acute setting
o Donepezil
94
Chapter seven: Conclusions
A reduction in „n‟ was observed in patients with likely-cSVD compared to age-, gender-
and education-matched controls using hyperoxia calibrated fMRI. These disparate
values of „n‟ signify altered NVC in patients with cSVD. The finding of altered NVC
indicates that BOLD signal change does not represent a good imaging marker for cSVD
as it does not accurately reflect neural activity in patients with this disease. „n‟ was
inversely correlated with „M‟, which might indirectly suggest that altered NVC is
associated with a raised OEF and loss of metabolic reserve. This is one of the first
studies to investigate „n‟ in cSVD and therefore these preliminary results, based upon a
small sample size, require validation in larger studies.
95
Appendix A: Summary of studies reviewed in chapter 2
Key:
* = Significant difference at p<0.05
N= Number; A= Mean age in years; E= Years of education in years; F= Percentage
females; HT= Percentage hypertensive; LI= Lacunar infarct (visible on MRI); LACS=
Lacunar syndrome (clinical diagnosis); LA= Leukoaraiosis (White matter
hyperintensities on MRI); CeVD= Cerebrovascular disease; VaD= Vascular dementia
Imaging modalities: CT= Computed Tomography; MRI= Magnetic Resonance
Imaging; MRS= Magnetic Resonance Spectroscopy; DWI= Diffusion Weighted
Imaging; DTI= Diffusion Tensor Imaging; PET= Positron Emission Tomography;
SPECT= Single Photon Emission Computed Tomography; fNIRS= Functional Near-
Infrared Spectroscopy
NAWM= Normal appearing white matter; FA= Fractional anisotropy; MD= Mean
diffusivity; NAA/Cr= N-acetylaspartate/ creatine ratio
Neuropsychological tests: EXEC= Executive (test); MMSE= Mini-Mental State
Examination; CDR= Clinical Dementia Rating scale; WAIS-R= Wechsler Adult
Intelligence Scale-Revised; TMT= Trail making Test; WCST= Wisconsin Card Sorting
Test; DRS= Dementia Rating Scale; MDRS I/P= Mattis Dementia Rating Scale-
Initiation Perseveration; BADS= Behavioural Assessment of the Dysexecutive
Syndrome; NART-R= National Adult Reading Test-Restandardised; SSS=
Scandinavian Stroke Scale
NINDS= NINDS-AIREN criteria for the diagnosis of vascular dementia
96
97
98
99
Appendix B: Participant consent form
January 2008, version 2
CONSENT FORM
Title of Project:
Cerebral small vessel disease: novel MRI markers and cognitive dysfunction
Name of Researchers:
Dr H Emsley (University of Liverpool and The Walton Centre)
Dr L Parkes (University of Liverpool)
Please initial box
1. I confirm that I have read and understand the information sheet dated January
2008 (version 2) for the above study and have had the opportunity to ask
questions. I agree to comply with the procedures as set out in the information
sheet.
2. I understand that my participation is voluntary and that I am free to withdraw at
any time, without giving any reason, without my medical care or legal rights
being affected.
3. I understand that sections of any of my medical notes may be looked at by
responsible individuals from regulatory authorities where it is relevant to my
taking part in research. I give permission for these individuals to have access to
my records.
4. I agree that personal data relating to myself (as defined by the Data
Protection Act, 1998), being used for research purposes only. I
understand that my personal information will be kept for up to fifteen
years and then will be confidentially destroyed.
5. I consent to my GP being informed of my participation in this study and that
they may be contacted for any relevant medical information as required.
6. I agree to be contacted at six monthly intervals over the next two years, to be
informed of the progress of the study and to discuss any health changes.
100
7. I agree to take part in the above study.
____________________________ ____________ ____________
Name of Participant Date Signature
____________________________ ____________ ____________
Name of Person taking consent Date
Signature
(if different from researcher)
____________________________ ____________ ____________
Researcher Date Signature
1 copy for participant, 1 copy for researcher, 1 copy for hospital notes (if a patient)
101
Appendix C: Heading letter for participants
Mr Guy Lumley
Research Assistant Mr Guy Lumley
MARIARC
Pembroke Place
Liverpool
15th
July 2009 L69 3GE
Dear Mr ,
I write with reference to a trial we are running at the University of Liverpool into
lacunar strokes and wondered if you would consider volunteering.
Please find enclosed details on the trial and what your participation would involve, as
well as directions to the Magnetic Resonance and Image Analysis Research Centre
(MARIARC). If you‟ve any questions, please don‟t hesitate to contact me by letter,
telephone 0151 7954622, or email [email protected].
If happy to participate, you would be asked to come into MARIARC for approximately
3 – 3½ hours. The session would involve a clinical assessment, including a brief
examination. We would test cognitive function (thinking abilities), which includes a
series of complicated computerised tests. These assessments usually take 1 hour in
total. We would also need you to have a detailed MRI scan, which also last about 1
hour.
Prior to scanning you would need to be screened to check that it is safe to scan you.
This involves filling in a form with a qualified radiographer at MARIARC. For the
scan, you would be asked to change into a gown in the changing rooms provided. If
you wish, you are welcome to bring a T-shirt and pyjama bottoms to wear under the
gown, providing they contain no metal. There are lockers available for your personal
belongings.
For part of the scanning procedure you would be asked to wear a mask that allows us to
deliver oxygen-rich air. You would also be asked to do another test inside the scanner.
Yours sincerely,
Guy Lumley (Research Assistant)
102
Appendix D: Patient information sheet
January 2008
Patient Information
CEREBRAL SMALL VESSEL DISEASE:
NOVEL MRI MARKERS & COGNITIVE DYSFUNCTION
You are being invited to take part in a research study. Before you decide it is
important for you to understand why the research is being done and what it will
involve. Please take time to read the following information carefully and
discuss it with others if you wish. Please ask us if there is anything that is not
clear or if you would like more information on. Thank you for reading this.
What is the purpose of the study?
Stroke is an important cause of damage to the brain leading to death or
disability. Around one quarter of strokes are due to „small vessel disease‟. This
type of stroke has been studied less and we know little about what happens to
patients in the longer term after having a small vessel stroke, including what
happens to their cognitive function (or thinking processes). We are also unsure
about what the best scanning tests are for small vessel strokes. We aim to
investigate new and revised scanning tests using magnetic resonance imaging
(MRI) techniques. Gaining information as to whether these different tests can
be related to any possible problems with thinking processes. The whole study
will last for at least 2 years.
Why have I been chosen?
You have been chosen because you have either had a lacunar stroke (a type of
small vessel stroke) or your brain scan shows some changes like those that we
see in this type of stroke.
Do I have to take part?
It is up to you to decide whether or not to take part. If you do decide to take
part you will be given this information sheet to keep and be asked to sign a
consent form. If you decide to take part you are still free to withdraw at any
time and without giving a reason. A decision to withdraw at any time, or a
decision not to take part, will not affect the standard of care you receive now or
in the future.
103
What will happen to me if I take part?
If you have had a stroke we will ask you about this, we will also obtain details
from your medical notes.
We will perform a clinical assessment, including brief examination. We will
also perform an assessment of cognitive function (thinking abilities), including
a range of computerised tests. These assessments would take approximately 1
hour in total. We would also need you to have detailed MRI scans, also lasting
approximately 1 hour.
We would arrange for you to attend the Magnetic Resonance and Image
Analysis Research Centre (MARIARC) at the University of Liverpool, where
the clinical assessment, tests of cognitive function and MRI scans will be
performed. There will be a nurse at MARIARC (but no doctor present) to guide
you through these tests and scans. We are willing to pay for your transport to
and from MARIARC, or organise a taxi if this is preferable. The scans at
MARIARC will be in addition to any other scans that are part of your care.
Prior to scanning you will need to be screened to check that you are safe to be
scanned. This involves filling in a form with a qualified radiographer at
MARIARC. Prior to the scanning procedure, you will be asked to change into a
gown in the changing rooms provided. If you wish you may bring a t-shirt and
pyjama bottoms to wear under the gown, providing it contains no metal. There
will be lockers available for your personal belongings which can be locked
during the period of the study.
For part of the scanning procedure you will be asked to wear a mask that will
allow us to deliver oxygen-rich air. You will also be asked to perform a number
of tasks inside the scanner, such as making judgments about words presented on
a screen.
We would like to contact you again in the future, at 6 monthly intervals by
telephone or in writing, and after two years so that we can arrange to see you
again to repeat the clinical assessment, tests of cognitive function and MRI
scanning so that we can compare these tests with the first tests you had.
What do I have to do?
Participation in this study does not require any additional visits to the hospital,
and only requires you to attend MARIARC for the clinical assessment, tests of
104
cognitive function and MRI scans. No special preparations or changes in
medication or lifestyle are needed.
What are the possible disadvantages and risks of taking part?
There are no particular disadvantages or risks because MRI is generally safe.
The scanner experience is noisy, and has been reported as „claustrophobic‟,
causing some anxiety in a small number of individuals. If you experience any
distress, the scanning process will be immediately discontinued.
The MRI scans will be looked at and reported by a consultant neuroradiologist.
If the neuroradiologist notices something abnormal (in addition to the effects of
any small vessel disease) and if this is considered to be clinically important then
a report will be sent to your General Practitioner (GP). The images collected for
this study are not a typical clinical MRI series and would not however be made
available for diagnostic purposes.
What are the possible benefits of taking part?
There are no direct benefits to you personally from participation in the study.
However we hope that the information we gain from the study will help us to
identify what the best scans are to obtain improved MRI markers of small vessel
stroke, and whether these may relate to possible problems with cognitive
function (thinking processes).
What if new information becomes available?
Sometimes during the course of a research project, relevant information
becomes available. If this happens, a member of the research team will tell you
about it. Also, on receiving new information the research team might consider it
to be in your best interests to pass this information to your own doctor, or even
withdraw you from the study. The research team will explain the reasons and
arrange for your care to continue.
What if something goes wrong?
If you are harmed by taking part in this research project, there are no special
compensation arrangements. If you are harmed due to someone‟s negligence,
then you may have grounds for a legal action but you may have to pay for it.
Regardless of this, if you wish to complain about any aspect of the way you
have been approached or treated during the course of this study, the normal
complaints mechanisms should be available to you.
105
Will my taking part in this study be kept confidential?
All information which is collected about you during the course of this research
will be kept strictly confidential in line with the data Protection Act 1998. Any
information about you which leaves the hospital will have your name and
address removed so that you cannot be recognised from it. Your GP will be
informed of your participation in this study. The results of your tests will be
made available to other doctors in this hospital or other hospitals who may be
directly involved in your future medical care.
What will happen to the results of the research study?
After analysis of the results of all the participants in the study, we aim to
publish the conclusions in a peer-reviewed medical journal. Your individual
results will not be identifiable in any report or publication. You will however be
able to obtain a copy of any published report by contacting us in the future.
Who is organising and funding the research?
The research is organised by Dr Emsley (University of Liverpool and The
Walton Centre) and Dr Parkes (University of Liverpool). None of the research
team receives any direct payment for this research. The research is being funded
through financial support from Unilever. It is possible that additional or
alternative sources of funding will be identified during the course of the
research.
Who has reviewed the study?
The Local Research Ethics Committee has reviewed this study.
Contact for further information
We suggest you contact Dr Parkes if you require further information. She is
contactable on 0161 2755577. Thank you for taking part in this study.
You will be given a copy of this Participant Information Sheet and a signed
consent form to keep.
Patient Information number for this trial:
106
Appendix E: Presentations resulting from this work
1. Parkes LM, Lumley G, Mohtasib RS, Emsley H, Goodwin JA „Calibrated fMRI
reveals altered neurovascular coupling with age during a cognitive Stroop task‟
ISMRM Honolulu [2009].
2. Goodwin JA, Lumley G, Irwin A, Emsley H, Parkes LM „Hyperoxia calibrated
fMRI of cerebral small vessel disease during a cognitive Stroop task‟ ISMRM
Honolulu [2009].
3. Lumley G, Goodwin J, Parkes LM, Emsley HCA. Executive and frontal
dysfunction in an fMRI study of cerebral small vessel disease; ABN
(Association of British Neurologists) conference Liverpool [2009]. To be
published in abstract form in JNNP (Journal of Neurology, Neurosurgery and
Psychiatry). [Oral presentation by the first author].
4. Lumley G, Goodwin J, Parkes LM, Emsley HCA. Executive and frontal
dysfunction in an fMRI study of cerebral small vessel disease; North England
Neurological Association annual conference Harrogate [2009]. [Oral
presentation by the first author].
5. Lumley G, Goodwin J, Parkes LM, Emsley HCA. Executive and frontal
dysfunction in an fMRI study of cerebral small vessel disease; UK Stroke Forum
conference Glasgow [2009]. [Oral presentation by the first author].
107
Glossary
ASL Arterial spin labelling
BBB Blood-brain barrier
BOLD Blood-oxygen-level-dependent
BP Blood pressure
CADASIL Cerebral autosomal dominant arteriopathy with subcortical infarcts and
leukoencephalopathy
CBF Cerebral blood flow
CBV Cerebral blood volume
Cho Choline
CDR Clinical Dementia Rating
CHS Cardiovascular health study
CMA Cerebral microangiopathy
CMRO2 Metabolic rate of oxygen
COPD Chronic obstructive pulmonary disease
Cr Creatine
CSF Cerebrospinal fluid
cSVD Cerebral small vessel disease
CT Computed tomography
DSST Digit symbol substitution test
DTI Diffusion tensor imaging
DWI Diffusion-weighted imaging
EEG Electroencephalography
EPVS Enlarged perivascular space
FLAIR Fluid attenuated inversion recovery
fMRI Functional magnetic resonance imaging
(f)NIRS (Functional) near-infrared spectroscopy
FWHM Full width at half maximum
GABA Gamma-aminobutyric acid
GLM General linear model
HDL High-density lipoprotein
108
LA Leukoaraiosis
LDL Low-density lipoprotein
LI Lacunar infarct
LILACS Liverpool lacunar stroke study
LIPL Left inferior parietal lobule
(L/R)MeFG (Left/ right) medial frontal gyrus
MARIARC Magnetic resonance and image analysis research centre
MDRS-I/P Mattis dementia rating scale-initiation/ perseveration subscale
MEG Magnetoencephalography
MiFG Middle frontal gyrus
MMSE Mini-mental state examination
MRA Magnetic resonance angiography
MRI Magnetic resonance imaging
MRS Magnetic resonance spectroscopy
NAA N-acetylaspartate
NIHSS National Institute of Health Stroke Scale
NO Nitric oxide
NVC Neurovascular coupling
NVU Neurovascular unit
PCG Precentral gyrus
PET Positron emission tomography
PL Parietal lobe
PROGRESS Perindopril protection against recurrent stroke study
PROSPER Pravastatin in elderly individuals at risk of vascular disease (study)
RCT Randomised controlled trial
RF Radiofrequency
ROI Region of interest
SNR Signal to noise ratio
SPECT Single photon emission computed tomography
SRN Stroke Research Network
SWI Susceptibility-weighted imaging
109
Syst-Eur Systolic hypertension in Europe (study)
TE Echo time
TI Inversion time
TR Repetition time
UHA University Hospital Aintree
VaD Vascular dementia
VMR Vasomotor reactivity
WML White matter lesion
110
References
(1) Norrving B. Lacunar infarcts: no black holes in the brain are benign. Pract Neurol
2008 Aug;8(4):222-228.
(2) Norrving B. Long-term prognosis after lacunar infarction. Lancet Neurol 2003
Apr;2(4):238-245.
(3) Arauz A, Murillo L, Cantu C, Barinagarrementeria F, Higuera J. Prospective study
of single and multiple lacunar infarcts using magnetic resonance imaging: risk factors,
recurrence, and outcome in 175 consecutive cases. Stroke 2003 Oct;34(10):2453-2458.
(4) Wardlaw JM, Sandercock PA, Dennis MS, Starr J. Is breakdown of the blood-brain
barrier responsible for lacunar stroke, leukoaraiosis, and dementia? Stroke 2003
Mar;34(3):806-812.
(5) Nitkunan A, Barrick TR, Charlton RA, Clark CA, Markus HS. Multimodal MRI in
cerebral small vessel disease: its relationship with cognition and sensitivity to change
over time. Stroke 2008 Jul;39(7):1999-2005.
(6) McManus J, Stott DJ. Small-vessel cerebrovascular disease in older people. Reviews
in Clinical Gerontology 2005 Aug;15(3-4):207.
(7) Roman GC, Erkinjuntti T, Wallin A, Pantoni L, Chui HC. Subcortical ischaemic
vascular dementia. Lancet Neurol 2002 Nov;1(7):426-436.
(8) Schmidt R, Scheltens P, Erkinjuntti T, et al. White matter lesion progression: a
surrogate endpoint for trials in cerebral small-vessel disease. Neurology 2004 Jul
13;63(1):139-144.
(9) O'Sullivan M, Morris RG, Markus HS. Brief cognitive assessment for patients with
cerebral small vessel disease. J Neurol Neurosurg Psychiatry 2005 Aug;76(8):1140-
1145.
(10) O'Sullivan M, Morris RG, Huckstep B, Jones DK, Williams SC, Markus HS.
Diffusion tensor MRI correlates with executive dysfunction in patients with ischaemic
leukoaraiosis. J Neurol Neurosurg Psychiatry 2004 Mar;75(3):441-447.
(11) Kinoshita T, Okudera T, Tamura H, Ogawa T, Hatazawa J. Assessment of lacunar
hemorrhage associated with hypertensive stroke by echo-planar gradient-echo T2*-
weighted MRI. Stroke 2000 Jul;31(7):1646-1650.
(12) Lammie GA, Brannan F, Wardlaw JM. Incomplete lacunar infarction (Type Ib
lacunes). Acta Neuropathol 1998 Aug;96(2):163-171.
111
(13) Roman GC. From UBOs to Binswanger's disease. Impact of magnetic resonance
imaging on vascular dementia research. Stroke 1996 Aug;27(8):1269-1273.
(14) Longstreth WT,Jr, Manolio TA, Arnold A, et al. Clinical correlates of white matter
findings on cranial magnetic resonance imaging of 3301 elderly people. The
Cardiovascular Health Study. Stroke 1996 Aug;27(8):1274-1282.
(15) Pantoni L, Garcia JH. The significance of cerebral white matter abnormalities 100
years after Binswanger's report. A review. Stroke 1995 Jul;26(7):1293-1301.
(16) Longstreth WT,Jr, Arnold AM, Beauchamp NJ, et al. Incidence, manifestations,
and predictors of worsening white matter on serial cranial magnetic resonance imaging
in the elderly: the Cardiovascular Health Study. Stroke 2005 Jan;36(1):56-61.
(17) Kuller LH, Arnold AM, Longstreth WT, et al. White matter grade and ventricular
volume on brain MRI as markers of longevity in the cardiovascular health study.
Neurobiol Aging 2007 Sep;28(9):1307-1315.
(18) Bowler JV. Vascular cognitive impairment. J Neurol Neurosurg Psychiatry 2005
Dec;76 Suppl 5:v35-44.
(19) Chillon J-M and Baumbach G. Autoregulation of cerebral blood flow. In: Welch k,
Caplan L, Reis D, Siesjo B, Weir B, editor. Primer on cerebrovascular diseases. First ed.
California: Academic Press; 1997. p. 51-54.
(20) Pantoni L, Garcia JH. Pathogenesis of leukoaraiosis: a review. Stroke 1997
Mar;28(3):652-659.
(21) Birns J, Markus H, Kalra L. Blood pressure reduction for vascular risk: is there a
price to be paid? Stroke 2005 Jun;36(6):1308-1313.
(22) Markus HS, Lythgoe DJ, Ostegaard L, O'Sullivan M, Williams SC. Reduced
cerebral blood flow in white matter in ischaemic leukoaraiosis demonstrated using
quantitative exogenous contrast based perfusion MRI. J Neurol Neurosurg Psychiatry
2000 Jul;69(1):48-53.
(23) Isaka Y, Okamoto M, Ashida K, Imaizumi M. Decreased cerebrovascular dilatory
capacity in subjects with asymptomatic periventricular hyperintensities. Stroke 1994
Feb;25(2):375-381.
(24) Marstrand JR, Garde E, Rostrup E, et al. Cerebral perfusion and cerebrovascular
reactivity are reduced in white matter hyperintensities. Stroke 2002 Apr;33(4):972-976.
(25) Bakker SL, de Leeuw FE, de Groot JC, Hofman A, Koudstaal PJ, Breteler MM.
Cerebral vasomotor reactivity and cerebral white matter lesions in the elderly.
Neurology 1999 Feb;52(3):578-583.
112
(26) Terborg C, Gora F, Weiller C, Rother J. Reduced vasomotor reactivity in cerebral
microangiopathy : a study with near-infrared spectroscopy and transcranial Doppler
sonography. Stroke 2000 Apr;31(4):924-929.
(27) Hassan A, Hunt BJ, O'Sullivan M, et al. Markers of endothelial dysfunction in
lacunar infarction and ischaemic leukoaraiosis. Brain 2003 Feb;126(Pt 2):424-432.
(28) Lin JX, Tomimoto H, Akiguchi I, et al. Vascular cell components of the medullary
arteries in Binswanger's disease brains: a morphometric and immunoelectron
microscopic study. Stroke 2000 Aug;31(8):1838-1842.
(29) Tomimoto H, Akiguchi I, Suenaga T, et al. Alterations of the blood-brain barrier
and glial cells in white-matter lesions in cerebrovascular and Alzheimer's disease
patients. Stroke 1996 Nov;27(11):2069-2074.
(30) Stevenson SF, Doubal FN, Shuler K, Wardlaw JM. A systematic review of
dynamic cerebral and peripheral endothelial function in lacunar stroke versus controls.
Stroke 2010 Jun;41(6):e434-42.
(31) Werring DJ, Frazer DW, Coward LJ, et al. Cognitive dysfunction in patients with
cerebral microbleeds on T2*-weighted gradient-echo MRI. Brain 2004 Oct;127(Pt
10):2265-2275.
(32) Greenberg SM, Vernooij MW, Cordonnier C, et al. Cerebral microbleeds: a guide
to detection and interpretation. Lancet Neurol 2009 Feb;8(2):165-174.
(33) Fisher CM. Lacunar strokes and infarcts: a review. Neurology 1982
Aug;32(8):871-876.
(34) Dichgans M. CADASIL: a monogenic condition causing stroke and subcortical
vascular dementia. Cerebrovasc Dis 2002;13 Suppl 2:37-41.
(35) Wardlaw JM. What is a lacune? Stroke 2008 Nov;39(11):2921-2922.
(36) Poirier J, Derouesne C. Cerebral lacunae. A proposed new classification. Clin
Neuropathol 1984 Nov-Dec;3(6):266.
(37) Lastilla M. Lacunar infarct. Clin Exp Hypertens 2006 Apr-May;28(3-4):205-215.
(38) Arboix A, Marti-Vilalta JL, Garcia JH. Clinical study of 227 patients with lacunar
infarcts. Stroke 1990 Jun;21(6):842-847.
(39) Norrving B. Lacunar infarction: embolism is the key: against. Stroke 2004
Jul;35(7):1779-1780.
113
(40) Gorman MJ, Dafer R, Levine SR. Ataxic hemiparesis: critical appraisal of a
lacunar syndrome. Stroke 1998 Dec;29(12):2549-2555.
(41) Vermeer SE, Longstreth WT,Jr, Koudstaal PJ. Silent brain infarcts: a systematic
review. Lancet Neurol 2007 Jul;6(7):611-619.
(42) Elkins JS, Longstreth WT,Jr, Manolio TA, Newman AB, Bhadelia RA, Johnston
SC. Education and the cognitive decline associated with MRI-defined brain infarct.
Neurology 2006 Aug 8;67(3):435-440.
(43) Braffman BH, Zimmerman RA, Trojanowski JQ, Gonatas NK, Hickey WF,
Schlaepfer WW. Brain MR: pathologic correlation with gross and histopathology. 1.
Lacunar infarction and Virchow-Robin spaces. AJR Am J Roentgenol 1988
Sep;151(3):551-558.
(44) Wardlaw JM. Blood-brain barrier and cerebral small vessel disease. J Neurol Sci
2010 Dec;299(1-2):66.
(45) Maclullich AM, Wardlaw JM, Ferguson KJ, Starr JM, Seckl JR, Deary IJ. Enlarged
perivascular spaces are associated with cognitive function in healthy elderly men. J
Neurol Neurosurg Psychiatry 2004 Nov;75(11):1519-1523.
(46) Wuerfel J, Haertle M, Waiczies H, et al. Perivascular spaces--MRI marker of
inflammatory activity in the brain? Brain 2008 Sep;131(Pt 9):2332-2340.
(47) Doubal FN, MacLullich AM, Ferguson KJ, Dennis MS, Wardlaw JM. Enlarged
perivascular spaces on MRI are a feature of cerebral small vessel disease. Stroke 2010
Mar;41(3):450-454.
(48) Wardlaw JM, Doubal F, Armitage P, et al. Lacunar stroke is associated with diffuse
blood-brain barrier dysfunction. Ann Neurol 2009 Feb;65(2):194-202.
(49) Schroeter ML, Cutini S, Wahl MM, Scheid R, Yves von Cramon D. Neurovascular
coupling is impaired in cerebral microangiopathy--An event-related Stroop study.
Neuroimage 2007 Jan 1;34(1):26-34.
(50) O'Sullivan M, Barrick TR, Morris RG, Clark CA, Markus HS. Damage within a
network of white matter regions underlies executive dysfunction in CADASIL.
Neurology 2005 Nov 22;65(10):1584-1590.
(51) Schmidtke K, Hull M. Cerebral small vessel disease: how does it progress? J
Neurol Sci 2005 Mar 15;229-230:13-20.
(52) Schmidt R, Enzinger C, Ropele S, Schmidt H, Fazekas F. Subcortical vascular
cognitive impairment: similarities and differences with multiple sclerosis. J Neurol Sci
2006 Jun 15;245(1-2):3-7.
114
(53) Markus HS. Mild cognitive impairment after lacunar infarction: voxel-based
morphometry and neuropsychological assessment. Cerebrovasc Dis 2007 May;23(5-
6):323-324.
(54) Mok VC, Wong A, Lam WW, Baum LW, Ng HK, Wong L. A case-controlled
study of cognitive progression in Chinese lacunar stroke patients. Clin Neurol
Neurosurg 2008 Jul;110(7):649-656.
(55) Shim YS, Yang DW, Kim BS, Shon YM, Chung YA. Comparison of regional
cerebral blood flow in two subsets of subcortical ischemic vascular dementia: statistical
parametric mapping analysis of SPECT. J Neurol Sci 2006 Dec 1;250(1-2):85-91.
(56) Girouard H, Iadecola C. Neurovascular coupling in the normal brain and in
hypertension, stroke, and Alzheimer disease. J Appl Physiol 2006 Jan;100(1):328-335.
(57) Patankar TF, Mitra D, Varma A, Snowden J, Neary D, Jackson A. Dilatation of the
Virchow-Robin space is a sensitive indicator of cerebral microvascular disease: study in
elderly patients with dementia. AJNR Am J Neuroradiol 2005 Jun-Jul;26(6):1512-1520.
(58) Farrall AJ, Wardlaw JM. Blood-brain barrier: ageing and microvascular disease -
systematic review and meta-analysis. Neurobiol Aging 2009 Mar;30(3):337-352.
(59) Jackson C, Sudlow C. Are lacunar strokes really different? A systematic review of
differences in risk factor profiles between lacunar and nonlacunar infarcts. Stroke 2005
Apr;36(4):891-901.
(60) Sironi L, Guerrini U, Tremoli E, et al. Analysis of pathological events at the onset
of brain damage in stroke-prone rats: a proteomics and magnetic resonance imaging
approach. J Neurosci Res 2004 Oct 1;78(1):115-122.
(61) Serena J, Leira R, Castillo J, Pumar JM, Castellanos M, Davalos A. Neurological
deterioration in acute lacunar infarctions: the role of excitatory and inhibitory
neurotransmitters. Stroke 2001 May;32(5):1154-1161.
(62) Castellanos M, Castillo J, Garcia MM, et al. Inflammation-mediated damage in
progressing lacunar infarctions: a potential therapeutic target. Stroke 2002
Apr;33(4):982-987.
(63) Audebert HJ, Pellkofer TS, Wimmer ML, Haberl RL. Progression in lacunar stroke
is related to elevated acute phase parameters. Eur Neurol 2004 Feb;51(3):125-131.
(64) Skoog I, Lernfelt B, Landahl S, et al. 15-Year Longitudinal Study of Blood
Pressure and Dementia. Lancet 1996 Apr 27;347(9009):1141-1145.
115
(65) Fornage M, Chiang YA, O'Meara ES, et al. Biomarkers of Inflammation and MRI-
Defined Small Vessel Disease of the Brain: The Cardiovascular Health Study. Stroke
2008 Jul;39(7):1952-1959.
(66) Schmidt R, Schmidt H, Curb JD, Masaki K, White LR, Launer LJ. Early
inflammation and dementia: a 25-year follow-up of the Honolulu-Asia Aging Study.
Ann Neurol 2002 Aug;52(2):168-174.
(67) Hara K, Shiga A, Fukutake T, et al. Association of HTRA1 mutations and familial
ischemic cerebral small-vessel disease. N Engl J Med 2009 Apr 23;360(17):1729-1739.
(68) Gould DB, Phalan FC, van Mil SE, et al. Role of COL4A1 in small-vessel disease
and hemorrhagic stroke. N Engl J Med 2006 Apr 6;354(14):1489-1496.
(69) Forette F, Seux ML, Staessen JA, et al. Prevention of dementia in randomised
double-blind placebo-controlled Systolic Hypertension in Europe (Syst-Eur) trial.
Lancet 1998 Oct 24;352(9137):1347-1351.
(70) Tzourio C, Anderson C, Chapman N, et al. Effects of blood pressure lowering with
perindopril and indapamide therapy on dementia and cognitive decline in patients with
cerebrovascular disease. Arch Intern Med 2003 May 12;163(9):1069-1075.
(71) PROGRESS Collaborative Group. Randomised trial of a perindopril-based blood-
pressure-lowering regimen among 6,105 individuals with previous stroke or transient
ischaemic attack. Lancet 2001 Sep 29;358(9287):1033-1041.
(72) Hasegawa Y, Yamaguchi T, Omae T, Woodward M, Chalmers J, Investigators P.
Effects of perindopril-based blood pressure lowering and of patient characteristics on
the progression of silent brain infarct: the Perindopril Protection against Recurrent
Stroke Study (PROGRESS) CT Substudy in Japan. Hypertension research : official
journal of the Japanese Society of Hypertension 2004 Mar;27(3):147-156.
(73) McGuinness B, Todd S, Passmore P, Bullock R. The effects of blood pressure
lowering on development of cognitive impairment and dementia in patients without
apparent prior cerebrovascular disease. Cochrane Database Syst Rev 2006 Apr
19;(2)(2):CD004034.
(74) Perry H, Davis B, Price T, et al. Effect of treating isolated systolic hypertension on
the risk of developing various types and subtypes of stroke: the Systolic Hypertension in
the Elderly Program (SHEP). JAMA : the journal of the American Medical Association
2000 Jul;284(4):465-471.
(75) Walters M, Muir S, Shah I, Lees K. Perindopril improves cerebral vasomotor
reactivity in patients with lacunar infarction. Stroke 2004 Aug;35(8):1899.
116
(76) Rossini PM, Altamura C, Ferretti A, et al. Does cerebrovascular disease affect the
coupling between neuronal activity and local haemodynamics? Brain 2004 Jan;127(Pt
1):99-110.
(77) Williams PS, Spector A, Orrell M, Rands G. Aspirin for vascular dementia.
Cochrane Database Syst Rev 2000;(2)(2):CD001296.
(78) Kronmal RA, Hart RG, Manolio TA, Talbert RL, Beauchamp NJ, Newman A.
Aspirin use and incident stroke in the cardiovascular health study. CHS Collaborative
Research Group. Stroke 1998 May;29(5):887-894.
(79) Bernick C, Katz R, Smith NL, et al. Statins and cognitive function in the elderly:
the Cardiovascular Health Study. Neurology 2005 Nov 8;65(9):1388-1394.
(80) Rea TD, Breitner JC, Psaty BM, et al. Statin use and the risk of incident dementia:
the Cardiovascular Health Study. Arch Neurol 2005 Jul;62(7):1047-1051.
(81) Shepherd J. A prospective study of pravastatin in the elderly at risk: new hope for
older persons. Am J Geriatr Cardiol 2004 May-Jun;13(3 Suppl 1):17-24.
(82) Lavallee P, Gongora-Rivera F, Jaramillo A, et al. The Lacunar Brain Infarction
Cerebral Hypereactivity and Atorvastatin Trial (LACUNAR-B.I.C.H.A.T.).
Cerebrovasc Dis 2007;23(Suppl 2):37.
(83) Jick H, Zornberg GL, Jick SS, Seshadri S, Drachman DA. Statins and the risk of
dementia. Lancet 2000 Nov 11;356(9242):1627-1631.
(84) Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of
cholesterol lowering with simvastatin in 20,536 high-risk individuals: a randomised
placebo-controlled trial. Lancet 2002 Jul 6;360(9326):7-22.
(85) Cocho D, Belvis R, Marti-Fabregas J, et al. Does thrombolysis benefit patients with
lacunar syndrome? Eur Neurol 2006 Mar;55(2):70-73.
(86) Albers GW, Amarenco P, Easton JD, Sacco RL, Teal P, American College of
Chest Physicians. Antithrombotic and thrombolytic therapy for ischemic stroke:
American College of Chest Physicians Evidence-Based Clinical Practice Guidelines
(8th Edition). Chest 2008 Jun;133(6 Suppl):630S-669S.
(87) Aslanyan S, Weir CJ, Muir KW, Lees KR, IMAGES Study Investigators.
Magnesium for treatment of acute lacunar stroke syndromes: further analysis of the
IMAGES trial. Stroke 2007 Apr;38(4):1269-1273.
(88) Clare L, Woods B. Cognitive rehabilitation and cognitive training for early-stage
Alzheimer's disease and vascular dementia. Cochrane Database Syst Rev
2003;CD003260(4).
117
(89) McShane R, Areosa Sastre A, Minakaran N. Memantine for dementia. Cochrane
Database Syst Rev 2006 Apr 19;(2)(2):CD003154.
(90) Craig D, Birks J. Galantamine for vascular cognitive impairment. Cochrane
Database Syst Rev 2006 Jan 25;(1)(1):CD004746.
(91) Malouf R, Birks J. Donepezil for vascular cognitive impairment. Cochrane
Database Syst Rev 2004;(1)(1):CD004395.
(92) Lopez-Arrieta JM, Birks J. Nimodipine for primary degenerative, mixed and
vascular dementia. Cochrane Database Syst Rev 2002;(3)(3):CD000147.
(93) Moretti R, Torre P, Antonello RM, et al. Rivastigmine superior to aspirin plus
nimodipine in subcortical vascular dementia: an open, 16-month, comparative study. Int
J Clin Pract 2004 Apr;58(4):346-353.
(94) Jackson C, Sudlow C. Comparing risks of death and recurrent vascular events
between lacunar and non-lacunar infarction. Brain 2005 Nov;128(Pt 11):2507-2517.
(95) Samuelsson M, Soderfeldt B, Olsson GB. Functional outcome in patients with
lacunar infarction. Stroke 1996 May;27(5):842-846.
(96) Appelros P, Samuelsson M, Lindell D. Lacunar infarcts: functional and cognitive
outcomes at five years in relation to MRI findings. Cerebrovasc Dis 2005 Jul;20(1):34-
40.
(97) Eriksson SE, Olsson JE. Survival and recurrent strokes in patients with different
subtypes of stroke: a fourteen-year follow-up study. Cerebrovasc Dis 2001
Oct;12(3):171-180.
(98) Staaf G, Lindgren A, Norrving B. Pure motor stroke from presumed lacunar
infarct: long-term prognosis for survival and risk of recurrent stroke. Stroke 2001
Nov;32(11):2592-2596.
(99) Salgado AV, Ferro JM, Gouveia-Oliveira A. Long-term prognosis of first-ever
lacunar strokes. A hospital-based study. Stroke 1996 Apr;27(4):661-666.
(100) Berman KF, Ostrem JL, Randolph C, et al. Physiological activation of a cortical
network during performance of the Wisconsin Card Sorting Test: a positron emission
tomography study. Neuropsychologia 1995 Aug;33(8):1027-1046.
(101) Aharon-Peretz J, Daskovski E, Mashiach T, Tomer R. Natural history of dementia
associated with lacunar infarctions. J Neurol Sci 2002 Nov 15;203-204:53-55.
118
(102) Schroeter ML, Zysset S, Kruggel F, von Cramon DY. Age dependency of the
hemodynamic response as measured by functional near-infrared spectroscopy.
Neuroimage 2003 Jul;19(3):555-564.
(103) Zysset S, Muller K, Lohmann G, von Cramon DY. Color-word matching stroop
task: separating interference and response conflict. Neuroimage 2001 Jan;13(1):29-36.
(104) D'Esposito M, Zarahn E, Aguirre GK, Rypma B. The effect of normal aging on
the coupling of neural activity to the bold hemodynamic response. Neuroimage 1999
Jul;10(1):6-14.
(105) O'Sullivan M, Jones DK, Summers PE, Morris RG, Williams SC, Markus HS.
Evidence for cortical "disconnection" as a mechanism of age-related cognitive decline.
Neurology 2001 Aug 28;57(4):632-638.
(106) Tang Y, Nyengaard JR, Pakkenberg B, Gundersen HJ. Age-induced white matter
changes in the human brain: a stereological investigation. Neurobiol Aging 1997 Nov-
Dec;18(6):609-615.
(107) van Dijk EJ, Prins ND, Vrooman HA, Hofman A, Koudstaal PJ, Breteler MM.
Progression of cerebral small vessel disease in relation to risk factors and cognitive
consequences: Rotterdam Scan study. Stroke 2008 Oct;39(10):2712-2719.
(108) Mungas D, Jagust WJ, Reed BR, et al. MRI predictors of cognition in subcortical
ischemic vascular disease and Alzheimer's disease. Neurology 2001 Dec
26;57(12):2229-2235.
(109) Fein G, Di Sclafani V, Tanabe J, et al. Hippocampal and cortical atrophy predict
dementia in subcortical ischemic vascular disease. Neurology 2000 Dec
12;55(11):1626-1635.
(110) de Groot JC, de Leeuw FE, Oudkerk M, et al. Cerebral white matter lesions and
cognitive function: the Rotterdam Scan Study. Ann Neurol 2000 Feb;47(2):145-151.
(111) Junque C, Pujol J, Vendrell P, et al. Leuko-araiosis on magnetic resonance
imaging and speed of mental processing. Arch Neurol 1990 Feb;47(2):151-156.
(112) Hunt AL, Orrison WW, Yeo RA, et al. Clinical significance of MRI white matter
lesions in the elderly. Neurology 1989 Nov;39(11):1470-1474.
(113) Sabri O, Ringelstein EB, Hellwig D, et al. Neuropsychological impairment
correlates with hypoperfusion and hypometabolism but not with severity of white matter
lesions on MRI in patients with cerebral microangiopathy. Stroke 1999 Mar;30(3):556-
566.
119
(114) Lopez OL, Kuller LH, Becker JT, et al. Classification of vascular dementia in the
Cardiovascular Health Study Cognition Study. Neurology 2005 May 10;64(9):1539-
1547.
(115) Schmand B, Smit JH, Geerlings MI, Lindeboom J. The effects of intelligence and
education on the development of dementia. A test of the brain reserve hypothesis.
Psychol Med 1997 Nov;27(6):1337-1344.
(116) Wen HM, Mok VC, Fan YH, et al. Effect of white matter changes on cognitive
impairment in patients with lacunar infarcts. Stroke 2004 Aug;35(8):1826-1830.
(117) Mok VC, Liu T, Lam WW, et al. Neuroimaging predictors of cognitive
impairment in confluent white matter lesion: volumetric analyses of 99 brain regions.
Dement Geriatr Cogn Disord 2008 Dec;25(1):67-73.
(118) Hund-Georgiadis M, Ballaschke O, Scheid R, Norris DG, von Cramon DY.
Characterization of cerebral microangiopathy using 3 Tesla MRI: correlation with
neurological impairment and vascular risk factors. J Magn Reson Imaging 2002
Jan;15(1):1-7.
(119) Hund-Georgiadis M, Norris DG, Guthke T, von Cramon DY. Characterization of
cerebral small vessel disease by proton spectroscopy and morphological magnetic
resonance. Cerebrovasc Dis 2001 Aug;12(2):82-90.
(120) Loeb C, Gandolfo C, Croce R, Conti M. Dementia associated with lacunar
infarction. Stroke 1992 Sep;23(9):1225-1229.
(121) Mok V, Chang C, Wong A, et al. Neuroimaging determinants of cognitive
performances in stroke associated with small vessel disease. J Neuroimaging 2005
Apr;15(2):129-137.
(122) Reed BR, Mungas DM, Kramer JH, et al. Profiles of neuropsychological
impairment in autopsy-defined Alzheimer's disease and cerebrovascular disease. Brain
2007 Mar;130(Pt 3):731-739.
(123) Gold G, Kovari E, Herrmann FR, et al. Cognitive consequences of thalamic, basal
ganglia, and deep white matter lacunes in brain aging and dementia. Stroke 2005
Jun;36(6):1184-1188.
(124) Tatemichi TK, Desmond DW, Prohovnik I, et al. Confusion and memory loss
from capsular genu infarction: a thalamocortical disconnection syndrome? Neurology
1992 Oct;42(10):1966-1979.
(125) de Oliveira Lanna ME, Madeira DM, Alves G, et al. Vascular dementia by
thalamic strategic infarct. Arq Neuropsiquiatr 2008 Jun;66(2B):412-414.
120
(126) Grau-Olivares M, Bartres-Faz D, Arboix A, et al. Mild cognitive impairment after
lacunar infarction: voxel-based morphometry and neuropsychological assessment.
Cerebrovasc Dis 2007 May;23(5-6):353-361.
(127) Wolf H, Jelic V, Gertz HJ, Nordberg A, Julin P, Wahlund LO. A critical
discussion of the role of neuroimaging in mild cognitive impairment. Acta Neurol
Scand Suppl 2003 Jul;179:52-76.
(128) O'Sullivan M, Summers PE, Jones DK, Jarosz JM, Williams SC, Markus HS.
Normal-appearing white matter in ischemic leukoaraiosis: a diffusion tensor MRI study.
Neurology 2001 Dec 26;57(12):2307-2310.
(129) Nitkunan A, Charlton RA, Barrick TR, McIntyre DJ, Howe FA, Markus HS.
Reduced N-acetylaspartate is consistent with axonal dysfunction in cerebral small vessel
disease. NMR Biomed 2009 Apr;22(3):285-291.
(130) van Zandvoort MJ, van der Grond J, Kappelle LJ, de Haan EH. Cognitive deficits
and changes in neurometabolites after a lacunar infarct. J Neurol 2005 Feb;252(2):183-
190.
(131) Yang DW, Kim BS, Park JK, Kim SY, Kim EN, Sohn HS. Analysis of cerebral
blood flow of subcortical vascular dementia with single photon emission computed
tomography: adaptation of statistical parametric mapping. J Neurol Sci 2002 Nov
15;203-204:199-205.
(132) Pineiro R, Pendlebury S, Johansen-Berg H, Matthews PM. Altered hemodynamic
responses in patients after subcortical stroke measured by functional MRI. Stroke 2002
Jan;33(1):103-109.
(133) 'Hashemi R', 'Bradley W', 'Lisanti C'. MRI: the basics. 2nd ed. Philadelphia:
Lippincott Williams & Wilkins; 2003.
(134) 'Westbrook C', 'Roth C', 'Talbot J'. MRI in practice. 3rd ed. Oxford: Blackwell
Publishing Ltd; 2005.
(135) Sands MJ, Levitin A. Basics of magnetic resonance imaging. Semin Vasc Surg
2004 Jun;17(2):66-82.
(136) Kates R, Atkinson D, Brant-Zawadzki M. Fluid-attenuated inversion recovery
(FLAIR): clinical prospectus of current and future applications. Top Magn Reson
Imaging 1996 Dec;8(6):389-396.
(137) Barkhof F, Scheltens P. Imaging of white matter lesions. Cerebrovasc Dis
2002;13 Suppl 2:21-30.
121
(138) Liu TT, Brown GG. Measurement of cerebral perfusion with arterial spin
labeling: Part 1. Methods. J Int Neuropsychol Soc 2007 May;13(3):517-525.
(139) Wintermark M, Sesay M, Barbier E, et al. Comparative overview of brain
perfusion imaging techniques. J Neuroradiol 2005 Dec;32(5):294-314.
(140) Brown GG, Clark C, Liu TT. Measurement of cerebral perfusion with arterial spin
labeling: Part 2. Applications. J Int Neuropsychol Soc 2007 May;13(3):526-538.
(141) Jezzard P, Buxton RB. The clinical potential of functional magnetic resonance
imaging. J Magn Reson Imaging 2006 Jun;23(6):787-793.
(142) Matthews PM, Honey GD, Bullmore ET. Applications of fMRI in translational
medicine and clinical practice. Nat Rev Neurosci 2006 Sep;7(9):732-744.
(143) Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion imaging. Magn Reson
Med 1992 Jan;23(1):37-45.
(144) Detre JA, Zhang W, Roberts DA, et al. Tissue specific perfusion imaging using
arterial spin labeling. NMR Biomed 1994 Mar;7(1-2):75-82.
(145) Roberts DA, Detre JA, Bolinger L, Insko EK, Leigh JS,Jr. Quantitative magnetic
resonance imaging of human brain perfusion at 1.5 T using steady-state inversion of
arterial water. Proc Natl Acad Sci U S A 1994 Jan 4;91(1):33-37.
(146) Silva AC, Zhang W, Williams DS, Koretsky AP. Multi-slice MRI of rat brain
perfusion during amphetamine stimulation using arterial spin labeling. Magn Reson
Med 1995 Feb;33(2):209-214.
(147) Zhang W, Silva AC, Williams DS, Koretsky AP. NMR measurement of perfusion
using arterial spin labeling without saturation of macromolecular spins. Magn Reson
Med 1995 Mar;33(3):370-376.
(148) Zhang W, Williams DS, Detre JA, Koretsky AP. Measurement of brain perfusion
by volume-localized NMR spectroscopy using inversion of arterial water spins:
accounting for transit time and cross-relaxation. Magn Reson Med 1992 Jun;25(2):362-
371.
(149) Williams DS, Detre JA, Leigh JS, Koretsky AP. Magnetic resonance imaging of
perfusion using spin inversion of arterial water. Proc Natl Acad Sci U S A 1992 Jan
1;89(1):212-216.
(150) Detre JA. Clinical applicability of functional MRI. J Magn Reson Imaging 2006
Jun;23(6):808-815.
122
(151) Campbell AM, Beaulieu C. Comparison of multislice and single-slice acquisitions
for pulsed arterial spin labeling measurements of cerebral perfusion. Magn Reson
Imaging 2006 Sep;24(7):869-876.
(152) Wolf RL, Detre JA. Clinical neuroimaging using arterial spin-labeled perfusion
magnetic resonance imaging. Neurotherapeutics 2007 Jul;4(3):346-359.
(153) Wolf RL, Alsop DC, McGarvey ML, Maldjian JA, Wang J, Detre JA.
Susceptibility contrast and arterial spin labeled perfusion MRI in cerebrovascular
disease. J Neuroimaging 2003 Jan;13(1):17-27.
(154) Ye FQ, Berman KF, Ellmore T, et al. H(2)(15)O PET validation of steady-state
arterial spin tagging cerebral blood flow measurements in humans. Magn Reson Med
2000 Sep;44(3):450-456.
(155) Feng CM, Narayana S, Lancaster JL, et al. CBF changes during brain activation:
fMRI vs. PET. Neuroimage 2004 May;22(1):443-446.
(156) Detre JA, Alsop DC, Vives LR, Maccotta L, Teener JW, Raps EC. Noninvasive
MRI evaluation of cerebral blood flow in cerebrovascular disease. Neurology 1998
Mar;50(3):633-641.
(157) Chalela JA, Alsop DC, Gonzalez-Atavales JB, Maldjian JA, Kasner SE, Detre JA.
Magnetic resonance perfusion imaging in acute ischemic stroke using continuous
arterial spin labeling. Stroke 2000 Mar;31(3):680-687.
(158) Chalela JA, Kasner SE, McGarvey M, Alsop DC, Detre JA. Continuous arterial
spin labeling perfusion magnetic resonance imaging findings in postpartum
vasculopathy. J Neuroimaging 2001 Oct;11(4):444-446.
(159) Buxton RB, Uludag K, Dubowitz DJ, Liu TT. Modeling the hemodynamic
response to brain activation. Neuroimage 2004 Sep;23 Suppl 1:S220-33.
(160) Bolan PJ, Yacoub E, Garwood M, Ugurbil K, Harel N. In vivo micro-MRI of
intracortical neurovasculature. Neuroimage 2006 Aug 1;32(1):62-69.
(161) Matthews PM, Jezzard P. Functional magnetic resonance imaging. J Neurol
Neurosurg Psychiatry 2004 Jan;75(1):6-12.
(162) Heeger DJ, Ress D. What does fMRI tell us about neuronal activity? Nat Rev
Neurosci 2002 Feb;3(2):142-151.
(163) Logothetis NK, Pauls J, Augath M, Trinath T, Oeltermann A. Neurophysiological
investigation of the basis of the fMRI signal. Nature 2001 Jul 12;412(6843):150-157.
123
(164) Chuang KH, van Gelderen P, Merkle H, et al. Mapping resting-state functional
connectivity using perfusion MRI. Neuroimage 2008 May 1;40(4):1595-1605.
(165) Brinker G, Bock C, Busch E, Krep H, Hossmann KA, Hoehn-Berlage M.
Simultaneous recording of evoked potentials and T2*-weighted MR images during
somatosensory stimulation of rat. Magn Reson Med 1999 Mar;41(3):469-473.
(166) Ogawa S, Lee TM, Stepnoski R, Chen W, Zhu XH, Ugurbil K. An approach to
probe some neural systems interaction by functional MRI at neural time scale down to
milliseconds. Proc Natl Acad Sci U S A 2000 Sep 26;97(20):11026-11031.
(167) Bangen KJ, Restom K, Liu TT, et al. Differential age effects on cerebral blood
flow and BOLD response to encoding: Associations with cognition and stroke risk.
Neurobiol Aging 2009 Aug;30(8):1276.
(168) Hoge RD, Atkinson J, Gill B, Crelier GR, Marrett S, Pike GB. Linear coupling
between cerebral blood flow and oxygen consumption in activated human cortex. Proc
Natl Acad Sci U S A 1999 Aug 3;96(16):9403-9408.
(169) Davis TL, Kwong KK, Weisskoff RM, Rosen BR. Calibrated functional MRI:
mapping the dynamics of oxidative metabolism. Proc Natl Acad Sci U S A 1998 Feb
17;95(4):1834-1839.
(170) Chiarelli PA, Bulte DP, Wise R, Gallichan D, Jezzard P. A calibration method for
quantitative BOLD fMRI based on hyperoxia. Neuroimage 2007 Sep 1;37(3):808-820.
(171) Kliefoth AB, Grubb RL,Jr, Raichle ME. Depression of cerebral oxygen utilization
by hypercapnia in the rhesus monkey. J Neurochem 1979 Feb;32(2):661-663.
(172) Rostrup E, Law I, Blinkenberg M, et al. Regional differences in the CBF and
BOLD responses to hypercapnia: a combined PET and fMRI study. Neuroimage 2000
Feb;11(2):87-97.
(173) Kim J, Whyte J, Wang J, Rao H, Tang KZ, Detre JA. Continuous ASL perfusion
fMRI investigation of higher cognition: quantification of tonic CBF changes during
sustained attention and working memory tasks. Neuroimage 2006 May 15;31(1):376-
385.
(174) Kim SG, Richter W, Ugurbil K. Limitations of temporal resolution in functional
MRI. Magn Reson Med 1997 Apr;37(4):631-636.
(175) McGonigle DJ, Howseman AM, Athwal BS, Friston KJ, Frackowiak RS, Holmes
AP. Variability in fMRI: an examination of intersession differences. Neuroimage 2000
Jun;11(6 Pt 1):708-734.
124
(176) Aguirre GK, Zarahn E, D'esposito M. The variability of human, BOLD
hemodynamic responses. Neuroimage 1998 Nov;8(4):360-369.
(177) Aguirre GK, Detre JA, Zarahn E, Alsop DC. Experimental design and the relative
sensitivity of BOLD and perfusion fMRI. Neuroimage 2002 Mar;15(3):488-500.
(178) Kemeny S, Ye FQ, Birn R, Braun AR. Comparison of continuous overt speech
fMRI using BOLD and arterial spin labeling. Hum Brain Mapp 2005 Mar;24(3):173-
183.
(179) Tjandra T, Brooks JC, Figueiredo P, Wise R, Matthews PM, Tracey I.
Quantitative assessment of the reproducibility of functional activation measured with
BOLD and MR perfusion imaging: implications for clinical trial design. Neuroimage
2005 Aug 15;27(2):393-401.
(180) Wang J, Li L, Roc AC, et al. Reduced susceptibility effects in perfusion fMRI
with single-shot spin-echo EPI acquisitions at 1.5 Tesla. Magn Reson Imaging 2004
Jan;22(1):1-7.
(181) Duong TQ, Kim DS, Ugurbil K, Kim SG. Localized cerebral blood flow response
at submillimeter columnar resolution. Proc Natl Acad Sci U S A 2001 Sep
11;98(19):10904-10909.
(182) Luh WM, Wong EC, Bandettini PA, Ward BD, Hyde JS. Comparison of
simultaneously measured perfusion and BOLD signal increases during brain activation
with T(1)-based tissue identification. Magn Reson Med 2000 Jul;44(1):137-143.
(183) Davis SM, Donnan GA. Why lacunar syndromes are different and important.
Stroke 2004 Jul;35(7):1780-1781.
(184) Gerraty RP, Parsons MW, Barber PA, et al. Examining the lacunar hypothesis
with diffusion and perfusion magnetic resonance imaging. Stroke 2002 Aug;33(8):2019-
2024.
(185) Markus H. A DNA RESOURCE FOR LACUNAR (SMALL VESSEL DISEASE)
STROKE study protocol. 2004; Available at:
http://www.sgul.ac.uk/research/divisions/cardiovascular-sciences/clinical-
neuroscience/stroke/dna-resource/Study%20Protocol.pdf. Accessed November, 2008.
(186) Goodwin JA, Vidyasagar R, Balanos GM, Bulte D, Parkes LM. Quantitative
fMRI using hyperoxia calibration: Reproducibility during a cognitive Stroop task.
Neuroimage 2009 May 3;47(2):573.
(187) Luh WM, Wong EC, Bandettini PA, Hyde JS. QUIPSS II with thin-slice TI1
periodic saturation: a method for improving accuracy of quantitative perfusion imaging
using pulsed arterial spin labeling. Magn Reson Med 1999 Jun;41(6):1246-1254.
125
(188) Grubb RL,Jr, Raichle ME, Eichling JO, Ter-Pogossian MM. The effects of
changes in PaCO2 on cerebral blood volume, blood flow, and vascular mean transit
time. Stroke 1974 Sep-Oct;5(5):630-639.
(189) Wahlund LO, Barkhof F, Fazekas F, et al. A new rating scale for age-related
white matter changes applicable to MRI and CT. Stroke 2001 Jun;32(6):1318-1322.
(190) Gregoire SM, Chaudhary UJ, Brown MM, et al. The Microbleed Anatomical
Rating Scale (MARS): reliability of a tool to map brain microbleeds. Neurology 2009
Nov 24;73(21):1759-1766.
(191) Gao T, Wang Y, Zhang Z. Silent cerebral microbleeds on susceptibility-weighted
imaging of patients with ischemic stroke and leukoaraiosis. Neurol Res 2008
Apr;30(3):272-276.
(192) Mori N, Miki Y, Kikuta K, et al. Microbleeds in moyamoya disease:
susceptibility-weighted imaging versus T2*-weighted imaging at 3 Tesla. Invest Radiol
2008 Aug;43(8):574-579.
(193) Genovese CR, Lazar NA, Nichols T. Thresholding of statistical maps in
functional neuroimaging using the false discovery rate. Neuroimage 2002
Apr;15(4):870-878.
(194) Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and
Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society Series
B (Methodological) 1995;57(1):289.
(195) Lancaster JL, Woldorff MG, Parsons LM, et al. Automated Talairach atlas labels
for functional brain mapping. Hum Brain Mapp 2000 Jul;10(3):120-131.
(196) De Reuck J, Decoo D, Hasenbroekx MC, et al. Acetazolamide vasoreactivity in
vascular dementia: a positron emission tomographic study. Eur Neurol 1999
Jan;41(1):31-36.
(197) Schroeter ML, Bucheler MM, Preul C, et al. Spontaneous slow hemodynamic
oscillations are impaired in cerebral microangiopathy. J Cereb Blood Flow Metab 2005
Dec;25(12):1675-1684.
(198) Hund-Georgiadis M, Zysset S, Naganawa S, Norris DG, Von Cramon DY.
Determination of cerebrovascular reactivity by means of FMRI signal changes in
cerebral microangiopathy: a correlation with morphological abnormalities. Cerebrovasc
Dis 2003 Jun;16(2):158-165.
(199) Krainik A, Hund-Georgiadis M, Zysset S, von Cramon DY. Regional impairment
of cerebrovascular reactivity and BOLD signal in adults after stroke. Stroke 2005
Jun;36(6):1146-1152.
126
(200) Ross MH, Yurgelun-Todd DA, Renshaw PF, et al. Age-related reduction in
functional MRI response to photic stimulation. Neurology 1997 Jan;48(1):173-176.
(201) De Reuck J, Decoo D, Marchau M, Santens P, Lemahieu I, Strijckmans K.
Positron emission tomography in vascular dementia. J Neurol Sci 1998 Jan
21;154(1):55-61.
(202) O'Sullivan M, Lythgoe DJ, Pereira AC, et al. Patterns of cerebral blood flow
reduction in patients with ischemic leukoaraiosis. Neurology 2002 Aug 13;59(3):321-
326.
(203) Yao H, Sadoshima S, Kuwabara Y, Ichiya Y, Fujishima M. Cerebral blood flow
and oxygen metabolism in patients with vascular dementia of the Binswanger type.
Stroke 1990 Dec;21(12):1694-1699.
(204) Bastos-Leite AJ, Kuijer JP, Rombouts SA, et al. Cerebral blood flow by using
pulsed arterial spin-labeling in elderly subjects with white matter hyperintensities.
AJNR Am J Neuroradiol 2008 Aug;29(7):1296-1301.
(205) Kitagawa K, Oku N, Kimura Y, et al. Relationship between cerebral blood flow
and later cognitive decline in hypertensive patients with cerebral small vessel disease.
Hypertens Res 2009;32(9):816.
(206) Yoshikawa T, Murase K, Oku N, et al. Statistical image analysis of cerebral blood
flow in vascular dementia with small-vessel disease. J Nucl Med 2003 Apr;44(4):505-
511.
(207) Zaharchuk G, Martin AJ, Dillon WP. Noninvasive imaging of quantitative
cerebral blood flow changes during 100% oxygen inhalation using arterial spin-labeling
MR imaging. AJNR Am J Neuroradiol 2008 Apr;29(4):663-667.
(208) Bulte DP, Chiarelli PA, Wise RG, Jezzard P. Cerebral perfusion response to
hyperoxia. J Cereb Blood Flow Metab 2007 Jan;27(1):69-75.
(209) Watson NA, Beards SC, Altaf N, Kassner A, Jackson A. The effect of hyperoxia
on cerebral blood flow: a study in healthy volunteers using magnetic resonance phase-
contrast angiography. Eur J Anaesthesiol 2000 Mar;17(3):152-159.
(210) Mark CI, Fisher JA, Pike GB. Improved fMRI calibration: Precisely controlled
hyperoxic versus hypercapnic stimuli. Neuroimage 2010 Sep 7;54(2):1102.
(211) Tak S, Yoon SJ, Jang J, Yoo K, Jeong Y, Ye JC. Quantitative analysis of
hemodynamic and metabolic changes in subcortical vascular dementia using
simultaneous near-infrared spectroscopy and fMRI measurements. Neuroimage 2011
Mar;55(1):176-184.
127
(212) Schmidt R, Ropele S, Enzinger C, et al. White matter lesion progression, brain
atrophy, and cognitive decline: the Austrian stroke prevention study. Ann Neurol 2005
Oct;58(4):610-616.
(213) Kramer JH, Mungas D, Reed BR, et al. Longitudinal MRI and cognitive change in
healthy elderly. Neuropsychology 2007 Jul;21(4):412-418.
(214) Corbett A, Bennett H, Kos S. Cognitive dysfunction following subcortical
infarction. Arch Neurol 1994 Oct;51(10):999-1007.
(215) Prins ND, van Dijk EJ, den Heijer T, et al. Cerebral small-vessel disease and
decline in information processing speed, executive function and memory. Brain 2005
Sep;128(Pt 9):2034-2041.
(216) Mungas D, Reed BR, Jagust WJ, et al. Volumetric MRI predicts rate of cognitive
decline related to AD and cerebrovascular disease. Neurology 2002 Sep 24;59(6):867-
873.
(217) Jagust WJ, Zheng L, Harvey DJ, et al. Neuropathological basis of magnetic
resonance images in aging and dementia. Ann Neurol 2008 Jan;63(1):72-80.
(218) Cohen RA, Paul RH, Ott BR, et al. The relationship of subcortical MRI
hyperintensities and brain volume to cognitive function in vascular dementia. J Int
Neuropsychol Soc 2002 Sep;8(6):743-752.
(219) Jackson CA, Hutchison A, Dennis MS, et al. Differing risk factor profiles of
ischemic stroke subtypes: evidence for a distinct lacunar arteriopathy? Stroke 2010
Apr;41(4):624-629.
(220) Potter G, Doubal F, Jackson C, Sudlow C, Dennis M, Wardlaw J. Associations of
clinical stroke misclassification ('clinical-imaging dissociation') in acute ischemic
stroke. Cerebrovasc Dis 2010 Feb;29(4):395-402.
(221) Gorbet DJ, Sergio LE. Preliminary sex differences in human cortical BOLD fMRI
activity during the preparation of increasingly complex visually guided movements. Eur
J Neurosci 2007 Feb;25(4):1228-1239.
(222) Kastrup A, Li TQ, Glover GH, Kruger G, Moseley ME. Gender differences in
cerebral blood flow and oxygenation response during focal physiologic neural activity. J
Cereb Blood Flow Metab 1999 Oct;19(10):1066-1071.
(223) Kaufmann C, Elbel GK, Gossl C, Putz B, Auer DP. Frequency dependence and
gender effects in visual cortical regions involved in temporal frequency dependent
pattern processing. Hum Brain Mapp 2001 Sep;14(1):28-38.
128
(224) Marcar VL, Loenneker T, Straessle A, Girard F, Martin E. What the little
differences between men and women tells us about the BOLD response. Magn Reson
Imaging 2004 Sep;22(7):913-919.
(225) Moulton EA, Keaser ML, Gullapalli RP, Maitra R, Greenspan JD. Sex differences
in the cerebral BOLD signal response to painful heat stimuli. Am J Physiol Regul Integr
Comp Physiol 2006 Aug;291(2):R257-67.
(226) O'Carroll RE, Baikie EM, Whittick JE. Does the National Adult Reading Test
hold in dementia? Br J Clin Psychol 1987 Nov;26(Pt 4):315-316.
(227) Nakane H, Ibayashi S, Fujii K, et al. Cerebral blood flow and metabolism in
patients with silent brain infarction: occult misery perfusion in the cerebral cortex. J
Neurol Neurosurg Psychiatry 1998 Sep;65(3):317-321.
