.2 I z I
.13.. all

 

J
fail...

.lvcbn.

. wfifim 7.
Ha. .. . . in...
$5.... :Mr gum

awn:

é“

1:»Mrh3034 .
. .u .322...
. .wu Gr.

‘
affl um N
. “nxswmfi www.11-
mm. why. ...
:U‘I‘l
* s

.1 Lara

 

 

 

 

 

 

This is to certify that the
dissertation entitled

EVIDENCE FOR HETEROGENEITY IN MILD COGNITIVE
IMPAIRMENT (MCI): DIFFERENCES IN
NEUROPSYCHOLOGICAL PROFILE AND ASSOCIATED
WHITE MATTER LESION PATHOLOGY

presented by
LISA M. DELANO-WOOD

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Clinical Psychol_ogL

 

(M lflU/nv Mtg C)

Major Professor’s Signature
/
5/16 03
I I

Date

MSU is an Affinnative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:/ClRC/DateDue.lndd-p.15

EVIDENCE FOR HETEROGENEITY 1N MILD COGNITIVE IMPAIRMENT (MCI):
DIFFERENCES IN NEUROPSYCHOLOGICAL PROFILE AND ASSOCIATED
WHITE MATTER LESION PATHOLOGY
By

Lisa M. Delano-Wood

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

2005

EVIDENCE FOR I-IETEROGENEITY 1N MILD COGNITIVE IMPAIRMENT (MCI):
DWCES IN NEUROPSYCHOLOGICAL PROFILE AND ASSOCIATED
WHITE MATTER LESION PATHOLOGY
By
Lisa M. Delano-Wood

Since Alzheimer’s disease (AD) and vascular dementia (V aD) are associated with
considerable morbidity and mortality and together comprise roughly 80% of all dementia
cases worldwide (Morris, 1994; Small. Rabins, Barry, & Buckholtz, 1997), delineating
and elucidating early, pro-clinical manifestations of these dementia syndromes is
essential. Although no reliable treatments currently exist for dementia, it is important to
identifyand define early stagwofdiedemaitingpmcwswiththe goal ofmore clearly
understanding how and why dementia progresses; in addition, identifying early
manifestations of dementia is important in terms of possibly slaving 03‘ future cognitive
decline with existing pharmacological treatment. Mild Cognitive Impairment (MCI), a
reliable clinical entity that has received considerable attention since its inception in 1997,
is thought to represent a vulnerable population lying between normal aging and dementia
(Morris et al., 2001; Petersen et al., 2001 ). Indeed, \mderstanding what constitutes MCI
is impaative as several studies have indicated that individuals in this at-n'sk group are at
dramatically increased risk for dementia and mortality (Palmer et al., 2003; Peta'sen et
al., 1999).

This study examined whether distinct nemopsychological profiles could be
delineated within an MCI population as well as to investigate the contribution of white

matter lesions (WML) to associated cognitive impairment. A clinical sample of 70 olda

adults diagnosed with MCI (age range 55-88) was assessed using neurOpsychological test

scores (CERAD battery, Trails A and B, and Stroop Color Word Test). Additionally,
WML found on structural MRI was measured using a semi-automated volumetric
approach (Pixel Thresholding) with T2-weighted FLAIR (fluid-attenuated inversion
recovery) images. Using cluster and discriminant analyses, three distinct groups
(Cortical, Subcortical, and Amnestic) were formed based on neuropsychological scores.
Results showed that each group differed on white matter lesion load, with the Subcortical
group demonstrating the highest level of WML pathology. Finally, using regression
analyses, the effect of lesion type (deep white matter (DWML) versus periventricular
(PVL)) on neuropsychological performance was investigated. Only DWML was
associated with greater cognitive impairment, likely due to frontal-subcortical circuitry
disruption. Taken together, findings suggest that distinct neuropsychological profiles
exist within MCI and that these profiles differ according to levels of WML. Clinical,

theoretical, and methodological implications of these results are discussed.

Thisdissertationisdedicatedtomyfatherwhoisandalwayshasbeen
the strongest influence in my life.

ACKNOWLEDGMENTS

I would like to thank my family and fiiends for their support and encomagement
throughout graduate school and the writing of this dissertation. Most especially, I would
like to thank my husband, Bob Wood, for his support throughout the past decade. This
dissertation would not have been possible forme to embark upon and complete
successfully without his patience and kindness. I would also like to thank Dr. Andrea
Bozoki whomadethisstudyposm’ble. Hainsightsandexpertisein the
fields of neurology and aging have been invaluable. Finally, I would like to thank the
chair of my committee—Dr. Norman Abeles—who, with all of his guidance, knowledge,

and support has been one of the most wonderful professors from which to learn.

 

 

 

 

 

 

 

A.‘

M

All:

CH
ME;

 

TABLE OF CONTENTS
LISTOFFIGURES ............................................................................... ix

CHAPTER 1

INTRODUCTION"...................................... .............................................................. 1

Mild Cognitive Impairment (MCI) .................................................... 2
Clinical and Diagnostic Definition... ........................................................ 2
Possible Subtypes... 3
Clinical Outcome" 4

Distinguishing Alzheimer’s Disease (AD) and Vascular Dementia (V aD) .............. 6
Neuroanatomical/Neuropsychological Differences m AD and VaD... ...8
Problems with the DSM- IV VaD Diagnosis... 10

A New Preposal: Vascular Cognitive Impairment (V CI) .. . . . 13
Continuum of Impairment m Vascular Cognitive Impairment (V CI) ........... l4
Neuropsychological Impairment m VCI.. l6
Outcome m VCI.. ..................................................................... 17

White Matter Lesion (WML) Pathology... 18
Potential Causes of WML ........................................................................ 18
Major Subtypes of WML ......................................................................... 19
Relationship between WML and Vascular Risk Factors........................ 20
VCI and White Matter Lesions ...................................................... 23
Neuropsychological Deficits Associated with WML .................................... 24
WML and Depression 27

 

CHAPTER 2
Hypotheses ............................................................................ 31

CHAPTER 3
METHOD ......................................................................................... 34
Sample ................................................................................. 34
Procedures ............................................................................. 35
Diagnosis ofMild Cognitive Impairment 35
Neuropsychological Test Battery 36
Quantification of White Matter Lesions ........................................... 38
Statistical Analysis. ..................................................................... 40

CHAPTER 4

RESULTS ...................................................................................... 44
Hypothesis 1 ............................................................................. 44
Hypothesis 2 ............................................................................. 44

Hypothesis 3 ........................................................................... 49

 

Hypothesis .................................................................................................... 50

CHAPTER 5

DISCUSSION ........................................................................................................... 52
Relationship of WMLVolume to Overall Neuropsychological F Imctioning.52
Neuropsychological Performance and MCI Group Membership ................. 53
Total WML Volume and MCI Group Membership ...................................... 57
Periventricular and Deep White Matter Volumes and MCI Group
Membership ...................................................... _ - -- 6O
LimitationsandFutureDirections........... .................................................. 63

Conclusion 64

REFERENCES ................................................................................... 84

vii

 

 

 

 

 

Tat

Tat

Tab

Tab

Tab]

LIST OF TABLES

Table 1: Descriptive statistics and intercorrelations among demographic,
neuropsychological, and WML variables... ....................................................... 66

Table 2: Group means and standard deviations on neuropsychological variables by group
membership ....................................................................................................... 67

Table 3: Demographic data (means and standard deviations) by cluster
Groupings .................................................................................. 68

Table 4: Structure ofthe discriminant functions . ....69
Table 5: Functions at group centroids (means) ................................................................ 70
Table 6: Means and standard deviations on WML total volume across clusters ............. 71

Table 7: Hierarchical regression for PVL as a predictor of overall neuropsychological
fimctioning (NP) ................................................................................................ 72

Table 8: Hierarchical regression for DWML as a predictor of overall neuropsychological
functioning(NP)... .73

Table 9: Hierarchical regression for DWML as a predictor of executive functioning

Table 10: Hierarchical regression for DWML as a predictor of processing speed

Table 11: Hierarchical regression for DWML as a predictor of visuospatial/construction

Table 12: Hierarchical regression for DWML as a predictor of confi'ontation naming
(naming)... . 77

Table 13: Hierarchical regression for DWML as a predictor of verbal fluency
(fluency)... 78

Table 14: Hierarchical regression for DWML as a predictor of memory (mem) .......... 79

viii

LIST OF FIGURES
Figure 1: Nemopsychological profile by cluster group ................................................... 80
Figure 2: Depression levels as a function of group81
Figure 3: Biterritorial map of discriminant functions. .................................................. 82

Figure 4: WML as a function of group membershrp83

ix

CHAPTER 1
INTRODUCTION

Recent population statistics have demonstrated that the “graying of America” is
imminent. Indeed, current population estimates have revealed that over 51 million people
in the United States (21% of the total population) are at least 55 years of age (U .8.
Bureau of Census, 2000), a 10-fold increase over estimates in 1900. Furthermore, the
“oldest old,” or the 85 and older demographic, is expected to more than double from 4
million today to 8.5 million by 2030, making this age group the fastest growmg segment
of the population (J este, 1997).

It is striking that approximately 50% of those aged 80 and older demonstrate
some kind of cognitive impairment or dementia (Hachinski, 1992). Indeed, older adults
are at increasing risk for developing Alzheimer’s disease (AD) or vascular dementia
(V aD), two highly prevalent diseases that together comprise roughly 80% of all cases
worldwide (Morris, 1994; Small, Rabins, Barry, & Buckholtz,, 1997). It has become
glaringly apparent that increased research and resources need to be devoted to our older
adult population if morbidity and mortality and the expense and suffering inherent to
these dementias are to be appropriately controlled and attenuated.

Clearly, understanding and delineating the types and causes of dementia remain a
significant priority for the medical, psychological, and neuropsychological fields. Before
discussing white matter lesions and their potential role in neuropsychological
impairment, a brief synopsis of Mild Cognitive Impairment (MCI)--a milder form of
cognitive impairment which is thought to represent a pre-dementia state-will be
provided. In addition, diagnostic criteria for AD and VaD will be discussed with the

central aim of highlighting similarities and differences between the two dementia types;

 

 

Ill

Ct

pf
ml

Tc

 

1h:

C0

this will provide the structure for the conceptual background of the present research,
which aims to further elucidate the nature of white matter lesions and their significance
with respect to neuropsychological functioning in MCI.

Mild Cognitive Impairment

Clinical and Diaggostic Definition

Mild Cognitive Impairment (MCI) is a condition representing an area between
normal aging and dementia in which there are subtle, clinically manifest signs of
cognitive impairment in the absence of overt deficits meeting criteria for dementia
(Petersen, 1999, 2001). Research has shown that MCI represents an underlying disease
process which can be differentiated from normal aging-related physiological changes and
thus represents a transitional area between normal aging and dementia (Ritchie &

T ouchon, 2000; Almkvist et al, 1998; Jacobs et al., 1995j; Morris et al., 2001; Petersen et

al., 2001; Petersen et al., 1999).

Mild Cognitive Impairment is a medical diagnosis currently in its infancy, and
thus diagnostic criteria are somewhat inconsistent and there is no current agreed upon
consensus in the field on a Single set of criteria (Monis et al., 2001; Petersen et al., 2001;
2003). Current criteria for MCI include the following: 1) memory complaints,
preferably corroborated by an informant, 2) objective memory impairment for age, 3)
largely preserved general cognition, 4) essentially normal activities of daily living, and 5)
no dementia (Petersen, 1997; 2001; 2003). However, as will be discussed in further
detail below, Petersen has recently expanded to include deficits in at least two
neuropsychological domains, after controlling for age and educational achievement

(Petersen et al., 2001, 2003). Despite the variability in criteria, recent studies have

 

m
((3.

sub

COS

002

)1:

CK

validated MCI as a useful, separate entity that is qualitatively different from both normal
aging and dementia (Petersen et al., 1999; 2004; Bennett, Wilson, & Schneider et al.,
2002; Larrieu, Letenneur, & Orgogozo et al., 2002; Smith & Ivnik, 2003). In the absence
of a reliable biological marker, individual performance on neuropsychological tests is
currently the most reliable way of detecting early phases of dementia.
Possible Sungpes of MCI

AS the concept of an intermediate state of cognitive impairment between normal
aging and dementia has advanced, it has recently been suggested that several clinical
subtypes of MCI may exist (Petersen, 2003). Since some patients present with mild
cognitive impairment in a cognitive domain other than memory or within multiple
domains, the idea that MCI may represent other etiologies aside from Alzheimer’s
disease has begun to receive considerable attention (Petersen 2001, Petersen 2003;
Rockwood et al., 2003). To this end, the concept of MCI has recently been expanded to
include at least three subtypes: 1) Amnestic MCI in which a person has impairment only
in memory; 2) Multiple-Domain MCI in which a person may have mild impairments in
several cognitive domains with or without a memory impairment; and 3) Single Non-
Memory Domain MCI in which a person is impaired in a non-memory area such as
executive function or language.

In addition to the clinical subtypes, Petersen et al. (2003) posits that there may
be multiple etiologies or causes for each subtype. For example, he has argued that the
Multiple-Domain type of MCI may either progress to AD or, alternatively, it may be a
phenotype of early vascular dementia. While research investigating the Non-Amnestic

types of MCI is virtually absent from the field at this time, most research has focused

almost exclusively on the Amnestic type which has been shown to overwhelmingly
progress to AD (Bennett et al., 2002; Lanieu et al., 2002; Petersen et al., 1999).

Clinical Outcome of MCI

Unlike previous concepts, MCI is based on a pathological model of cognitive

change (Ritchie & Touchon, 2000). Indeed, reviews of several studies have indicated
that individuals with MCI are at risk of higher rates of death (Bennett et al., 2002; Palmer
et al., 2003) and it has been shown that up to 25% of those with MCI develop dementia
annually (Petersen et al., 1999; Palmer et al., 2003). The most common converting type
of dementia appears to be Alzheimer’s disease with an annual converting rate that is ten

times higher than in normal aging (Petersen et a1, 2001).

Several studies have been conducted in recent years in a variety of research
settings to measure outcome in subjects with MCI (Bozoki, Giordani, & Heidebrink et
al., 2001; Rasquin, Lodder and Visser et al., 2005; Fiducia & DeFilippis, 2004). For
example, a group of 220 older adults with a mean age of 79 were followed at the Mayo
Alzheimer’s Disease Research Center/Alzheimer’s Disease Patient Registry for 3-6 years
post diagnosis of MCI (Petersen et al., 1999). It has been shown that subjects in this
study have progressed to dementia at a rate of approximately 12% per year (Petersen et
al., 2001), indicating that roughly 80% of the individuals in this study will have
converted to dementia within 6 years. Moreover, in the first population-based study of
Amnestic and Multiple-Domain MCI (Cardiovascular Health Study), results indicated
that the overall prevalence for the Amnestic subtype of MCI was 6% and Multiple-
Domain MCI represented 16% (Lopez, Jagust, & Dekosky, 2003). These statistics

highlight a greater need to better understand and characterize Multiple-Domain MCI

 

(It

«it.

given its high prevalence; additionally, they underscore the high risk nature of this group
as a whole.

In general, individuals with MCI function well in their activities of daily living
and differentiation of any mild functional difficulty is often challenging. Thus, early
symptoms of dementia are commonly overlooked because they are relatively mild, do not
call for immediate medical attention, and are often dismissed as signs of old age
(Hachinski, 1994; Bowler & Hachinski, 2003). Quite frequently, recognition of dementia
comes too late when too much cerebral damage has occurred. At this point, cognitive
decline progresses at an extremely rapid pace, the individual moves into severe stages of
impairment, and any available treatment cannot be effective. Thus, early detection is
critical for individuals afflicted with early dementia so that intervention can be beneficial.
To this end, MCI research is currently focusing on recognition of the risk factors of
disease progression for the purpose of identifying therapeutic interventions that may
delay or possibly prevent the onset of dementia. One major goal in this field is to more
clearly define MCI and to decipher whether the diagnosis primarily represents a risk
factor for Alzheimer’s disease Specifically, or whether MCI is more accurately described
as comprised of a heterogeneous group of individuals with different pathologies,
neurological markers, and/or neuropsychological profiles (Looi & Sachdev, 2000a;
Bowler, 2002; Rockwood, K., Davis, H., & MacKnight, C., et al., 2003). Before
discussing research relevant to this question, a brief description of the two most prevalent

types of dementia—Alzheimer’s disease and vascular dementia—follows.

 

Distinguishing Alzheimer’s Disease and Vascular Dementia

There is an increasing emphasis on the early diagnosis of dementia and, as
preventive and early intervention strategies become better established, clinical
differentiation between the two most common dementia types (Alzheimer’s disease (AD)
and vascular dementia (V aD)) will be essential. Indeed, AD and VaD differ in many
respects and precision in diagnosis is thus very important (Chui & Gonthier, 1999;
Rockwood et al., 2003). Specifically, research has demonstrated that the rate of
progression and prognosis differ for the two disorders (Chui & Gonthier, 1999) and
differential diagnosis has implications for effective long-tenn management and resources.
While the etiology and risk factors are not completely understood in AD, risk factors for
VaD such as hypertension and stroke are better established and are more likely to be
modifiable. Indeed, it has been speculated by many that VaD may be more amenable
than other dementia types such as AD to early intervention (Loci & Sachdev, 2000a;

Bowler, 2002; Rockwood et al., 2003).

All dementias are characterized by a progression of cognitive and functional
capacity. Conventionally, AD and VaD can be distinguished by separate risk factors,
clinical course, neuroirnaging characteristics, and neurological markers (Roman,
Tatemichi, Erkinjuntti, Cummings, et a1, 2001; Roman, 2001). For example, AD has
been described as having an insidious onset that is followed by slow, progressive
cognitive deterioration. It tends to have no or very little obvious focal neurological
symptoms and signs and its neuropathology is well characterized as primarily

degenerative (Braak and Braak, 1991).

Conversely, VaD has been traditionally thought of as having an abrupt onset of
dementia symptoms, followed by stepwise deterioration of cognitive performance
reflecting neurological signs and symptoms consistent with focal brain lesions (Roman et
al., 2001). Current DSM-IV criteria indicate that cognitive loss in combination with
evidence of stroke warrants diagnosis of VaD. However, recent research has
demonstrated that VaD appears to be more heterogeneous in origin, pathogenesis, and
clinical course compared with AD (Rockwood etal., 1999; Loeb & Meyer, 1996;
Erkinjuntti, 2000; and Hachinski, 1994). In fact, Meyer et al. (2002) argue that VaD can
be separated into two groups, with one group having abrupt onset caused predominantly
by multi-infarct, strategic-infarct, or intracranial hemorrhage (in line with the traditional
view of VaD); the other form of VaD appears to have an insidious onset, caused
predominantly by subcortical small—vessel disease. Indeed, recent research has
demonstrated that subcortical small-vessel disease appears to be an important etiology for
VaD (Erkinjuntti, Inzitari, & Pantoni et al., 2000; Chui, 2001), with prevalence rates
ranging from 36% to 67% (Erkinjuntti, 1987; Cummings, 1994; Brun, 1994; Ross,
Petrovitch, & White, et al., 1999).

Unlike typical multi-infarct dementia, particularly that caused by large cortical
strokes, VaD caused by subcortical small-vessel disease has a relatively insidious onset
which can mimic the clinical course of AD (Chui, 2001, Erkinjuntti et al., 2000; Roman,
Tatemichi, & Erkinjuntti, et al., 1993). Slow progression of neuropsychological deficits
is thought to arise due to gradually progressive microvascular changes in the brain.
Specifically, it has been posited that hypertensive arteriolar lipohyalinosis involving

small penetrating vessels (cerebral microangiopathy) is responsible for underlying

 

 

cl

 

 

in ;
pref
6X6

C0n'

Info

 

neuropathological changes that promote VaD (Erkinjuntti & Hachinski, 1993; Filey,
Franklin, & Heaton, et al., 1989; Chui, 2001). It is thought that these neuropathological
changes associated with vascular risk factors (i.e., hypertension, hyperlipidemia, and
diabetes) are present a relatively long time before VaD becomes diagnosable and may
cause cognitive impairment over time (Meyer et al., 2002; Bowler & Hachinski, 2003).
The relationship between vascular risk factors and subsequent cognitive decline will be
discussed in detail in the sections that follow.

Given that there are different subgroupings of VaD, it has been somewhat
challenging to identify homogeneous populations to study in terms of neuropsychological
performance. Most research to date has focused on multi-infarct dementia and, although
there is some overlap between the major subgroups, efforts to distinguish them according
to dementia onset have clinical merit. Early identification of insidious-onset VaD
(subcortical VaD) introduces an opportunity for clinical interventions such as control of
vascular risk factors that may minimize, arrest, or even reverse cognitive deterioration
(Skoog, 1998; Bowler & Hachinski, 2003; & Li et al., 2002).

Neuroanatomical Neuropsychological Differences in AD and VaD

The neuropathological basis and pattern of cognitive deficits differs considerably
in Alzheimer’s disease and vascular dementia. Specifically, early VaD appears to
preferentially affect subcortical and fiontal lobe functions (i.e. Speed of processing and
executive fimctioning), while early AD primarily involves the medial temporal lobes and,
consequently, memory firnctions (Bowler, 2002). Indeed, memory disturbance is an early
feature in AD, with problems in new learning (Lafosse et al., 1997), a faster rate of

information decay (Carlesirno et al., 1995), and a reduced ability to benefit from cues to

facilitate retrieval (Del Re et al., 1994). However, there is greater heterogeneity in the
various neocortical regions affected and, consequently, neuropsychological profiles may
vary somewhat, particularly later in the disease process (Fisher, Rourke, & Bieliauskas, et
al., 1997).

Memory is supported by multiple neural systems, with particular involvement of
the medial temporal and diencephalic structures (U ngerleider, 1995; Curran and Schacter,
1996; Haxby, 1996). In AD, the neuropathological lesions impact directly and early on
structures closely associated with memory such as the hippocampus (Squire and
Shirnarnura, 1996). During progression of the dementia process, widespread pathology
spreads across the medial temporal lobe (Garrard et al., 1997) which, in addition to
memory, is largely responsible for language and semantic knowledge. In contrast,
memory and language abilities may be relatively spared in subcortical VaD due to greater
heterogeneity of neuropathology and the frequent sparing of medial temporal lobe
structures.

VaD is thought to primarily affect subcortical and fi'ontal lobe functions (Bowler,
2002; Rockwood et al., 2003) and patients with this dementia type have been described in
the literature to exhibit early executive impairment, slowed psychomotor functioning, and
fiequent depressive features (Bartres-Faz, Clemente, & Junque, 2001; Kramer-Ginsberg
et al., 1999; Libon et al., 2001; Naarding et al., 2003; & O’Sullivan et al., 2001). Frontal-
subcortical circuits have been an area of recent research interest (Mega and Cummings,
1994; Bartrez-Faz et al., 2001; Libon et al., 2001; Bigler et al., 2003) and the circuits
disrupted in VaD appear to include: the dorsolateral prefrontal neuronal circuit mediating

executive functioning, orbit-fi'ontal circuit mediating emotional lability, and the anterior

cingulate circuit responsible for motivation and initiation (Almkvist, 1994; Cummings,
1994). Moreover, correlated imaging studies lend support to the conceptualization of
VaD as being characterized by a greater degree of frontal-subcortical dysfunction than in
AD of comparable severity (Aharon-Peretz et al., 1988; Starkstein et al., 1996; Lafosse et
al., 1997; Libon et al., 1997; Villardita, 1997).
Problems with the DSM-IV VaD Diagnosis

More recently, the traditional DSM-IV diagnosis of vascular dementia (VaD) has
been challenged as problematic and the foundations for the diagnosis are under scrutiny
(Hachinski, 1994; Rockwood, 1997; Rockwood, Parhad, & Hachinski et al., 1994;
Bowler, Eliasziw, & Steenhuis et al., 1997; Verhey, Lodder, & Rozendaal et al., 1996;
Nolan, Lino, Seligmann, & Blass, 1998; Rockwood, Wentzel, & Hachinski et al., 2000).
Indeed, it has been noted that the diagnosis of VaD is clouded by many false perceptions;
specifically, these assumptions are 1) that the course of VaD is chronic and progressive,
2) that a large stroke or many strokes must precede VaD; and 3) that vascular dementias
are quite rare (Brust, 1988). Contrary to present day criteria, focal neurological signs,
sudden onset, step-wise progression, and relationship to known stroke(s) are argued to be
unnecessary (Erkinjuntti & Rockwood, 2001; Bowler & Hachinski, 2003; Rockwood et
al., 2000; Sachdev & Looi, 2003).

The DSM-IV criteria for VaD have been criticized as “Alzheimerized” (Sachdev
& Looi, 2003) and thus inappropriate since the criteria were based primarily on AD
(McKhann et al., 1984; Hachinksi, 1990, 1994, 1999). First, there is a requirement of
memory impairment which is restrictive when applied to VaD since recent research has

shown that memory impairment may not be an early or most salient feature of this

10

dementia type (Cosentino et al., 2004; Garrett et al., 2004; Bowler & Hachinski, 2003).
Second, while individuals with AD typically evidence significant deficits on language
and semantic knowledge tasks, recent research has shown that those with VaD often
demonstrate few if any such deficits (Rockwood et al., 2001; Garrett, 2004; Cosentino,
2004). Instead, some studies have shown that, in contrast to AD, VaD is often
characterized by dramatic impairment on tests of executive fimctioning, or higher-order
drinking, encompassing domains such as cognitive flexibility, planning, and inhibition
(Cosentino et al., 2004; Garrett et al., 2004; Rockwood et al., 2001; Bowler & Hachinski,
2000). The application of the ‘Alzheimerized’ definition to VaD diagnosis may lead to
VaD being diagnosed at a relatively advanced state, or not being diagnosed at all (Bowler
& Hachinski, 2003; Rockwood et al., 2000; Erkinjuntti et al., 2000).

The diagnosis of VaD is also problematic because there is currently no ageed
upon set of criteria (Chui et al., 2000; Gold, Bouras, & Canuto et al., 2002). Several
diagnostic criteria are in widespread use, including the Hachinski Ischemic Score (HIS),
Diagnostic and Statistical Manual of Mental Disorders (DSM-IIIR, DSM-IV),
International Classification of Diseases (ICD-lO), State of Califorrnia Alzheimer’s
Disease Diagnostic and Treatment Centers (ADDTC), and the National Institute for
Neurological Disorders and Stroke—Association Internationale pour la Recherche et
l’Enseignement enNeurosciences (N INDS-AIREN). Unfortunately, the criteria are not
interchangeable and can lead to a five-fold difference in the fi'equency of classifying VaD
(V erhey et al., 1996; Wetterling et al., 1996; Chui et al., 2000; Gold et al., 2002). A
recent retrospective analysis using four classification systems for the clinical diagnosis of

VaD indicated that classification varied widely as a function of criteria used such that

11

patients diagnosed with VaD using one set of criteria were not necessarily diagnosed with
VaD using other criteria (Cosentino et al., 2004).

While criteria such as the ICD-lO allow for a differentiation of vascular dementia
into subtypes (vascular dementia of acute onset, multi-infarct dementia, subcortical
vascular dementia, and mixed or unspecified types), most do not allow for subtyping and
require evidence of significant ischemic brain injury by structural neuroirnaging (i.e.,
multiple infarcts). The limitation that strokes be clinically evident is highlighted by the
fact that only 40% of patients with VaD have such focal signs (Bowler, 2002). It is thus
argued that the problems inherent in the diagnosis of VaD may be associated with
inadequate detection and diagnosis (Bowler & Hachinski, 2003). Indeed, the current
criteria may fail to identify patients with significant cognitive loss caused by
cerebrovascular disease and the emphasis on dementia may underestimate the burden of
disease associated with early VaD. (Hachinski, 1992; 1994).

Failing to identify early VaD may distract from the focus on prevention which
may have negative consequences with respect to treatment since, in contrast to AD,
vascular risk factors are treatable and strokes can be preventable (Hachinski & Bowler,
2003; Erkinjuntti et a1, 2000). Such patients will likely not have the opportunity for
secondary preventive measures to delay or stop their progession to dementia (Hachinski,
1991, 1992, 1994; Bowler, 1993; Erkinjuntti & Hachinski, 1993; Hachinski & Bowler,
2003; Rockwood et al., 2004). Calling for an overhaul of current VaD criteria, Bowler
(2002) fervently argued that VaD reflects an “outmoded concept” since it is based on
criteria for Alzeirner’s disease and urged for increased attention to be focused on subtler

forms of vascular-related impairment.

12

A New Proposal: Vascular Cognitive Impairment (V CI)

Although vascular dementia (V aD) has been the subject of increased clinical
research (Rockwood et al., 2004; Hachirnski & Bowler, 2003), far less is known about the
factors that contribute to the development of dementia among people with chronic
cerebrovascular disease (CVD) compared to Alzheimer’s disease (AD). Bowler and
Hachinski (2002) report that the total burden of cognitive impairment of cerebrovascular
origin may be very high. Specifically, 78% of elderly patients at autopsy have evidence
of cerebrovascular disease as do over 80% of those who are demented (Bowler &
Hachinski, 2002) and about 9 in 1000 people aged 85 years or over may develop VaD
each year (Hachinski & Bowler, 2003). Just as subtle cognitive impairment may be
present several years before the clinical diagnosis of AD (Meguro et al., 2001; Masur et
al., 1994), such understated vascular-related cognitive deficits may be evident in early
VaD. The crucial question is whether this common and important problem can be
detected and halted before Significant cognitive loss has occurred.

It has been argued that vascular-related cognitive impairment (V CI) be used to
describe the early stage of VaD (Bowler, 2002; Hachinski, 1994; Erkinjuntti &
Rockwood, 2001). However, currently, no universally accepted criteria have been
formalized (mostly due to a relative lack of data with respect to early cognitive loss) and
it has been suggested that, until formal criteria are proposed, criteria for VCI should be
wide-ranging and broad (Bowler & Hachinski, 2003). Specifically, Bowler and
Hachinski (2002; 2003) indicate that individuals Should be considered for VCI when

there is any evidence of cerebrovascular disease (with or without known risk factors or

13

neuroimaging findings) in the presence of associated cognitive impairment that does not
reach the level of dementia (Bowler & Hachinski, 2003).

Using these broad criteria, preliminary data appear to support the validity of this
concept of VCI (Ballard et al., 2003; Meyer et al., 2002; Wentzel et al., 2001). For
example, Wentzel et a1. (2001) recently reported that 50% of the subjects in their sample
who met criteria for VCI developed dementia over five years. Additionally, Meyer et a1.
(2002) attempted to answer the question of whether an MCI stage precedes some cases of
VaD by longitudinally following 291 volunteer subjects over one to seven years.
Subjects were irnitially screened with the Mini-Mental Status Examination (MMSE) and
tested annually with neuroimaging (MRI or CT). Follow-up of the 73 subjects who were
diagnosed with MCI occurred every 3 to 6 months and results indicated that, of the 27
subjects who developed VaD witlnin 7 years, roughly 56% had prodromal MCI.
Similarly, Ballard et a1. (2003) showed that approximately 9% of their MCI sample
developed VaD over one year and almost half of the patients developed dementia over 5
years of follow-up. Taken together, these findings suggest that individuals with cognitive
difficulties secondary to CVD are at increased risk for further cognitive decline and
conversion to dementia. Indeed, it appears that MCI may represent a clinically
heterogeneous goup of individuals who are at elevated risk for dementia that is not
solely restricted to AD and thus the concept of VCI appears to be a clinically useful and
relevant.

Continuum of Impairment in Vascular Cogm'tive Impairment (VCI)
Hachinski & Bowler (2000) and others (Bowler, Steenhuis, & Hachinski, 1999;

Devasenapathy & Hachinski, 2000; Bowler, 2000; Bowler, Steenhuis, & Hachinski,

l4

1999; Garrett, Browndyke, & Whelihan et al., 2004) proposed that VCI exists on a
continuum that is comprised of three primary stages (brain-at-risk (R-CVD; vascular
cognitive impairment—no dementia (V aCIND); and vascular dementia (V aD)). The first
stage--brain-at-risk (R-CVD)--is comprised of individuals with cardiovascular or other
system disease processes who are at risk for developing cerebrovascular disease (with no
clinically significant cognitive or functional impairments). VaCIN D is the first
hypothesized clinical stage preceding VaD (Hachinski, 1993) and is analogous to mild
cognitive impairment (MCI) (Bowler & Hachinski, 2003). Like MCI, individuals with
VaCIND do not have Significant impairment in their ability to complete basic activities of
daily living.

Chui, Mack, Varpetian, and Mungas (2003) have conceptualized VCI with finer
degadations within the course of VaD. Their stages include:

1. brain at risk (presence of vascular risk factors)

2. early ischemic brain injury, but clinically asymptomatic

3. early symptomatic, but prior to diagnosis (cognitive decline by

neuropsychological testing)

4. clinical diagnosis of VaCIN D

5. clinical diagnosis of VaD

6. advanced dementia (severe VaD); and

7. death

According to Chui et a1 (2003) and others (Hachinski & Bowler, 2003; Rockwood
et al., 2003; Skoog & Gustafson, 2003; Inzitari, Larnassa, & Pantoni, 2003), detection

and attempts at intervention do not typically occur until individuals reach stage five or six

15

of Chui et al.’s (2003) VaD continuum, mostly due to the restrictive nature of the current
diagnostic criteria for VaD, as discussed above. Unfortunately, adequate intervention at
these stages is futile given that excessive cerebrovascular damage has typically occurred
at these later stages. While research on vascular-related cognitive impairment (V C1) is in
its infancy and only a few studies have been conducted to date, VaCIND is thought to be
the best established concept for identifying at-risk patients with cerebrovascular disease
(Stephens et al., 2004). Empirical support for the clinical concept has begun to deveIOp--
particularly in Asia and Europe-with a strong interest and emphasis on the early phases
of the disease process.
Neuropsychological Impairment in VCI

A geater understanding of neuropsychological impairment evidenced during the
early stages of VaD may have Significant clinical implications since therapeutic
interventions initiated early in the disease process may prevent progession to dementia
(Bowler, 2000; Devasenapathy & Hachinski, 2000; Garret etal., 2004). However, to
date, the neuropsychological profile of VaCIN D has not received Significant study. It has
been shown that neurocognitive difficulties associated with cardiac disease and mild
cerebrovascular disease include reduced information processing speed and reduced
cognitive flexibility (Kilander et al., 1998; Rockwood, Dobbs, Rule, Howlett, & Back,
1992; Waldstein et al., 1996). Additionally, recent neuroimaging studies conclude that
the magnitude of cognitive impairment among patients with mild cardiovascular disease
is significantly associated with neuropathological changes secondary to vascular damage
(Chui & Gonthier, 1999; Raz, Rodrigue, & Acker, 2003). As will be discussed in further

detail, research has suggested that cardiac risk factors can be associated with cognitive

l6

and neuropatlnological changes in the absence of frarnk stroke or dementia (Skoog, 1994;
DeCarli et al., 2001; Swan et al., 1998).

The finding that cognitive functions mediated by frontostriatal circuits (i.e.,
executive functioning) are most disturbed during the late phases of VaD raises the
possibility that mild changes in these cognitive functions may occur during the early
stages of this dementia process. However, it remains unclear whether clinically
mearningful changes are evident among patients in early stages of VaD. Furthermore, it
is of clinical interest to determine if the difficulties on measures of frontal lobe
functioning exist in the context of preserved function in other cognitive domains such as
memory. As noted by Garrett et al. (2004), a pattern of strengths and weaknesses on
cognitive measures may prove to be a critical determinant of diagnostic differentiation
between the various stages of vascular-related cognitive impairment, as well as
distinguishing vascular etiologies from other dementing conditions. Thus, the
identification of a pattern of neuropsychological changes associated with the early stage
of VaD (i.e., Va CIND) would greatly benefit diagnosis and facilitate intervention
(Garrett et al., 2004).

Outcome of VCI

Vascular cognitive impairment (V CI) is associated with an increased risk for
adverse outcomes. Rockwood et al. (2000) found that a failure to consider VCI
underestimates the prevalence of innpainnent and the associated risk for adverse
outcomes. Specifically, this Study compared the rates of adverse outcomes for older
patients with no cognitive impairment, VCI, and probable AD. After reassessment 5

years later, VCI was the most prevalent form of cognitive impairment among Older adults

17

aged 65 to 84 years. Most strikingly, rates of institutionalization for those with VCI were
similar to that of those with VaD, and the mortality rate for VCI patients was similar to
that of patients with AD. Data such as these support the view that increased attention on
subtle vascular-related cognitive impairment is important and clinnically relevant.
Moreover, criteria which require the diagnosis of dementia likely underestimate the
prevalence and burden of vascular cognitive disease (Hachinski, 1994; Rockwood et al.,
1994)
White Matter Lesion (W ML) Pathology

In 1986, Hachinski coined the term “leukoaraiosis” to describe abnormal cerebral
white matter changes as seen on CT scans (Hachinski, Potter, & Merskey, 1986;
Hachiski, Potter, & Merskey, 1987). At about the same time, Awad, Spetzler, & Hodak
et a1. (1986) characterized these white matter changes as seen on MRI as “Incidental
Subcortical Lesions.” Initially termed “UBO’S,” or Unidentified Bright Objects, these
changes have more recently been termed white matter lesions (WML) or hyperintensities.
They are typically found in deep cerebral areas and as cappings on the lateral ventricles
(Drayer, 1988) and are identified as signal hyperintensities on T2- and proton-density
weighted magnetic resonance (MR) scans.
Potential Causes of WML

The precise etiology of WML is debated since numerous potential causes have
been identified (Chui, 2001; Pantoni, Inzitari, & Wallin, 2000; Brown, 2000). For
example, it is thought that WML can be caused by either arteriosclerosis causing direct
occlusion of small arteries or by partial occlusion of small arteries in combination with

cardiovascular failure (ortlnostatic hypotension, carotid sinus sensitivity, and congestive

18

heart failure) (Chui, 2001). In addition, WML appear to be attributed to other alterations
in the brain, including increased fluid in the white matter, infarcts (stroke), and enlarged
perivascular spaces (Munoz et al., 1993). Moreover, WML have been posited to be
caused by subclirnical ischemia secondary to cerebral hypoperfusion (Pantoni, Inzitari, &
Wallin, 2000). Indeed, the cerebral white matter is the least irrigated compartment of the
brain and thus it may be more vulnerable to the effects of ischemia and hypoperfusion
(Brown, 2000; Pantoni & Garcia, 1997). Interestingly, the prefrontal brain regions
appear to be most vulnerable given lower vasodilatory capacity versus white matter found
in other brain regions (Brown, 2000).

Recent studies Show that the most frequently observed pathological correlates of
WML include gliosis (F azekas, Schmidt, & Scheltens, 1998), myelin pallor (Fazekas et
al., 1993, 1998; Takao et al., 1999), atrophy ofthe neuropil (F azekas et al., 1998),
breakdown of the ependymal ventricular lining (Scarpelli et al., 1994), and ischemia in
the white matter (Chui et al., 2001). Considered to be rare before neuroimaging advanced
to the MRI, WML pathology is now receiving considerable attention with respect to the
understanding of cognitive, behavioral, and motor decline seen in dementia and, more
recently, subtler neuropsychological impairment (Raz etal., 2003; Stephens et al., 2004;
Garrett et al., 2004).
Major subtypes of WML

WML, which are best identified on T2-weighted, fluid-attenuated inversion
recovery (FLAIR) and proton density magnetic resonance images, comprise both
periventricular hyperintensities (PVL) and deep white matter hyperintensities (DWML).

PVL are defined as signal abnormalities directly lining the ventricular lumen and involve

19

the white matter around the ventricles. This type of lesion is most frequently found
adjacent to the frontal horns of the lateral ventricle, although these lesions can also be
found laterally around the ventricle as well as around the posterior horns (Raz et al.,
2003)

Although WML occur in different disorders and arise from a range of
pathological processes (Everall et a1, 1999), DWML appear to be vascular in origin,
representing areas of ischennia as well as infarction (Awad et a1, 1986; F azekas et a1,
1993; Thomas et al., 2002) and thus it has been argued that PVL and DWML may differ
in their pathogenesis. DeCarli & Scheltens (2002) stress the need for these lesion types
to be analyzed separately, a recommendation which was recently corroborated by a
clinical study among patients with AD and dementia with Lewy Body (DLB) (Barber,
Gholkar, & Scheltens, et al., 2000). Specifically, in this study, total brain, ventricular
volumes, and white matter lesions were visually rated and quantified. PVL were found to
independently correlate with advancing age and increasing ventricular size (suggesting
atrophy) in all subjects. In contrast, DWML did not correlate with measures of brain
atrophy or age, but were associated with a history of hypertension. These findings
support the hypothesis that PVL and DWML are likely pathologically diverse. In
particular, PVL appear to be linked to atrophic processes involving ventricular
enlargement while DWML appear to be associated with cerebrovascular risk factors.
Relationship Between WML aniV_ascular Riskj actors

WML appear to be related to physical illnesses such as diabetes mellitus
(Longstretln et al., 1996; Longstreth et al., 2000) and reach their highest prevalence in

patients who have vascular dementia (Smith, Snowdon, Wang & Markesbery, 2000) and

20

depression (Kumar, Bilker, Jin, & Udupa, 2000). There is little controversy that WML
increase with the aging process (Gunning-Dixon & Raz, 2000; Inzitari, 2000; Pantoni &
Garcia, 1995; Schmidt, Fazekas, & Kapeller et al., 1999) and a tlnreshold effect for
cognitive decline has been hypothesized to be present (Munoz et al., 2003). However,
the debate over identified WML centers on their clinical significance and it is still unclear
whether they represent a pathognomonic sign for brain disease (Munoz et al., 2003;
DeCarli & Scheltens, 2002; Inzitari, 2000).

While much attention has focused on the rising prevalence of AD (Marshall,
Bradley, & Marshall et al., 1988; Jones, Lythgoe, & Horsfield et al., 1999), the impact of
cerebrovascular risk factors on cognitive function is just being recognized (Markus,
Lythgoe, & Ostegaard et al., 2000; Schmidt, Fazekas, & Kapeller et al., 1998). Studies
have shown that advancing age and hypertension represent the two most significant risk
factors for stroke as well as cardiovascular and peripheral vascular disease, suggesting
that these disorders may share a common mechanism of vascular insult (Pantoni &
Garcia, 1995; Wahlund, Barkhof, & Fazekas et al., 2001; F azekas, Schmidt, & Alavi,
1998). Although less attention has been paid to the possible spectrum of brain injury
resulting fiom other cerebrovascular disease risk factors, hypertension has been shown to
be associated with increased WML volume (Chabriat, Pappata, & Poupon et al., 1999;
Hanyu, Asano, & Sakurai etal., 1999). Moreover, the fact that WML significantly
predict future stroke (Awad, Masaryk, & Magdinec et al., 1993) lends firrther support to
the notion that WML are likely part of a spectrum of vascular-related brain injury (Roob,

Lechner, & Schmidt et al., 2000).

21

In the Rotterdam Scan Study, a large population study in Europe, 27% of subjects
between the ages of 65-84 years of age had evidence of WML on MRI which increased in
prevalence Significantly with age (Breteler, van Amerongen, & van Swieten, etal., 1994).
Elevated blood pressure and cholesterol levels as well as hypertension were significantly
associated with the presence and severity of WML. Also related to increased WML were
history of stroke or myocardial infarction, and later studies Showed an association with
atrial fibrillation and carotid artery atherosclerosis (de Leeuw, de Groot, & Bots et al.,
2000; de Leeuw, de Groot, & Oudkerk, et al., 2000). Taken together, these studies
support previous observations that WML may represent a general measure of vascular
disease (Manolio, Burke, & O’Leary, etal., 1996; O’Leary, Polak, & Kronmal, et al.,
1996; O’Leary, Polak, & Kronmal, et a1. 1999).

A recent longitudinal study of the National Heart and Lung and Blood Institute
(NHLBI) found a relationship between cardiovascular risk factors (i.e., high blood
pressure) occurring in middle age and the development of later-life brain insult (DeCarli,
Miller, & Swan, et a1, 1999). Specifically, there were strong correlations between mid-
life systolic blood pressure (SBP) and later-life WML volume. Additionally, those with
higher levels of WML volume at advanced ages had Significantly higher rates of coronary
artery disease compared to those with no WML. Results of this study suggested that
WML may be the consequence of blood pressure changes beginning early (in middle age)
and provides further evidence that WML appear to be associated with incident vascular
disease. The authors of this study suggest that early and aggessive treatment of high
blood pressure in middle age might significantly reduce vascular disease in later life

(DeCarli et al., 1999).

22

VCI and White Matter Lesions

The contribution of WML to cognitive impairment is beginning to receive
increased attention. Recent research has shown that 33% to 97% of cases of VaD
evidence extensive WML (W etterling, Kanitz, & Borgis, 1996; Pohjasvaara, Mantyla,
Ylikoski, Kaste, & Erkinjuntti, 2000; Campbell & Coffey, 2001) and less extensive white
matter lesions are thouglnt to be evident in VCI (Erkinjuntti & Rockwood, 2001;
Pohjasvaara et al., 2000). Bowler and Hachinski (2003) have reported that there is a need
for research to focus on WML in terms of their association with subtle forms of cognitive
impairment as well as how best to define the borders between normal aging, VCI, and
VaD (Bowler & Hachinski, 2003).

The presence of vascular risk factors or cerebrovascular events in general may be
sufficient to be diagnosed with VCI (Erkinjuntti & Rockwood, 2001). Although current
criteria require evidence of stroke for a diagnosis of VaD, some studies are showing that
silent infarcts and white matter lesions are relevant and may comprise early cases of
VaD, or individuals with VCI (Erkinjuntti & Rockwood, 2001). Indeed, it has been
shown that evidence of ischennic disease on neuroimaging (in the absence of stroke and
atrophy) appear to be associated with cognitive impairment (Bowler, 2002). In fact,
some researchers (Phillips & Mate-Kole, 1997; Emory, Gillie, & Ramdev, 1994) have
argued that WML represent the beginning stage of a series of pathological processes
which precede tissue infarction and suggest the term “pre-infarct state” to account for
cognitive changes in patients who evidence WML, regardless of whether they evidence
cerebral infarcts. Current research is beginning to focus on more subtle lesions and how

they may be associated with neuropsychological impairment (Raz et al., 2003; Stephens

23

_—_

 

 

 

 

 

et al., 2004; Garrett et al., 2004). According to Sachdev & Looi (2003), more clearly
defined differentiation of VaD from AD is likely to be geatly influenced by the extent of
WML in the two goups.

In a longitudinal study of 27 individuals with MCI by Wolf et a1. (2000), it was
demonstrated that subjects who developed dementia afier 2-3 years had more severe
white matter lesions. In addition, there was an inverse relationship between WML and
the degree of temporal lobe atrophy in those who progessed to dementia during the
follow-up interval; thus, high WML severity was associated with lesser degee of
temporal lobe atrophy and higher global cognitive performance, as measured by the
SIDAM (Structured Interview for the Diagnosis of Dementia of Alzheimer Type). This
illustrates how Subcortical (V CI) and Cortical seem to differ in their course and clirnical
picture. Specifically, it appears likely that those with geater levels of WML’S and less
medial temporal lobe atrophy may progess to VaD, whereas those with little to no WML
and evidence of significant medial temporal lobe atrophy may be more likely progess to
AD. Thus, WML appear to play a role in the dementia process and may accelerate
decline in individuals with mild cognitive impairment.

Neuropsychologi cal Deficits Associated with WML

 

Early research involving white matter lesions was quite mixed. While some
studies pointed to a relationship between such lesions and neuropsychological
functioning (Almkvist et al., 1992; Breteler et al., 1994a, 1994b), many found no
relationship after controlling for age (Hunt et al., 1989; Tupler et al., 1992). It has been
suggested that earlier research was hampered by many factors, including small sample

sizes (Erkinjuntti et al., 1994; Mirsen etal., 1991), the use of inadequate and insensitive

24

tests that are rarely used with older populations (Mirsen et al., 1991; Hunte et al., 1998),
differences in subject settings (Rao et al., 1989; Hershey et al., 1987), and the use of CT
scans which are not very sensitive to WML (Diaz et al., 1991; Miyao et al., 1992; Skoog
etal., 1993; Hershey et al., 1987).

Contrary to previous studies, more recent reports have demonstrated that WML
appear to disrupt cognitive functioning and thus may have clinical significance (Amar et
al., 1996; Breteler et al., 1994a, 1994b, Garrett et al., 2004; Rockwood et al., 2001).
Indeed, although the effects are often subtle and mild, there has been a link between early
VCI and white matter lesions and cognitive loss (Bowler, Hachinksi, Steenhuis, & Lee,
1998; Garrett et al., 2004; Rockwood et al., 2001). Consistent with deficits often
evidenced in VaD, older individuals with numerous white matter lesions appear to
perform worse on tests of speed and visuospatial function, even after controlling for age
(Garrett et al., 2004; Cosentino etal., 2004). Additionally, while few studies have
examined the distribution of WML and their effect on cognitive function, there is some
evidence that the extent and pattern of WML is important in influencing cognitive
performance (Raz et al., 2003). Specifically, associated deficits appear to be of a frontal-
subcortical nature, affecting neuropsychological functions such as speed of processing
and executive functioning; this relationship appears to be stronger with respect to deep
white matter lesions (DWML) versus periventricular (PVL) WML (Garrett et al., 2004;

Raz et al., 2003), although data is preliminary.

Recent studies have noted that the neuropsychological impairment sometimes
seen in association with WML is similar to other types of subcortical dementing illnesses

(Giovannetti-Carew, Lamar, Cloud, & Libon, 1997; Padovani et al., 1995; Libon et al.,

25

2001). Libon et al (1997) found a dissociation between tests of verbal declarative
memory and executive systems functioning such that patients with vascular dementia
obtained lower scores and made more perseverative errors on executive systems tests, but
exhibited less forgetting, obtained higher tests scores on delayed free/cued recall and
recognition test conditions, and made fewer intrusion errors than AD patients. This study
also found that the pattern of neuropsychological impairment seen in dementia associated
with subcortical WML was more similar to other subcortical dementing illnesses, such as
Parkinson’s disease (PD) and Huntington disease, than cortical dementing illnesses such
as AD. Similar finding have been reported by Doody, Massman, Mawad, and Nance

(1998)

In a more recent study by Libon and colleagues (2001) subjects were assessed
with regards to WML quantity independent of clinical diagnosis. Individuals with a
diagnosis of either vascular or Alzheimer’s dementia were placed in goups representing
either minimal-mild WML or significant WML. After being compared to patients with
PD, it was found that neuropsychological testing failed to distinguish between those with
PD and those with significant WML; additionally, those with significant WML
demonstrated disproportionate innpairment on tests of visuoconstruction and executive
systems functioning, whereas patients with little WML showed geater impairment on
tests of declarative memory and semantic knowledge. These findings point to a pattern
Of cognitive impairment associated with significant WML as distinctly different when

compared to AD.

26

WML and Depression

Depression is a common psychological disorder in late life, with prevalence rates
between 2% and 3% for major depression and 12% -- 15% for all depressive syndromes
(Beekrnan et a1, 1999). Recent growing evidence suggests that there may be a subtype of
depression in later life—commonly termed “vascular depression”--that is characterized
by a distinct clirnical presentation and an association witln cerebrovascular disease and
thus the presence of white matter lesions (Alexopoulos et a1, 1997; Krishnan, Hys, &
Blazer, 1997; Krishnan, 2000). Similar to vascular dementia, it is thought that vascular
disease somehow predisposes, precipitates, or worsens pre-existing depression. Indeed,
depression is increased in dementia, with much higher rates in vascular dementia as
opposed to Alzheimer’s disease (Newman, 1999).

It is known that depression is highly prevalent in patients with diseases involving
the cardiovascular system, such as hypertension, diabetes, coronary artery disease, and
myocardial infarction (Carney et a1, 1987; Rabkin et a1, 1983). In addition, clirnical
Studies have shown compelling lirnks in both directions between some vascular disorders
and depression. Indeed, depression is fi'equently elevated following a myocardial
infarction and independently predicts an increase in mortality of about 350% during the
next six months (Glassman & Shapiro, 1998). Conversely, longitudinal studies of
initially healthy subjects have shown that depression itself can increase the risk of
subsequent myocardial infarction or coronary artery disease by 300-400% (Glassman &
Shapiro, 1998; Ford et al, 1998). Furthermore, hypertension is associated with a three-

fold increase in major depression (Rabkin, Charles, & Kass, 1983) and depressive

27

symptoms appear to double the rates of subsequent hypertension (Jonas, Franks, &
Ingam, 1997).

Other evidence that supports the vascular theory of late-life depression stems
from research which has demonstrated that depression is a common occurrence in many
patients in their first year post-stroke (House et a1, 1991), with rates of depression
reported ranging from 20-65% (Gordon & Hibbard, 1997; Pohjasvaara et a1, 1998). In
addition, the Alameda County Study recently demonstrated depressive symptoms to be an
independent risk factor for subsequent fatal stroke, increasing fatal stroke by 50% over
about 30 years (Everson et a1, 1998). Finally, silent stroke has been associated with
depression, even in the absence of a genetic predisposition or significant psychosocial
events (Fujikawa et a1, 1997).

In addition to the clinical work cited earlier, many MRI studies have consistently
shown a relationship between WML and depression (V idebech, 1997; De Groot, 2000).
WML associated with depression appear to be located predominantly in the deep white
matter (Alexopoulos, Myers, & Yong, et al., 1997; Krishari, Hays, & Blazer et al., 1997;
O’Brien et a1, 1996). Interestingly, De Groot (2000) found that this relationship occurred
independent of cognitive status, indicating that depressive symptoms may not be caused
largely by a psychological reaction to declining cognitive function. In addition, the
relationship between WML severity and depressive symptoms did not change when other
possible confounders were controlled for such as presence of stroke and cortical atrophy.

As life expectancy continues to increase dramatically, the severity and incidence
of neuropsychological impairment and dementia are expected to rise; the prevalence of

vascular disease is estimated to increase as well. Thus, there is a significant need to

28

elucidate the relationship between alterations in cognitive state and white matter lesions.
In addition, now that it is clear that Mild Cognitive Impairment represents a useful

clirnical entity, a focus must elucidate what constitutes MCI in an effort to identify those
patients who may be helped by medical intervention before a full dementia syndrome is

evidenced (i.e. those with cerebrovascular underpinnings).

29

CHAPTER 2
AIMS AND HYPOTHESIS

Primary Aims.

The primary aim of this study was to explore the relationship between white
matter lesions (WMLS) and neuropsychological functioning in a sample of at-risk, pre-
clinical, older adults who presented at the MSU Geriatric Neurology clirnic and were
diagnosed with Mild Cognitive Impairment (MCI) between the years 2000 and 2004.
Structural MRI scans were analyzed for presence/absence of white matter lesions and
volumetric analyses were used to assess the extent of pathology. In addition,
neuropsychological test performance using the CERAD neuropsychological battery,
Trails A and B, and the Stroop Color and Word Test were used to assess the relationship
between WMLS and cognitive functioning in individuals with MCI.

Specific aims of the current study include the following:

1. Provide a detailed description of an MCI population in terms of
neuropsychological performance. All cognitive data will be analyzed in order to evaluate
whether MCI is indicative of a homogeneous goup (i.e., evidencing a similar
neurocognitive profile across neuropsychological domains) or, alternatively, if those with
MCI represent a heterogeneous population with differing neuropsychological profiles.

11. Provided that reliable, distinct goups exist witlnin the sample (based on
neuropsychological performance), an analysis will be conducted to evaluate whether and
how the goups differ with respect to total WML volume. Identifiable differences would
provide evidence that MCI may arise through multiple etiologies rather than being
exclusively attributed to Alzheimer’s disease. While many different etiologies for MCI

could be expected, major goups anticipated to emerge from the analyses (based on

30

current base rates for dementia) include a subpopulation demonstrating features of a
cortical, degenerative etiology (i.e., pre-Alzheimer’s dementia with corresponding
neuropsychological deficits) as well as a subpopulation characterized by early vascular
dementia (based on increased WMLS and their associated neuropsychological profile).

111. To date, research investigating neuropsychological deficits associated with
WML types (periventricular and subcortical hyperintensities) has been largely
contradictory and inconclusive. Therefore, the current study seeks to elucidate Specific
patterns of impairment associated with both lesion types by applying newer methodology
(i.e., volumetric analyses of WML) and relating neuroimaging findings to
neuropsychological performance.

HypotheSes.

Based on the theory that cognitive impairment associated with white matter
lesions (WML) can be viewed within the context of a subcortical dementing syndrome
(i.e., vascular dementia) the following hypotheses will be tested in this study:

1) Since it is thought that WMLS interrupt neural communication, it is expected

that increased WML volumes will be associated with geater global cognitive

impairment as evidenced by decreased performance on the MMSE.

2) It is expected that MCI does not represent a homogeneous population (i.e.,

solely Cortical) and will be better characterized as a heterogeneous goup based

on neuropsychological scores. Using cluster analysis, it is expected that the
following two major goupings will emerge:
2a) One subpopulation is expected to evidence a neuropsychological

profile consistent with early vascular dementia (Subcortical), with primary

31

deficits demonstrated on tasks of executive functioning, processing Speed,
and visuospatial/constructional abilities. This particular profile is
expected given that vascular dementia arises from disruption of fi'ontal-
subcortical circuits, which are thought to be related to these specific
aspects of neuropsychological performance.
2b) A second goup with a profile similar to those with early Alzheimer’s
disease (Cortical) is hypothesized to exist within the sample. Based on
numerous studies suggesting early degeneration of the medial temporal
lobes in AD, specific impairments in memory and language (naming and
fluency) are expected. Indeed, the medial temporal lobes are known to
largely mediate memory and language functions.
3) If the expected goups are found within the sample (Subcortical and Cortical),
the goups will be furtlner assessed to determine whether they differ in terms of
WML volumes. A functional dissociation of neuropsychological performance is
expected with regards to levels and severity of WML as delineated below:
3a) Given that WML pathology has recently been shown to be associated
with various vascular conditions such as hypertension and stroke, it is
expected that the Subcortical goup (based on neuropsychological profile)
will evidence geater lesion load. This relationship is theorized to exist
based on disruption of fi'ontal-subcortical circuits witlnin the brain
secondary to WML pathology.
3b) Conversely, it is expected that subjects who appear more AD-like in

terms of neuropsychological profile (geater deficits in language and

32

memory relative to executive functioning, speed of processing, and
visuospatial skills), will evidence little or no WML pathology.
4) Although it is clear that two very distinct types of WML exist (deep white
matter lesions (DWML) and periventricular white matter lesions (PVL)), Specific
associations with neuropsychological impairment are not well known. Therefore,
both types of WML will be used to evaluate neuropsychological finnctioning in
this sample as a whole and the following relationships are hypothesized to exist:
4a) DWML will be associated with geater overall neuropsychological
impairment across cognitive domains versus PVL, irrespective of goup
membership. The relationship is posited to exist based on the theory that
DWML may interrupt a larger number of critical neuropatlnways that are
important for complex neuropsychological fimctioning.
4b) It is expected that DWMLs will more strongly predict executive
impairments, Slowed processing speed, and visuospatial/constructional
difficulties, but will not be related to impairments on tests tlnought to be
more sensitive to cortical or temporal lobe fimctions such as memory and
language. Consistent with neuropsychological deficits evidenced in VaD,
the relationship has been posited due to disruption between frontal-

subcortical circuits associated with subcortical WML.

33

CHAPTER 3
METHOD

Sannple

The sample for this study consisted of 70 community-dwelling older adults (aged
55 and older) who presented for cognitive complaints at the MSU Geriatric Neurology
Clinic (an outpatient subspecialty of the MSU Neurology and Opthalmology
Department). Subjects meeting criteria for MCI were included in this study after a
careful health screening (described below). The participants consisted of 33 males (47%)
and 37 females (53%), and ranged in age fi'om 55 - 88, with a mean age of 75 (SD = 7.9).
The mean years of education completed was 13.3 (range = 8 - 24, SD = 3.4). Finally,
MMSE scores ranged from 24-30 (per outlined criteria for MCI described below), with a
mean of 26.3 (SD = 1.6). All patients who underwent neuropsychological testing and
neuroimaging (MRI) as part of their clinical care and data collected between February,
2000 and July, 2004 were included in this study. The average number of months between
neuroimaging and neuropsychological assessment was 4 months (range 1 week to 13
months), with neuropsychological testing typically taking place before imaging was
completed. The procedures in the present study were approved by the Michigan State

University Committee on Research Involving Human Subjects (U CRIHS).

Medical history was obtained for the patients through clinic records and patients
were carefully screened for and excluded from the study based on any current or past
diagnoses of neurological or psychiatric disorder, stroke, thyroid disease, diabetes, known
head injury, or any significant visual or auditory impairment which precluded them from
participating in neuropsychological testing. Additionally, individuals not meeting

neuropsychological criteria for MCI (i.e., demented or cognitively normal) were not

34

included in this study. Neuropsychological classification of MCI was conducted
independently of the white matter lesion (WML) quantitative analyses. From the original
sample of 99 patients who met criteria for MCI using Petersen’s guidelines (Petersen,
1999, 2001; see below), 15 were excluded fiom the study due to severe health conditions
such as multiple sclerosis or stroke, as well as psychiatric reasons (i.e., schizophrenia,
severe depression). Seven subjects were excluded because only CT scans were
available, making reliable identification of WMLS difficult. Three individuals identified
as outliers were excluded based on extreme white matter pathology found on MRI
indicative of white matter disease. Finally, 4 subjects were excluded based on poor scan
quality due to movement-artifact or insufficient image sequences to perform all of the

analyses required for this study.
Procedures
Diagnosis of Mild Cognitive Impairment (MCI)

A diagnosis of MCI was based on meeting all of the following criteria delineated
by Petersen (2004): 1) subjective patient memory complaints; 2) normal activities of
daily living; 3) absence of diagnosable dementia; 4) normal MMSE score as determined
by a score of 24 or geater; and 5) mild quantifiable impairments of cognitive function;
specifically, geater or equal to 1.2 standard deviations below the mean on any

neuropsychological subtest.

To identify participants with MCI, all neuropsychological scores were
standardized with a z-score transformation on the basis of the CERAD or other normative
data of the neuropsychological tests (Welsh et al., 1994), following the guidelines for

such calculations as outlined by Moser et a1. (2000). Scores that reflected number of

35

errors or response times were multiplied by -1, so that negative z-scores consistently
reflected poor performance. Impairment on a single test or subtest was operationally
defined as a z-score of -l .2 or lower (after controlling for age, education, and gender
using adjusted norms), indicating a level of performance worse than 88.5% of the
population (mildly to moderately impaired range).
Neuropsychological Test Battery

Prior to testing, all participants were interviewed in order to obtain necessary
demogaphic information. A neuropsychological test battery and a depression screen
(Geriatric Depression Scale) were administered. All cognitive tasks were administered to
each participant individually by a technician who was blind to the participant’s medical
status and MRI results. All tests were administered in standard paper/pencil format and
the sequence of neuropsychological tests was the same for all subjects. The battery of
tests administered was selected to assess a broad range of neuropsychological functions
including attention, executive functioning, language abilities, memory, and
visuospatial/construction. Specifically, the neuropsychological battery of the Consortium
to Establish a Registry for Alzheimer’s Disease (CERAD) neuropsychological battery
(Welsh et al., 1994), Trails A and B, and Stroop Color Word Test were administered.
General cognitive status was assessed with the Mini-Mental Status Examination (MMSE)

(Folstein, Folstein, & McHugh, 1976).

The CERAD is a brief neuropsychological assessment battery which was
developed for the diagnosis and staging of dementia (Morris et al., 1989). Altogether, it
is a reliable, well-standardized, and normed battery which includes seven individual tests

that tap several domains including executive functioning, memory (Word List Memory

36

Task; Welsh, Butters, et al., 1994), praxis/visuoconstruction, attention/orientation, and

language functioning (naming, animal fluency).

Specific test components of the CERAD include: Verbal fluency (Isaacs &
Akhtar, 1972) that requires naming as many animals as possible in 60 seconds; Modified
Boston Naming (Kaplan, Goodglass, & Weintraub, 1983) that requires nanning 15
drawings (of the original 60) ranging in fi'equency of occurrence in common usage; Word
List Learning that requires immediate recall of words in 10-item list, presented in random
order on three consecutive occasions; Word List Recall that requires free recall of words
on 10-item list; Word List Recognition that requires identification of the original 10
words of Word List from a list of 20 words; Mini-Mental State Examination (MMSE)
(Folstein, Folstein, & McHugh, 1975) that requires answering questions and/or
performing activities related to several areas of cognitive functioning; Constructional
Praxis (Rosen, Mohs, & Davis, 1984) that requires copying four basic geometric designs;
and Constructional Praxis Recall that requires recall of the geometric designs drawn in

the original condition;

In this study, the CERAD battery was supplemented by the inclusion of
additional common neuropsychological tests to augnent the assessment of processing
Speed and executive functioning (Trails A and B and Stroop Color Word Test). Trails A,
a test of psychomotor Speed, involves visual scanning and drawing lines between
numbers fiom 1-20. Trails B, a measure of executive finnctioning, is a more complex
version of Trails A which requires the subject to mentally alternate drawing lines
between numbers and letters, in chronological and alphabetical sequence. To control for

Speed of processing, time to complete Trails A was subtracted from time to complete

37

Trails B. Finally, the Stroop Color and Word Test is a measure of cognitive flexibility
and inhibition ability. This timed test measures the subject’s ability to inhibit the

tendency to read words and, instead, focus on the color of inconguent word/color pairs.
MRI Protocol

All MR imaging was performed on 1.5-T Signa scanners (General Electric,
Milwaukee, WI). WML volumes were estimated fiom T2-weighted axial F luid-
attenuated inversion recovery (FLAIR) images with the following parameters: FOV = 20
x 20 cm; matrix = 256 x 256; flip angle = 90 degrees, TE=142 ms, TR=10000 ms, 5 mm
slice thickness with no interslice gap. All MR scans were examined for space-occupying
lesions, and all questionable cases (lesions or other neurological conditions such as white

matter disease) were evaluated by a neuroradiologist (Mark Delano, MD).
Quantification of Wlnite Matter Lesions

WMLS have been shown to be better identified on MR versus CT scans (Kirshner
et al., 1985; Paty et al., 1998); tlnerefore, only MR scans were considered for evaluation.
The current study employed a semi-automated volumetric approach (Pixel Thresholding)
using T2-weighted FLAIR images, a methodology which was recently shown to be the
most reliable approach for the analysis of WML when compared to other image types and
traditional quantitative visual rating scales (Price, Schmalfuss, Sistrom, 2005). Training
to reliably measure WML took place at Wayne State University, following the protocol
established by Raz et a1. (2001) and Gunning-Dixon and Raz (2000) who have published
extensively on white matter pathology. Volumetric measures of WML were obtained
based on 17-21 axial, T2-weighted FLAIR images per subject using GE's Advantage

Workstation software (v. 4.2). This user-automated progarn is widely used to complete

38

volumetric analyses on MR images. Hyperintense regions, defined as circumscribed
areas of increased signal intensity within the white matter, were identified and measured
on axial slices of the T2-weighted FLAIR images beginning at the most inferior slice on
which the inferior horn of the lateral ventricles could be seen. Hyperintensities were
coded according to their presence, volume, and type (periventricular (PVL) and deep
white matter (DWML). Briefly, WML were considered to be PVL if the largest diameter
was abutted next to the ventricular lining; otherwise, they were considered to be
subcortical. All questionable cases were resolved by consulting an experienced
neuroradiologist and the correlative images in an MRI neuroanatomy atlas (Damasio,
1995; DeArmond, Fusco & Dewey, 1976; Montemurro & Brurni, 1988; Ono, Kubik, &
Abernathy, 1990).

As described in detail by Raz, Rodrigue, and Acker (2003), the total volume of
each region in cubic centimeters was calculated by multiplying the summed pixel cross-
sectional area in square centimeters by slice thickness in centimeters. Intra-rater and
inter-rater reliabilities for this method have been shown to be high (DeCarli et al.,
1992b). Since two operators completed the WML measurements, inter-rater reliability
coefficients were computed based on a random sample of 5 traced brains. The intraclass
correlation formula for two random raters (ICC(2); Shrout & Fleiss, 1979) was used, and
the resulting reliability estimates for all regions of interest and types of WML exceeded
0.80. Finally, relative head size was controlled for by taking into account gender in all

WML analyses.

39

Statistical Analyses

A number of statistical procedures were used to address our hypotheses and a)
evaluate the presence of distinct goups within the sample based on neuropsychological
performance across domains, b) examine the role of white matter lesions on cognitive
performance, and c) furtlner characterize tlnese goups based on neuropsychological
profile and white matter pathology. Additionally, statistical analyses were conducted to
investigate the role of deep white matter lesions (DWML) and periventricular white
matter lesions (PVL) on cognitive performance. Cluster analysis using Ward’s method,
discriminant analysis, analysis of variance (ANOVA), and multiple regession were used
for this study and all statistical analyses were performed using the Statistical Package for
the Social Sciences for Windows (SPSS, 2004). Before beginning any statistical analyses,
missing data were estimated based on the expected maximization algorithm using the
computer software program Systat (Version 11). Approximately 5% of the data for two
variables (Stroop and GDS) needed to be estimated (because these instruments were
added to the clirnic’s neuropsychological battery a few months after data collection
began).

Correlational analyses were conducted to determine whether total lesion load was
associated with poorer performance on the MMSE, a global measure of cognition
(Hypothesis 1). Cluster analysis was performed in order to assess for the presence of
goups based on neuropsychological performance (Hypothesis 2). Specifically, Ward’s
minimum variance cluster analysis (Ward, 1963; Kaufman & Rousseau 1990) was used
to identify patterns of impairment. This type of hierarchical agglomerative method takes

into consideration all possible combinations of profiles, computes a distance measure

40

(squared Euclidean distance) between profiles, and combines the pair with the smallest
sum of squared difference to the first cluster. Profiles are continued to be combined
which mirnimizes the increase to the witlnin-goup Stun of squared variance. Ward’s
technique parcels the total sum of squares into within-goup and between-goup
variability, thus yielding a measure of the percentage of variability accounted for by the
clustering solution.

Before beginning the cluster analysis, composite scores were computed for the
various cognitive domains of interest. When performance in the various cognitive
domains was assessed by more than two tests, the test scores were combined (their
significant correlations supported this approach). Specifically, two composite scores
were computed (one for executive functioning and one for memory). A composite score
for executive finnctioning was computed due to the presence of a high correlation
between total time to complete Trails B and the Stroop Color-Word Test Interference
score (r = .69, p < 0.001). Additionally, a composite score for memory was computed by
comng the CERAD word list delayed and recognition memory scores as well as the
praxis (visual) memory score, all of which were moderately correlated (between .26 and
.42; p < .05).

All neuropsychological scores were converted to z-scores before being entered as
variables for the cluster analysis [(Memory composite (Mem), Executive Functioning
composite (EF), Processing Speed (Speed), Visuospatial/Construction (VS),
Confi'ontation Naming (Naming), and Verbal Fluency (Fluency)]. The number of
clusters was set a priori to two to reflect the expected tlneoretical distinction between

cortical dysfunction (Cortical) and subcortical impairment (Subcortical). However, given

41

that patients may exhibit cognitive profiles consistent with both types of pathology, an
additional cluster analysis was performed with the number of clusters set to three
(allowing for the possibility of a “mixed” profile to emerge). A comparison was made
between all cluster structures to identify best data fit. In all cluster analyses, no
constraints were imposed on the numbers of participants in each cluster or the profiles of
neuropsychological performance across goups. Thus, the data were allowed to
empirically define the subgoups to address the question of whether the cognitive profiles
exhibited by patients with MCI segegate into theoretically meaningful subsets. Finally,
to assess whether goup differences based on cognitive performance were statistically
meaningful, a 3 (goup) X 6 (cognitive domain) MANOVA was conducted. Post hoc
analyses were employed using Tukey’s HSD to evaluate how each goup differed with
respect to neuropsychological functioning.

To determine whether neuropsychological scores (predictors) can be combined to
predict goup membership reliably and whether neuropsychological test scores accurately
predicted the cluster goups (classification), discriminant analysis was performed.
Discriminant analysis (DA) maximizes goup differences by creating a weighted linear
combination of the predictor variables (Kleinbaum et al., 1988), such that differences
between goups are maximized using weights in the discirninant function. The Wilk’s
lambda coefficient was computed to test the significance of the discriminant firnction and
the partial Wilk’s lambda was used as a measure of the contribution of individual
neuropsychological variables to the discriminant function (Norusis, 1988). The partial
Wilk’s lambda of a given variable ranges between 1 (when all goup means are equal)

and 0 (when between goups variability largely outweighs within goups variability).

42

To determine whether goups differed on white matter lesion load (Hypothesis 3),
a univariate analysis of variance (AN OVA) was conducted. Group membership was
entered as the independent variable and total lesion load was entered as the dependent
variable, while controlling for age, education, gender, and depression. All goup
differences were evaluated in post hoc analyses using independent samples t-tests. All
differences were considered statistically significant at p < 0.05.

T o evaluate Hypothesis 4, a series of linear regession analyses was used to
determine which lesion type (DWML or PVL) predicted poorer overall
neuropsychological performance across cognitive domains. Lesion type was entered as
the independent variable and overall neuropsychological functioning (NP) was entered as
the dependent variable (computed as a composite variable from all neuropsychological
tests), after controlling for age, education, gender, and depression in the first block.
Additionally, a 3 (goup) x 2 (lesion type) MANOVA was conducted to examine
differences in the severity of WML (DWML and PVL) across goup. Finally, several
linear regession analyses were performed to determine the effect of DWML on each

neuropsychological measure or domain.

43

CHAPTER 4
RESULTS

Table 1 provides descriptive statistics and intercorrelations for all variables of
interest in this study. As can be seen in the table, there was a reasonable amount of
variability in all of the measures with two exceptions. Variability was lacking for the
MMSE (possibly related to its truncrated range (24-30) per MCI criteria, and some of the
CERAD neuropsychological measures (i.e., low ranges for memory and naming).
Ceiling effects can be expected given that the sample is comprised of individuals who,
despite some mildly to moderately impaired scores, are relatively high functioning. As
might be expected, higher age was associated with overall poorer neuropsychological
functioning. In addition, higher levels of education and lower levels Of depression
appeared to be strongly associated with higher performance on overall
neuropsychological functioning as well as lower volumes Of white matter lesions
(regardless of type). The results of more detailed analyses, which take into account
Specific neuropsychological variables and cluster goupings are found below.
Hypothesis 1

A Pearson’s correlation coefficient was computed to test the hypothesis that
WML would be associated with MMSE performance. Results indicated that total WML
was not found to be significantly correlated with MMSE (r = -.122).

Hypothesis 2:

To determine whether distinct goups could be identified within the sample based
on cognitive performance, a cluster analysis was performed. Initially, the number of
clusters was set a priori to two, reflecting the theoretical distinction between cortical

dysfunction (Cortical) and subcortical impairment (Subcortical). Results indicated that

44

the 2-cluster solution produced widely unbalanced goups (goup l, n = 60; goup 2, n =
10) and the resulting dendogam indicated that goup 2 could be Split into two distinct
subgoups. Consequently, an additional cluster analysis was performed with the number
of clusters set to three, allowing for expression of geater heterogeneity in the sample.
The resulting solution produced three goups (n = l8, l8, and 34). The means and
standard deviations for each of the three goups on the six cognitive variables are

presented in Table 2.

The profiles of neuropsychological performance for the different clusters
conformed to expectations concerning the existence of both cortical and subcortical
goups. That is, based on group means, goup 1 (labeled “Cortical”), demonstrated a
neuropsychological profile consistent with a cortical process, performing significantly
worse on tests of memory and language (fluency and confrontation nanning). Conversely,
goup 2, labeled “Subcortical,” demonstrated the opposite pattern of results, with poorer
executive functioning (EF), visuoconstruction (VS), and processing speed (Speed). Both
the Cortical and Subcortical goups exhibited impaired fluency, with the Cortical goup
performing more poorly. The third goup exhibited only impaired memory and was
labeled “Amnestic.” Interestingly, the Amnestic goup performed better than the other
two goups in most domains (i.e., fluency, processing speed, and visuoconstruction). A
gaphical representation of the neuropsychological performance of each goup can be

found in Figure 1.

One-way AN OVA indicated that each cluster goup differed with respect to
overall neuropsychological functioning (F(2, 67) = 49.5, p < .001). Results from a

MAN OVA with goup as the independent variable Showed that the goups differed

45

significantly on EF, VS, Speed, VS, Fluency, and Naming (p < .001). Additionally,
goup differences existed based on memory performance. However, memory
performance was the weakest predictor (F = 3.26, p < .05) and post hoc analyses using
Tukey’s HSD demonstrated that the Cortical and Subcortical goups did not differ on this
domain. Furtlnermore, the Amnestic goup performed significantly better on fluency than
botln Cortical and Subcortical goups (p < .001). It is important to note that the Cortical
goup and the Subcortical goup did not differ significantly on fluency performance.
Finally, the Cortical goup performed significantly worse on naming than the other 2
goups (p < .001); however, no significant differences were found between the

Subcortical and Amnestic groups on this measure.

Demogaphically, the clusters differed significantly by age (F0, 67) = 4.19, p <
.02), education (Fa, (,7) = 8.21, p = .001, and levels of depression (Pa, (,7) = 9.24, p < .001).
However, no goup differences were found based on MMSE scores or gender. Table 3
includes the demogaphic data for the 3 goups defined by the cluster analysis. Post Hoc
tests using Tukey’s HSD demonstrated that the Cortical goup was significantly older
than the Amnestic goup (p < .02), but no difference in age was found between the
Subcortical goup and the Cortical goup or the Amnestic goup. The Amnestic goup
was significantly more educated than botln the Cortical (p = .001) and Subcortical goups
(p < .020), but there was no significant difference in education between the Cortical and
Subcortical goups (p > .05). Finally, the Subcortical goup was significantly more
depressed than the Amnestic goup (p < .001). There was also a trend toward Significant

difference based on depression between the Cortical and Amnestic goups on depression

46

(p = .075). Overall, a linear relationship was found for depression according to each

cluster goup as illustrated in Figure 2.

TO determine whether goup membership could be predicted from the
neuropsychological test scores as a whole, a direct discriminant function analysis was
performed using the six cognitive variables as predictors of membership in the three
cluster goups. Because the discriminant analysis was performed between three goups,
two discriminant functions were generated and subject classification was based on a
weighted combination of the scores along each function. The first discriminant function
accounted for 81 % of the differences among the three goups (F0, 67) = 144.4; Wilk’s
lambda = 0.085, p < .001). Individual discriminant scores correlated significantly (p <
0.001) with executive functioning, processing speed, and visuospatial/construction as
illustrated in the structure matrix (Table 4). The first discriminant function appears to be
consistent with what would be expected in subcortical, or vascular dementia (i.e., primary
deficits in executive functioning, processing speed, and visuoconstruction). AN OVA
results showed a significant goup effect of discriminant scores (F9, 67) = 144, p<0.001)
and post hoc analysis revealed that each goup differed significantly from each other.
The goup categorized as Subcortical had a Si gnificantly different pattern of
neuropsychological results with generally poorer executive functioning, processing
speed, and visuospatial/visuoconstructional skills than the other goups. The goup
discriminant score means (goup centroids) for the discriminant function are presented in

Table 5.

The second discriminant function explained the remaining variance (19%) and

was significant (Pg, 67) = 44.7; Wilk’s lambda = .476, p < 0.001). Individual discriminant

47

scores correlated significantly (p < 0.001) with naming, fluency, and memory,
neuropsychological domains which are largely implicated in a cortical dementia such as
Alzheimer’s disease. However, it is important to note that the absolute correlation for
memory was only .17, indicating that this variable was not generally important in
defining the discriminant function. The goup discriminant score means are listed in
Table 5. Post hoc analyses (Tukey’s HSD) demonstrated that, while the second function
discriminated between all other goups, it did not significantly discriminate between the
Subcortical and Amnestic goups. Using the full predictive model, the Six variables
correctly classified 18 Cortical subjects (100%), 18 Subcortical subjects (100%), and 33
Amnestic subjects (97.1%). The overall rate of correct classification was 98.6% and all

but one subject used to derive the discriminant fimction was correctly classified.

Examination of the biterritorial map of standard scores on the first and second
discriminant fimctions (Figure 3) revealed three distinct clusters consistent with a unique
neuropsychological pattern for each goup. The Amnestic goup clustered in the positive
range of the first function (Subcortical), but near zero in the second function (Cortical).
The Cortical goup clustered around zero on the first function and in the negative range
of the second function. Finally, the Subcortical group tended to cluster in the negative
range of the first function and around zero of the second function. As Shown in Figure 3,
the first discriminant function maximally separated VaD fi'om the other two goups,

while the second discriminant function discriminated AD from the other two goups.

The stability of the classification procedure was checked by a cross-validation run
using leave-one-out classification in SPSS. Classification for the originally derived cases

was 98.6% and fell somewhat to 94.3% for the cross-validation cases. One Cortical

48

person was misclassified as Amnestic and 3 Amnestics were classified as Cortical.

Overall, this pattern indicates a high degee of consistency in the classification scheme.

Although the Amnestic goup was not initially hypothesized to exist in the sample,
results provide support for Hypothesis 2 in terms of specific neuropsychological profiles
found for the other two goups in the sample (Cortical and Subcortical). Specifically, the
Cortical goup demonstrated deficits primarily in language and verbal memory with
relatively preserved executive functioning, processing speed, and visuoconstruction;

conversely, the Subcortical goup demonstrated the opposite pattern of results.

Hypothesis 3:

Given that distinct subgoups were identified within the MCI patient sample, all
three goups were assessed in terms of total WML volume. A one-way ANOVA
revealed that total WML volumes differed by goup membership (F(2, 67) = 30.82, p <
.001). Planned comparisons revealed that the Subcortical goup had significantly geater
WML volume load than both the Cortical (Q34) = -4.82, p < .001) and Amnestic goups
(Q50) = -9. l 3, p < .001). The Cortical and Amnestic goups did not differ significantly in
terms of total lesion load; however, there was evidence of a trend (t(50) = 2.00, p = .051).
Altlnough lesion volume between the Cortical and Amnestic goups did not significantly
differ, the Cortical goup demonstrated a Slightly geater lesion load than the Amnestic
goup. Moreover, the Subcortical exhibited roughly twice as much WML as the Cortical
goup and almost three times the total WML volume of the Amnestic goup as can be

seen in Figure 4. Means and standard deviations, and specific contrasts between means

that were significant can be seen in Table 6.

49

Hypothesis 4:
In order to assess whether DWML versus PVL is associated with geater

neuropsychological impairment, the effect of lesion type on overall neuropsychological
functioning (NP) was tested using correlation analyses and hierarchical multiple

regession. First, to assess overall neuropsychological functioning (NP), a composite
variable was created using all the means for the six neuropsychological variables.
Correlations were calculated to assess relationships between lesion type and overall NP.
Results demonstrated that PVL and DWML were higlnly correlated (r = .76, p < .001) and
that overall neuropsychological functioning was negatively associated with both PVL (r =
-.47, p < .001) and DWML (r = -.54, p < .001) as can be seen in Table 3. Although age
was strongly related to overall NP (r = -.41 , p < .001), no relationship was found for
gender. Additionally, higher levels of education were associated with better overall NP (r
= .42, p < .001). Interestingly, level of depression was more strongly associated with
overall NP than age or education (r = -.53, p < .001).

Linear regession was performed in order to assess prediction of overall
neuropsychological functioning fi‘om lesion types. Given the high correlations between
the predictors (PVL and DWML), separate regessions were conducted on each lesion
type. Age, education, gender, and depression were entered into the first block and lesion
type was entered in block two (see Table 7). For the first regession with PVL as the
predictor of interest, age, education, gender, and depression alone accounted for 45.9% of
the variance in the model (For, 69) = 13.81, p < .001). In block 2, PVL volume did not
significantly add to the prediction of overall NP functioning (AR2 = .09, p >.05). For the

second regression (see Table 8), with DWML as the main predictor of interest, the

50

prediction of NP functioning incremental to that of the predictors in Step 1 was
significant (AR2 = .034, p < .001), although the main effect was not large. Interestingly,
correlational analyses demonstrated that age appears to be positively associated with PVL
(r = .253, p < .05), but not DWML. Moreover, education was associated with both PVL
(r = -.346, p < .01) and DWML (r = .284, p < .02). Finally, level of depression was
strongly correlated with both lesion types (PVL, r = -.612, p < .001; DWML, r = -.63 l, p
< .001 ).

In order to assess the relationship between DWML and individual
neuropsychological variables, correlations were first examined. AS indicated by Table 3,
DWML was correlated significantly with executive functioning, processing speed, and
visuospatial/construction (p < .001). DWML was also negatively correlated with Fluency
(p < .05). A series of standard multiple regessions were performed with DWML
entered as the dependent variable and each of the neuropsychological variables as
independent variables, after controlling for age, education, gender, and depression. As
expected, DWML strongly predicted poorer Executive Functioning ([3 = -.65; AR2 = .25,

p < 0.001), Processing Speed ([3 = -.56, AR2 = .18, p < .001), and
Visuospatial/Construction (B = -.53, AR = .17, p < .001) as can be found in Tables 9-11.
In addition, DWML was not found to significantly predict Naming, Fluency, or Memory
(see Tables 12-14 which summarize the results of each of tlnese analyses). While the

relationship was not significant, there was a trend for Fluency to be negatively associated

with lesions in the deep white matter.

51

CHAPTER 5
DISCUSSION

The clirnical concept of mild cognitive impairment (MCI) has recently been
proposed to describe the transitional state between healthy aging and early dementia
(Petersen et al., 1999; 2001; 2003) and, while research to date has demonstrated
considerably high conversion rates to dementia (Petersen et al., 1999; Pahner et al.,

2003), many researchers and clinicians have conceptualized MCI as representative of pre-
clirnical AD, with little attention paid to possible heterogeneity of this vulnerable, at-risk
population. This study set out to examine whether heterogeneity exists in MCI in terms
of neuropsychological performance. Additionally, the role of white matter lesions was
investigated in order to better understand cerebrovascular contributions of pre-clinical
cognitive decline witlnin the sample. The current study investigated botln

neuropsychological and neurological correlates in a clirnical sample of patients diagnosed

with MCI in an effort to better characterize the population.

Relationship of White Matter Lesion Volume to Overall Cognitive Functioning

The first hypothesis set out to assess whether total WML volume in the brain is
associated with global cognitive impairment as measured by the MMSE. However,
contrary to expectations, the relationship was not significant. Given that the scores on the
MMSE ranged from only 24 to 30 in keeping with MCI criteria, it is likely that ceiling
effects occurred due to the truncated range of scores. Thus, identifying impairment in this
relatively high functioning sample based solely on the MMSE was difficult given the
narrow range of scores. Additionally, the MMSE is not sensitive to subtle
neuropsychological impairment and the measure largely taps verbal and memory

functions, cognitive domains which are known to be negatively impacted in early

52

manifestations of AD but not VaD. Indeed, the MMSE does not measure executive
functioning or psychomotor speed and minimally measures visuospatial/constructional
skills, major areas of neuropsychological functioning which are thought to be primarily
impacted irn VCI. Moreover, other recent studies have not found differences in AD and

VaD patients regarding the total amount of WML and MMSE scores (Aharon-Peretz et

al., 1998) and other measures of global cognition (Giubilei et al., 1997).
Neuropsychological Performance and MCI Group Membership

Results of the cluster analysis based on neuropsychological scores indicated
that, at least for the present sample, MCI can be described as a heterogeneous population.

Specific goups were found within the sample which generally conformed to
expectations. Specifically, one goup demonstrated deficits on tests of memory and
language, with relatively intact executive functioning, processing Speed, and
visuospatial/construction, appearing consistent with what might be expected with a
cortical etiology such as AD. A second goup emerging from the analysis (labeled
Subcortical), demonstrated the opposite pattern of results, with primary deficits revealed

on tests of executive functioning, processing speed, and visuospatial/construction.

The initial clustering of two forced clusters produced widely unbalanced groups
and suggested the inclusion of a third goup. It was initially presumed that the third

group would be indicative of subjects with a “mixed” pathology (i.e., subjects who
present with deficits commensurate with the combination of early Alzheimer’s disease
and concomitant vascular pathology); however, results indicated otherwise. Specifically,

evaluation of group means indicated that this goup performed within normal limits on all

53

neuropsychological domains, with the exception of memory. Interestingly, this goup,

labeled “Amnestic,” was the largest, with an n almost twice that of the other goups.

Given that a large percentage of individuals with MCI do not convert to dementia,

it is reasonable to posit that the identified third goup may itself be heterogeneous for a
variety of reasons, representative of subjects who present with isolated memory
impairment for numerous reasons. First, given that it is well documented that early AD is
marked by memory impairment, this goup may be comprised of individuals who are in
very early stages of early AD. Indeed, this goup evidenced the lowest volume of WML
which might be expected with a degenerative dementia such as AD. Second, although
subjects with severe depression were not included in the study, it is likely that individuals
in this goup may Show isolated memory impairment secondary to psychiatric reasons
such as mild to moderate depression or anxiety (the latter of which was not measured in
this sample). Furthermore, individuals in this goup may exhibit memory as a relative
weakness in relation to other neuropsychological functions as a function of what might be
expected in a normally distributed population (i.e., these individuals may fall on the low
end of the normal curve on memory functions). Alternatively, it may be that this third
goup possesses more cognitive reserve given significantly higher levels of education
compared to the other goups. Indeed, research has shown that higher levels of education
may be a protective factor which may delay or prevent the development of progessive
dementia (Bennett, Wilson, & Schneider et al., 2003; Cabeza, Anderson, & Locuntore,
2002; Mortimer, Snowdon, & Markesbery, 2003). Finally, the Amnestic goup presented

with significantly lower levels of depression which also may have served as a buffer

against cognnitive decline. Thus, the third goup is perhaps most interesting as it is likely

54

composed of at-risk individuals for dementia as well as those who show subtle memory

deficits secondary to normal aging or psychiatric reasons.

Discriminant analysis demonstrated that goup membership could be predicted
from neuropsychological performance and overall predictive classification was high. The
first discriminant fimction, with significant correlations with executive functioning,
Speed, and visuospatial/construction, maximally separated the Subcortical goup fi'om the
other goups, while the second discriminant function, with significant correlations with
memory, naming, and fluency discriminated the Cortical goup fiom the other goups.

Thus, the first discriminant function is consistent with hypothesized deficits which would

be expected in a population with subcortical or vascular etiology and the second
discriminant function is in line with a cortical process such as Cortical. In terms of
overall neuropsychological functioning based on severity of deficits, results indicated
that, while the Amnestic goup differed significantly fi'om both the AD and Subcortical
goups, there was no difference in severity between both the Cortical and Subcortical

goups. Thus, while considerable differences existed between the types of deficits

exhibited in these goups, overall severity did not differ.

The strongest predictor for the first discirninant function was executive
functioning, with speed and visuospatial/construction contributing equally. Conversely,
the strongest predictor for the second discriminant function was confi'ontation naming
ability which is known to be impaired early in the process of early AD. While poorer
memory correlated with this discriminant function, its overall contribution was ratlner
minnimal. This finding was somewhat unexpected given that memory has traditionally

been thought to be one of the earliest indicators of early AD. Thus, this study hi glnlights

55

the possibility that language (naming ability) and not memory may be more differentially
affected early in the process of a dementing syndrome such as AD. Indeed, a recent
study using flVIRI demonstrated that, despite adequate memory, language appears to be
affected early in normal aging (W ierenga, Perlstein, & Benjamin et al., 2005) and,
possibly, early dementia. In the current sample, education predicted naming ability more
strongly than any other neuropsychological variable and tlnus it is reasonable to assume
that the Cortical group performed more poorly than the other goups on this task given
their lower educational level. However, the Cortical goup demonstrated significantly

poorer naming ability than the other goups, even after controlling for education.

Direct comparisons between goups revealed that all goups differed significantly
with respect to executive functioning, visuospatial/ construction, and processing speed.
However, it is interesting to note that, while naming ability discriminated between the
two most impaired goups (Cortical and Subcortical), semantic fluency did not (they were
equally impaired). Recent research has shown that fluency impairment tends to be
equally impaired in botln vascular dementia and Alzheimer’s disease (Vuorinen, Laine,
and Rinne, 2000) and, while no differences appear to exist between the dementia types
with respect to semantic fluency, letter fluency appears to best discriminate between the
groups, with the VaD goup perfornning significantly worse (Canning, Leach, Stuss, Ngo,

& Black, 2004). Unfortunately, a limitation of the current study is that only semantic

fluency was measured in the sample.

Interestingly, although memory has received the most attention in terms of
diagnoses of mild cognitive impairment and dementia of the Alzheimer’s type, it only

discriminated between the Subcortical and Amnestic goups and, overall, served as the

56

weakest predictor in our analyses. While there was a trend for memory to discriminate
between the most impaired goups (Subcortical and Cortical) the relationship was not
significant. This relationship may exist for a few reasons, namely that memory testing as
part of this study was not extensive and ceiling effects may have occurred. Indeed, the
range of scores on learning and recall tasks was narrow, ranging from 0 to 10, which may
not have been large enough to detect subtle fluctuations in memory functioning in this

generally high functioning sample. Additionally, only verbal memory (and not visual

memory) was assessed which may have lead to limited findings.

Although age was strongly related to overall neuropsychological functioning, no
relationship was found for gender. Additionally, education appeared to be a protective
factor, with higher levels being associated with better overall neuropsychological
performance. It is interesting to note that level of depression was more strongly
associated with overall cognitive functioning than age or education. Overall, results
indicated that the two most impaired goups (Cortical and Subcortical) were not
significantly different in terms of age, education, or levels of depression; however,
qualitatively, the Cortical goup tended to be older, less educated, and less depressed. As
stated earlier, the Amnestic goup was younger than the other goups, with significantly

higher levels of education. Moreover, the Amnestic goup as a whole was significantly
less depressed than the other goups.

Total White Matter Lesion Volume and MCI Group Membership

Each cluster goup was evaluated in terms of total WML volume in order to
determine whether goup differences existed within the sample. As expected, the goup

with a neuropsychological profile consistent with what would be expected in vascular

57

dementia (Subcortical) evidenced significantly geater WML volume than the goup
whose neuropsychological profile appeared consistent with a cortical degenerative
etiology (Cortical). The third goup which was not irnitially hypothesized to exist within
the sample demonstrated the least total WML patlnology. Although the Cortical and

Amnestic groups did not differ statistically in terms of lesion load, the Amnestic

subgroup appeared to be the least impaired in terms of botln neuropsychological profile as
well as mean total WML pathology found on MRI.

Interestingly, all MCI subgoups evidenced some level of WML. Given the
relationship of WML with age (Papadernetriou, Narayan, & Rubins et al., 1998;
Jorgensen, Nakayama, & Raaschou et al., 1994) and common vascular risk factors such
as hypertension (Knopman et al., 2001) and atherosclerosis (Phillips & Mote-Kole,
1997), it is reasonable to assume that all goups as a whole would demonstrate WML on

MRI. However, in terms of expression of significant neuropsychological impairment

meeting criteria for MCI (and possibly later expression of dementia), a threshold may
exist whereby clinically salient cognitive deficits may be driven by the preponderance of
early WML patlnology. Indeed, some studies have shown a threshold effect where
extensive amounts of WML are necessary before cognitive impairments are seen

(DeCarli, Murphy, & Trank et al., 1995; Schmidt, Fazekas, & Koch et al., 1995; Boone,
Miller, & Lesser et al., 1992).

More recently, studies have shown WML in AD (Barber, Scheltens, & Gholkar et
al., 1999; Fazekas, Kapeller, & Schmidt et al., 1996; Scheltens, Barkhof, & Leys et al.,

1995); however, this relationship is not well understood. First, some researchers have

begun to question the prevalence of “pure” dementia syndromes and it may be that many

58

patients evaluated in this study are representative of mixed pathologies not limited solely
to vascular dementia or Alzheimer’s disease. Second, heterogeneity may exist within
AD. While research in this area is limited, a recent study (Leeuw, Barkhof, & Scheltens,
2004) argues for the identification of two syndromes within Alzheimer’s disease, one of
which evidences greater cerebrovascular disease and small vessel disease involvement.
More research in this area is clearly needed as this, to date, appears to be the only study
suggesting this relationship. Alternatively, WML in AD may be representative of a
different process or pathology. For example, WML may exert its deleterious effects on
the brain in a different manner (i.e., by primarily inducing neurodegeneration).
Alternatively, WML may interact with AD to increase the prevalence of dementia in late
life (Breteler, Bots, & Ott, 1998; DeCarli & Scheltens, 2003). Indeed, a recent study
suggests that WML may increase the likelihood of conversion to AD (Wolf et al., 2000).
Finally, there may be heterogeneity in terms of white matter changes such that other non-
vascular etiologies of dementia may lead to WML. Specifically, WML in AD may be the

consequence of amyloid deposition in cerebral vessels (Alonzo, Hyman, Rebeck, and
Greenberg, 1998).

While a relationship was expected between WML and depression based on previous
published studies, it was somewhat surprising that depression was the most significant
correlate and predictor of WML in this study (over and above the effects of age and
education). This association higlnlights the need for additional studies investigating the
relationship between vascular risk factors and depression in the context of aging (i.e.,
vascular depression hypothesis). Additionally, while a recent study demonstrated that

higher educational attainment appears to modulate the impact of WML on cognition (i.e.,

59

high level of education protects against the cognitive deterioration related to vascular
insults of the brain) (Dufouil, Alperovitch, & Tzourio, 2003), the current study
demonstrated that higher education may protect against the mere presence of WML seen
on IVLRI. This relationship is difficult to explain, however higher education is likely
associated with greater cognitive reserve as described above which may protect against
neurological insult. Additionally, higher education may be linked to lower levels of
cerebrovascular disease given environmental factors such as geater access to health care.
Periventricular and Deep White Matter Lesion Volumes and MCI Group Membership
Given that very few studies have analyzed white matter lesion types separately,
this study set out to examine the contribution of both PVL and DWML on
neuropsychological functioning in MCI. Results indicated that, while the Cortical and
Subcortical goups differed considerably in terms of DWML (with the Subcortical goup
evidencing significantly geater volume as expected), the goups did not differ

Significantly in terms of PVL. Additionally, only DWML significantly predicted overall

poorer neuropsychological functioning and this type of lesion significantly predicted

performance on tests of executive functioning, Speed of processing, and

visuospatial/constructional skills but not tests of verbal memory or language.

Overall, all goups demonstrated similar levels of PVL and, in these analyses, this

lesion type appeared to be less important in differentiating the goups. This relationship

may be explained by a relative lack of range of PVL volume scores as compared to
DWML volumes across goups. Additionally, it may that PVL is not representative of
pathological change in the brain and, instead, represents normal changes with age.

Indeed, some researchers have argued that WML located along the borders of the

60

ventricles reflect breakdown of the ependymal lining and are both more common and less
reflective of pathology than DWML (Coffey, 2000; Van Petten et al., 2004). Indeed,
PVL are commonly observed in non-demented, healthy elderly (Shinkawa, Ueda, &
Kiyohara et al., 1995; Ott, Stolk, & van Harskarnp et al., 1999). Finally, while DWML
appears to be related to ischennia, examinations comparing imaging results with
histopathological findings indicate that some PVL appear to be related to myelin pallor or
rarefaction without other convincing evidence for ischemia (Hachinski, Potter, &
Merskey et al., 1986; Hachinski, Potter, & Merskey, 1987; Yue, Arnold, Longstreth et al.,
1997). Indeed, PVL tends to correlate more strongly than DWML with advanced age and
atrophy than the Hachinski Index (Hachinski, Iliff, & Zilhka, 1975), a commonly
employed measure of vascular burden. Morever, the Hachinski Index was recently found

to be weakly related to PVL (Bigler et al., 2003).

Unlike PVL, there appears to be strong evidence that DWML is caused by
ischemia (Hachinski, Potter, & Merskey et al., 1987;Steingart, Hachinski, & Lau et al.,
1987) and correlative MRI-histopatlnological studies provide support for a relationship
between DWML and rrnicroangiopatlny. For example, extensive DWML are usually
associated with arteriolar vessel-wall changes (Hachinski et al., 1987; Steingart et al.,
1987) and the frequent observation of concomitant lesions such as lacunes is also in
support of vascular mechanisms leading to WML. Further compelling evidence comes
from the large number and extent of WML usually found in patients with known

microangiopathy such as those suffering from intracerebral hemorrhage (DeCarli, Miller,

& Swan et al., 1999).

61

Interestingly, a double dissociation was evidenced between type of WML and
neuropsychological profile (goup). Specifically, while the Cortical and Amnestic goups
demonstrated similar WML profiles (slightly higher levels of PVL versus DWML), the
Subcortical group exhibited the opposite lesion pattern (much higher level of DWML
versus PVL). To date, this appears to be the only existing study which has demonstrated
this pattern of relationship between neuropsychological profiles in MCI and their
association with WML. Since PVL appears to be related to atrophy and age, it may be
that this lesion type is associated with both aging and early AD, although additional

research is needed in this area in order to elucidate this relationship.

Given the strong association in other studies between DWML and ischemia and
other vascular risk factors (i.e., hypertension, atherosclerosis), combined with the
neuropsychological pattern of results associated with DWML in the current study, it may
be that early manifestations of vascular cognitive impairment associated with DWML
lead to deficits in neuropsychological impairments dependent on the integity of fiontal-
subcortical circuits (executive functioning, speed of processing, and
visuospatial/constructional skills). Specifically, these neuropsychological deficits may be
the result of small vessel disease disrupting frontal-subcortical pathways (Cohen et al.,
2002; Burton et al., 2003; Vataja et al., 2003 Meyer et al., 2004). More substantial global
cognitive impairment seen in later stages of dementia may be more multi-factorial and
related to a combination of specific infarcts coupled with concurrent atrophy and more
extensive small vessel disease. Recently, Stephens et a1. (2004) showed that attentional

and executive impairments occur early in the process of VaD, with memory and language

62

deficits more indicative of individuals who are more pro gessed in the process of

dementia.
Limitations and Future Directions

To our knowledge, the results presented here represent one of the only studies to
date employing cluster analysis to an MCI population and investigating the relationship
of WML to associated neuropsychological impairment. However, it is important that the

findings reported here be viewed within the context of several methodological
limitations. Specifically, in an effort to increase sensitivity to subtle cognitive
impairment, inclusion criteria may not have been appropriately strict and thus subjects
aging normally may have been included (i.e., MCI was often conceptualized as
impairment in only one neuropsychological domain). Additionally, the cross-sectional
nature of this study makes it difficult to assess the accuracy of goup membership as well
as whetlner and how goups convert to dementia. Finally, it should also be noted that
some of the neuropsychological measures used in this study (CERAD) were designed as
screening instruments in the assessment of cognitive deficits of aging and disease (Welsh
et al., 1994). As such, some of the tests used in this study represented abbreviated
versions of the original, which likely restricted the range of neuropsychological
performance. Indeed, using abbreviated tasks may result in less sensitivity in detecting

brain-behavior relationships.

There are a few confounds in terms of the neuroimaging aspect of this that are
worthy to note. For example, while this study attempted to assess heterogeneity in MCI,
no specific measure of cerebral atrophy was employed. This is important given that

atrophy is a common finding in AD. Additionally, MRI methods used in this study

63

cannot elucidate underlying cellular mecharnisms of WML. As mentioned above, some
studies have indicated that white matter abnormalities reflect a number of pathological
processes and structural MR1 prevents a reliable discrimination among such mechanisms.
Furthermore and, perhaps most importantly, the current technology does not allow
identification of specific affected white matter tracts. Better differentiation of white
matter lesions as well as clarification of the underlying pathological causes of WML
should be easier to attain with newer MR1 approaches that are becoming more available
(i.e., diffusion tensor imaging (DTI), magnetization transfer) (Kapeller, Ropele, &
Fazekas, 2000). Specifically, DTI allows for the assessment of the directionality of white
matter damage which should make clearer Specific associations of individual tracts and

bundles in the white matter (Peled, Gudbjartsson, & Westin et al., 1998).

Finally, the generalizability of the findings reported here is limited to those
individuals who are typically highly educated, independent, and relatively free of age-
associated cerebrovascular risk factors such as TIA or stroke. Therefore, this sample is
likely neither typical nor representative of the general population of older adults.
However, given that the selection criteria restricted the range of white matter

abnormalities observed in this study, these results may represent a conservative estimate

of the role of WML in MCI.

Conclusion

The concept of Mild Cognitive Impairment (MCI) has only recently been
proposed to describe the transitional state between normal cognitive functioning and
dementia. Given that tlnere is currently no reliable treatment or intervention available for

cognitive impairment that reaches the level of dementia, there has been a Shift towards

64

better identification and understanding of pre-clinical manifestations of pro gessive
cognitive impairment. Using the most reliable quantification approach (pixel-
tlnresholding) and MRI sequences (FLAIR) available for measuring and identifying WML
(Price et al., 2005), this study provided evidence for heterogeneity witlnin MCI and
related WML patlnology to identified groups witlnin the sample, demonstrating that

goups differed with respect to total WML as well as overall lesion profile. Moreover,
results of this study indicate that WML appears to be associated with specific
neuropsychological deficits dependent upon frontal-subcortical circuitry, including

executive fimctioning, processing Speed, and visuospatial/construction.

Future directions should attempt to address the possibility of the co-mingling of
vascular pathology, aging, and early AD pathology (Arvarnitakis & Hachinski, 1999) as
well as the use of newer techniques described above (i.e., diffusion tensor imaging).
Given the gowing prevalence of cognitive disorders in late life (associated with
population increases of older adults) and advances in healtln care, longitudinal studies
following older subjects (with and without vascular risk factors and associated WML)
fi'om very early stages of cognitive impairment will be important in order to further
elucidate and understand early, preclinical manifestations of cognitive impairment which

may progess to dementia.

65

.3. v a. .2. n 2 .358 0:39:00 3829.. 895 85m 0 8a m2 8.“. can 59.52 ...—m ..8 382 65:0:er 3320538832
.28. I .2 .8. 8280.53 a Ema "Somme. Bus: 823 noon n 4230 mneguo— 385:3:8 n 4>m ”mp—Eocene...“ .aommsosoxmaoaoz
a m2 ”328. .832: 823 a 42>» 35:385."— o>m88xm n mm "3335:09323835 n m> E03838 a new 3802

 

_ .8. .8. a... 2. .8. .E. .8. .3. .8. .:.. .8. a... .3..- 8. as .235:
_ .8. ..8. .8. 8. .8. .9... mm. .3”. .3... 8.. .8. 2. 3 E 80.2
_ .3. .2. .8. .8. .3. 8. .8. .8. 8.. .8. 2. 2 8 E2392
_ a. S. .8. an. 8.. an. .3... S. .8. .3 ...n 2 42.:
_ a. .2. .3. 8.. .8. .8. .. .8. a. 3 3. 423.2
_ 3.. .. a. 8.- 2.. .8. 8.- 8.- a. 3 access
_ .3. 8. .2. .:. .8. an .2. we. as ...:
_ Na. .2. .2. 2. .8. a. an as $.02:
_ 8. 8. 8. .8. 8.. ... 8.: 3.52.8
_ .2. 2. 2. 2.. 4.. ea m>.n
_ a. ...e .3. 3a ...8. e88...
_ ...e an. 8.. SN 822a
_ .. ...n «.2 same
_ 2 a? a}...
.3 .2 .N. .: .2 .e a a. e a ... .m .u ._ em 2

 

..8—nets, A23 use .Eomwofinoaaoed: .oEQEonoe mecca meomgoboeocfi new 8338... guanomon ; 03a...

66

Table 2: Group means and standard deviations on neuropsychological variables by goup

 

 

 

membership.

Clusters
Variable Subcortical Cortical Amnestic
Executive Functioning (EF) -1.27 (.45) .06 (.80) .64 (.60)
Memory (Mem) .49b (.74) -27" (.79) -.12" (1.14)
Visuospatial/Constructional (VS) -l.04 (1.04) -.25 (.72) .68 (.40)
Processing Speed (Speed) -1.13 (1.06) .04 (.43) .57 (.62)
Category Fluency (Fluency) -.44c (.94) -.72c (.68) .61 (.79)
Confrontation Naming (Naming) .41d (.65) -l .04 (1.08) .33d (.63)

Note. Within each row, means having the same letter in their superscripts are not
significantly dtficrent from each other at the .05 level.

67

Table 3: Dernogaphic data (means and standard deviations) by cluster goupings.

 

 

 

Clusters
Variable Subcortical Cortical Amnestic
Age 76.3 (7.2) 79.2 (5.8) 73.2 (7.8)
Gender 6M, 12W 9M, 9W 18M, 16W
Education 12.2 (2.9) 11.5 (1.8) 14.9 (3.7)
MMSE 25.8 (1.7) 25.9 (1.2) 26.8 (1.7)
Depression (GDS) 12.2 (6.3) 8.7 (8.2) 4.8 (4.4)

68

Table 4: Structure of the discriminant functions

 

 

 

Structure Matrix
Variable DFl DF2
.59* -.12
Executive Functioning (EF)
.49* .15
Visuospatial/Construction (VS)
.46* -.09
Processing Speed (Speed)
.00 .75*
Confiontation Naming (Naming)
.27 .48*
Verbal Fluency (Fluency)
-.12 .17*
Memory (Mern)

69

Table 5: Functions at Group Centroids (Means)

 

 

 

 

 

Groups DF 1 BF 2
Cortical/Cortical -.20 -1.74
Subcortical/Subcortical -3.26 .708
Amnestic 1 .83 .55a

 

 

 

 

Note: Within each column, means having the same letter in their superscripts are not
significantly dijfirentfi'om each other at the .05 level. Unstandardized canonical
discriminant fimctions were evaluated at group means.

70

Table 6: Means and standard deviations on WML Total Volume across clusters

 

 

Group Mean Standard Deviation
Cortical 1 1.4a 7.9
Subcortical 21 .0., b 3 .2
Amnestic 7.51, 5.8

 

Note: Means with same subscripts difler significantly from each other using the T ukey
post-hoc test @ < .001).

71

Table 7. Hierarchical regessions for PVL as a predictor of overall neuropsychological

 

 

 

functioning (NP).
Step and Predictors Statistics for Step Statistics for Predictors
R2 df AR2 Adf AF [3 t

Step 1 .459 4 — — 13.81"
Age -.34 -3.63**
Education .24 249*
Gender -.02 -.20
Depression -.40 -4.01 **

Step2 .468 5 .09 1 l 1.26"
PVL -.13 -l .02

Notes: ' p < .02 and "p < .001. Dependent variable is Overall Neuropsychological
Functioning. PVL = Periventricular lesions.

72

Table 8. Hierarchical regessions for DWML as a predictor of overall

neuropsychological functioning (NP).

 

 

 

Step and Predictors Statistics for Step Statistics for Predictors
R2 (if AR2 Adf AF [3 T
Step 1 .459 4 — — 13.81"
Age -.34 -3.64**
Education .24 2.49*
Gender -.02 -.20
Depression -.40 -4.01 **
Step2 .493 5 .034 l 12.47“
DWML -.24 -2.07*

Notes: *p < .05 and "p < .001. Dependent variable is Overall Neuropsychological
Functioning. DWML = Deep white matter lesions.

73

Table 9. Hierarchical regessions for DWML as a predictor of executive functioning

 

 

 

(EF).
Step and Predictors Statistics for Step Statistics for Predictors
R2 df AR2 Adf AF [3 T
Step 1 .388 4 — — 10.32“
Age -.25 -2.49*
Education .1 1 1 .01
Gender .01 .13
Depression -.50 -4.66**
Step2 .635 5 .247 1 22.23"
DWML -.65 -6.57**

m: ' p < .05 and "p < .001. DWML = Deep White Matter Lesions.

74

Table 10. Hierarchical regessions for DWML as a predictor of Processing Speed

 

 

 

(Speed).
Step and Predictors Statistics for Step Statistics for Predictors
R2 Df AR2 Adf AF [3 T
Step 1 .202 4 -— — 4122*
Age -. l 9 -l .63
Education .10 .85
Gender -.02 -.15
Depression -.33 -2.75*
Step2 .379 5 .177 1 7.800"
DWML -.56 -4.26**

Notes: ' p = .02 and "p < .001. DWML = Deep White Matter Lesions.

75

Table 11. Hierarchical regessions for DWML as a predictor of

Visuospatial/Construction (VS).

 

 

 

Step and Predictors Statistics for Step Statistics for Predictors
R2 Df AR2 Adf AF [3 T
Step 1 .136 4 — — 2.55"
Age -.12 -l .02
Education .07 .53
Gender -.00 -.01
Depression -.30 -2.40*
Step2 .303 5 .167 l 5.56“
DWML -.53 -3.92**

Notes: *p < .001; **p<.05. DWML = Deep white matter lesions.

76

Table 12. Hierarchical regessions for DWML as a predictor of Confrontation Naming

 

 

 

(Naming).
Step and Predictors Statistics for Step Statistics for Predictors
R2 Df AR2 Adf AF [3 T
Step 1 .126 4 —- -- 2.35
Age -.16 -l .31
Education ..23 1 .81
Gender .02 .14
Depression -.14 -1 .08
Step2 .134 5 .008 l 1.43
DWML .10 1.12

Notes: DWML = Deep white matter lesions.

 

77

Table 13. Hierarchical regessions for DWML as a predictor of Verbal Fluency

 

 

 

(Fluency).
Step and Predictors Statistics for Step Statistics for Predictors
R2 Df AR2 Adf AF B T
Step 1 .320 4 — — 7.66"
Age -.19 -1.78
Education .34 310*
Gender -.1 5 -1 .41
Depression -.22 -1.94
Step2 .324 5 .004 1 6.13**
DWML .08 .57

Notes: ' p < .02 and I""‘p < .001. DWML = Deep white matter lesions.

78

Table 14. Hierarchical regessions for DWML as a predictor of Memory (Mem).

 

 

 

Step and Predictors Statistics for Step Statistics for Predictors
R2 Df AR2 Adf AF [3 T
Step 1 .102 4 — — 1.85
Age -.29 -2.34*
Education -.01 .08
Gender .08 .60
Depression . 10 .77
Step2 .135 5 .033 1 1.54
DWML .15 .82

Notes: ' p < .05. DWML = Deep white matter lesions.

79

Figure l: Neuropsychological Profiles by Cluster Group

 

 

 

 

 

 

 

 

 

 

' - -o- - Subcortical

_'. - Cortical
+Amnasfic

l

 

80

 

Figure 2: Depression Levels as a Function of Group

 

 

     

14
12d
10 .
8 I
8 6'
Q
C
g 4:
an
o
L.
Q.
8 2.
Amnestic Cortical Subcortical
Groups

81

 

Figure 3: Biterritorial Map of Discriminant Functions

Function 2 (Cortical)

Canonical Discriminant Functions

 

 

 

 

4
o
2 r A A o 00
an o
A A A o 0 § 00
A B“ (9 3V o 0
AA 0
O r A 0 0 0
$ 0 00
A o 0 o
A o 00
0 °°o o
0
00 v Odo Ward Method
'2’ o o b
O V Group Centroids
_4 ' A Subcortical
0 Amnestic
-6 _ _ _ - _ 0 Cortical
-8 -6 —4 -2 0 2 4

Function 1 (Subcortical)

82

Figure 4: WML Volume as a Function of Group

 

22
20- /
1a-
16-
14-

12d

Mean WML Volume

10%

 

 

 

Mxedllbmal AD VAD

Groups

83

REFERENCES

Aharon-Peretz, J ., Cummings, J .L., & Hill, MA. (1988). Vascular dementia and
dementia of the Alzheimer type. Cognition, ventricular size, and leuko-ariaiosis.
Archives of Neurology, 45, 719-721.

Alexopoulos, G.S., Meyers, B.S., Yong, R.C., Campbell, S., Silbersweig, D., & Charlson,
M. (1997). ‘Vascualr depression’ hypothesis. Archives of General Psychiatry,
54, 915-922.

Almkvist, 0., Basun, H., Backrnan, L., Herlitz, A., Lannfelt, L., Small, 8., Viitanern, M.,
Walnlund, L., & Winblad, B. (1998). Mild cognitive impairment—an early stage
of Alzheimer’s disease? Journal of Neurological Transmission, 54(Suppl), 21-29.

Almkvist, O. (1994). Neuropsychological deficits in vascular dementia in relation to
Alzheimer’s disease: reviewing the evidernce for functional similarity or
divergence. Dementia, 5, 203-209.

Amar, K., Bucks, R.S.; Lewis, T., Scott, M., & Wilcock, GK. (1996). The effect
of white matter low attenuation on cognitive performance in dementia of
the Alzheimer type. Age and Ageing, 25, 443-448.

Awad, 1., Masaryk, T., & Magdinec, M. (1993). Pathogenesis of subcortical
hyperinternse lesions on magnetic resonance imaging of the brain.
Observations in patients undergoing controlled therapeutic internal carotid
artery occlusion. Stroke, 24, 1339-1346.

Ballard, C., Rowan, E., Stephens, S., Kalaria, R., Kenny, RA. (2003). Prospective
follow-up study between 3 arnd 15 months after stroke: Improvements and decline
in cognitive fnmction among dementia-free stroke survivors >75 years of age.
St_rpk§, 34, 2440-2444.

Bartres-Faz, D., Clemente, I.C., & Jnmque, C. (2001). White matter changes and
cognitive performance in aging. Review of Neurology, 33(4), 347-353.

Baum, K.A., Schulte, C., Girke, W., Reischiss, F.M., & Feliz, R. (1996).
Incidental white-matter foci on MRI in “healthy” subjects: evidence of subtle
cognitive dysfunction. Neuroradiology, 38, 755-760.

Bennett, D.A., Wilson, R.S., Schneider, J.A., Evans, D.A., Beckett, L.A., Aggarwal,

N.T. & Barnes, LL. (2002). Natural history of mild cognitive impairment in
older persons. Archives of Neurology, 59, 198-205.

Bigler, E.D., Kerr, 3., Tate, D.F., Hessel., C.D., Heath.D.E., Miller., MJ., Rice, S.A.,
& Snnith, K.H. et al. (2003). Role of white matter lesions, cerebral atrophy, and

84

APOE on cognition in older persons with and without dementia; The cache
county, Utah, Study of Memory and Aging, 17(3), 339-352.

Bowler, J .V. (1993). Cerebral infarction and To HMPAO SPECT. London:
University of London.

Bowler, J .V., Eliasziw, M., Steenhuis, R., et al. (1997). Comparative evolution of
Alzheimer’s disease, vascular dementia, and mixed dementia. Archives of
Neurology, 54, 697-703.

Bowler, J.V., Hachinksi, V., Steenhuis, R., & Lee, D. (1998). Vascular cognitive
impairment: clinical, neuropsychological and imaging findings in early
vascular dementia. Lang, 352(Sgppl 4), 63.

Bowler, J .V., Steenhuis, R., & Hachinski, V. (1999). Conceptual backgound to
vascular cognitive irnpairmennt. Alzheimer ’s Disease and Associated
Disorders, 13(3), 830-837.

Bowler, IN. (2000). Criteria for vascular dementia: replacing dogna with data.
Archives of Neurology, 57(2), 170-171.

Bowler, J .V. (2002). The concept of vascular cognitive impairment. Journal of
the Neurological Sciences, 203-204, 11-15.

Bowler, J.V. & Hachinski, V. (2003). Vascular cognitive impairment—a new
concept. In: Vascular Cognitive Impairment: Preventable Dementia. Oxford
University Press: New York

Bozoki, A.Giordani, B., Heidebrink, J.L., Berent, S., & Foster, NJ... (2001). Mild
cognitive impairments predict demerntia in nondemented elderly patients with
memory loss. Archives of Neurology, 58, 411-416.

Braak, H. & Braak, E. (1991). Neuropathological staging of Alzheimer-related
changes. Acta Neuropathologica, 82, 239—259.

Breteler, M.B., van Swieten, J.C., 8:. Bots, M.L. (1994). Cerebral white matter
lesions, vascular risk factors, and cognitive function in a population based study:
the Rotterdam study. Neurology, 44, 1246-1252.

Brown, M. (2000). White matter elmnges: Possible cerebral blood flow abnormalities.
In L. Pantoni, D. Inzitari, & A. Wallin (Eds), The matter of white matter:
Clinical and pathological aspects of white matter disease related to cognitive
decline and vascular dementia (pp. 299-312). Utrecht, the Netherlands:
Academic Pharmaceutical Publications.

85

Brun, A. (1994). Pathology and pathophysiology of cerebrovascular dementia: pure
subgoups of obstructive and hypoperfusive etiology. Dementia, 5, 145-47.

Brust, J .CM. (1988). Vascular dementia is overdiagnosed. Archives of
Neurology, 45, 799-801.

Cambell, J J. 111 & Coffey, C.E. Nennopsychiatric significance of subcortical
hyperintensity. Journal of Nempsychiatry and Clinical Neuroscience, 13(2),
261 -288.

Capdevila, A., Pujol, J., Junque, C., & Vendrell, P. (1991). Cognitive correlates of
ventricular enlarganents in vascular patients with leuko-araiosis. Acta
Neurologica Scandinavica, 84, 237-242.

Cariesimo, G.A., Sabbadini, M., Fadda, L., & Caltagirone, C. (1995). Forgetting from
long-term memory in dementia and pure amnesia: role of task, delay of
assessment and aetiology of cerebral damage. Cortex, 31, 285-300.

Carr, D.B., Gray, 8., Baty, J., & Morris, J.C. (2000). The value of informant versus
individual’s complaints of memory impairment in early dementia. Neurology, 55.
1724-1726.

Chui, H. & Gonthier, R. (1999). Natural history of vascular dennentia. Alzheimer ’s
Disease and Associated Disorders, 13(Suppl 3), 8124-8130.

Chui, H.C., Mack, W., 8: Jackson, 1.13., et al. (2000). Clinical criteria for the
diagnosis of vascular dementia: a multicenter study of comparability and
interrater reliability. Archives of Neurology, 57, 191-196.

Chui, H.C., Mack, W., Varpetian, A., & Mungas, D. (2003). Evidence-based
prognosis of vascular dementia: survival and rate of cognitive decline. In:
Vascular Cognitive [again-rent: Preventable Dementia. Oxford University
Press: New York.

Chui, H.C., Victorofl‘, 1.1., Margolin, D., Jagust, W., Shankle, R. & Katzman, R. (1992).
Criteriaforthediagnosis ofischemicvasculardementiaproposedbythe state of
California Alzheimer’s Dismse Diagnostic and Treatment Centers. Neurology,
42, 473-480.

Chui, H. (2001). Vascular dementia: a new beginning. Shifting focus from clinical
phenotype to ischemic brain injnn'y. Clinical Neurology, 18, 951-977.

Cosentino, S.A., Jefferson, A.L., Carey, M; Price, C.C., Davis-Garrett, K., Swanson, R.,
& Libon, DJ. (2004). The clinical diamosis of vascular dennentia: A

comparisonunongfomclassificationsystemsandaproposalforanew
paradigm. Clinical Neuropsychologist, 18(1), 6—21 .

86

Cullum, S., Huppert, F.A., McGee, M., Dening, T., Ahmed, A., Paykel., E.S., &
Brayne, C. (2000). Decline across different domains of cognitive function
in normal aging: results of a longitudinal population-based study using
CAMCOG. International Journal of Geriatric Psychiatry, 15, 853-862.

Cummings, J .L. (1993). Frontal-subcortical circuits and human behavior.
Archives of Neurology, 50, 873-880.

Cummings, J .L. (1994). Vascular subcortical dementias: clinical aspects.
Dementia, 5, 177-180.

Curran, T. & Schacter, D.L. (1996). Amnesia: cognitive neuropsychological
aspects. In: Behavioral Neurology and Neuropsychology (ed. T.E. Feinberg
and M.J. Farah), 463-471. New York: McGraw-Hill.

Damian, M.S., Schilling, G., Bachmann, 6., Simon, C., Stoppler, S., & Domdorf, W.
(1994). White matter lesions and cognitive deficits: relevance of lesion patterns.
Acta Neurologica Scandinavia, 90, 430-436.

Dawe, B., Procter, A., & Philpot, M. (1992). Concepts of mild memory impairment in
the elderly and their relationship to dementia: A review. International Journal of
Geriatric Psychiatry, 7(7), 473-479.

DeCarli, C., Miller, B.L., Swan, G.E., Reed, T., Wolf, P.A., & Carmelli, D. (2001).
Cerebrovascular and brain morphologic correlates of mild cognitive impairment
in the National Heart, Lung, and Blood Institute Twin Study. Archives of
Neurology, 58(4), 643-647.

De Jager, C.A., Hogervorst, E., Combrinck, M., & Budge, M.M. (2003). Sensitivity and
specificity of neuropsychological tests for mild cognitive impairment, vascular

cognitive impairment, and Alzheimer’s disease. Psychological Medicine, 33,
1039-1050.

Del Re, M.L., Pennese, F ., Ciurlinio, P., & Abate, G. (1993). Analysis of verbal
memory and learning by means of selective reminding procedure in
Alzheimer and multi-infarct dementia. Aging, 5, 185-192.

Devasenapathy, A. & Hachinski, V.C. (2000). Vascular cognitive innpairment.
Current Treatment Options in Neurology, 2(1), 61-72.

Diagnostic and statistical manual for mental disorders. (4"1 ed.). (1994). Washington,
DC: American Psychiatric Association.

Diaz, F., Hachinski, V., Merskey H., et al. ((1991). Leuko-araiosis and cognitive
impairment in Alzheimer’s disease. In: Iqbal K., McLachlan, RC,

87

Winblad B., & Wisniewski (eds.), Alzheimer ’s disease: basic mechanisms,
diagnosis and therapeutic strategies. Chichester: John Wiley.

Dick, LP, Guilofl‘, RJ., Stewart, A., Blackstock, J., Bielawska, C., Paul, E.A., 82.
Marsden, CD. (1984). Mini-Mental State Examination in neurological
patiems. Journal of Neurology, Neurosurgery, and Psychiatry, 4 7, 496- 499.

Drayer, BP. (1988). Imaging of the aging brain, 1: Normal findings. Radiology. I 66.
785-796.

Dufouil, C., Alperovitch, A., & Tzourio, C. (2003). Influence of education on the
relationship between white matter lesions and cognition. Neurology, 60(5)
831-836.

Emery, V.B., Gillie, E., & Ramdev, RT. (1994). Vascular dementia redefined. In: The
Johns Hopkins series in psychiatry and neuroscience. Oxman, T. E. & Emery,
V.B (Eds). Dementia: Presentations, dtflerential diagnosis, and nosology.
Baltimore, MD, US: Johns Hopkins University Press.

Erkinjuntti, T. (1987). Differential diagnosis between Alzheimer’s disease and
vascular dementia: evaluation of some common clinical methods. Acta
Neurologica Scandinavica, 76, 433-442.

Erkinjuntti, T., Haltia, M., Palo, J., Sulkava, R., & Paetau, A. (1988). Accuracy of the
clinical diagnosis of vascular dementia: a prospective clinical and post-modem

neuropathological study. Journal of Neurology, Neurosurgery, and Psychiatry,
51, 1 037-1044.

Erkinjuntti, T. & Hachinski, V. (1993). Rethinking vascular dementia. C erebrovasmlar
Disorders, 3, 3-23.

Erkinjuntti, T., Inzitari, D., Pantoni, L., Wallin, A., Scheltens, P., Rockwood, K.,
Roman, G.C., Chui, H., & Desmond, D.W. (2000). Research criteria for
subcortical vascular dementia in clinical trials. Journal of Neurological
T rasmission, 59, 23-30.

Erkinjuntti, T. (2000). Vascular dementia: An overview. In: O’Brien, J., Bruns, L., eds.
Dementia. 2“d ed. London, UK: Arnold Press, 623-634.

Fazekas, F., Kleinert, R., & Offenbackher, H., et al. (1993). Pathologic correlates of
incidental MR1 white matter signal hyperintensities. Neurology, 43, 1683-1689.

Fazekas, F., Schmidt, R., & Scheltens, P. (1998). Pathophysiologic mechanisms in

the development of age-related white matter changes of the brain. Dementia and
Geriatric Cognitive Disorders, 9(Suppl 1), 2—5.

88

F iducia, D.A. & DeFilippis, N.A. (2004). Mild cognitive impairment: Aging to
Alzheimer’s disease. Applied Neuropsychologv, I 1(3), 175-176.

Ferris, S.H. & Kluger, A. (1996). Commentary on age—associated memory
impairment, age—related cognitive decline and mild cognitive decline. Aging
and Neuropsychological Cognition, 3, 148-153.

Filley, C.M., Franklin, G.M., Heaton, R.K., & Rosenberg, N.L. (1988). White matter
dementia: clinical disorders and implications. Neuropsychiatry,
Neuropsychology, and Behavioral Neurology, 1(4), 23 9-254.

Fisher, N.J., Rourke, B.P., Bieliauskas, L.A., Giordani, B., Berent, S., & Foster, N.L.
(1997). Unmasking the heterogeneity of Alzheimer’s disease: case studies of
individuals fiom distinct nem'opsychological subgoups. Journal of Clinical and
Experimental Neuropsychology, 19(5), 713-754.

Folstein, M.F., Folstein, S.E., & McHugh, RR. (1975). Mini-mental state: a
practical method for grading the cognitive state of patients for the clirnician.
Journal of Psychiatry Research, 12, 189-198.

Galasko, D., Hansen, L.A., Katzman R., et a1. (1994). Clinical-neumpathological
correlalations in Alzheimer’s disease and related dementia. Archives of
Neurology, 51, 888-895.

Garrard, P., Perry, R., & Hodges, J .R. (1997). Disorders of semantic memory.
Journal of Neurology, Neurosurgery, and Psychiatry, 62, 431-435.

Garrett, K.D., Browndyke, J.N., Whelihan, W., Paul, K.H., DiCarlo, M., Moser, D.J.,
Cohen, R.A., & Ott, BR. (2004). The neuropsychological profile of vascular
cognitive impairment—no dementia: comparisons to patients at risk for
cerebrovascular disease and vascular dementia. Archives of Clinical
Neuropsychology, 19(6), 745-757.

Gold, G., Bouras, C., Canuto, A., Bergallo, M.F., Herrmann, F.R., Hof, RR, Mayor,
P.A., Michel, J .P., & Giannakopoulos, P. (2002). Clinicopathological
Validation Study of Four Sets of Clinical Criteria for Vascular Dementia.
American Journal of Psychiatry, 159, 82-87.

Grundman, M., Petersen, R., Morris J., et al. (1996). Rate of dementia of the
Alzheimer type (DAT) in subjects with mild cognitive impairment.
Neurology, 46, A403.

Gunning-Dixon, F.M. & Raz, N. (2000). The cognitive correlates of white matte

abnormalities in normal aging: A quantitative review. Neuropsychology, 14(2),
224-232.

89

Hachinski, V.C., Lassenn, N.A, & Marshall, J. (1974). Multi-infarct dementia: a cause
of mental deterioration in the elderly. Lancet, 2, 207-209.

Hachinski, V, Potter, P., & Merskey, H. (1987). Leukoaraiosis. Archives of
Neurology, 44, 21-23.

Hachinski, V.C. (1990). The decline and resurgence of vascular dementia.
Canadian Medical Association Journal, 142, 107-11 1.

Hachinski, V.C. (1991). Multi-infarct dementia: a reappraisal. Alzheimer ’s
Disease Associative Disorders, 5, 64-68.

Hachinski, V.C. (1992). Preventable senility: a call for action against the vascular
dementias. The Lancet, 340, 645-647.

Hachinski, V.C. & Bowler, J.V. (1993). Vascular dementia. Neurology, 43(20), 2159-
2160.

Hachinski, V.C. (2000). Vascular factors in cognitive impaim'nent—where are we now?
In: Annals of the New York Academy of Sciences. Ince, P. & Kalaria, R.N
(Eds). Vascular factors in Alzheimer ’s disease. New York, NY, US. New York
Academy of Sciernces.

Hachinski, V.C. (1994). Vascular dementia: a radical redefinition. Dementia, 5, 130-
132.

Haxby, J.V. (1996). Memory: medial temporal imaging. Nature, 380, 669-670.

Huppert, F.A., Brayne, C., Gill, C., Paykel., E.S., & Beardsall, L. (1995). CAMCOG—
A concise neuropsychological test to assist dementia diagnosis: socio-
dernogaphic determinants in an elderly papulation sample. British Journal of
Clinical Psychology, 34, 529-541.

Inzitari, D. (2000). Age-related white matter changes and cognitive impairment. Annals
of Neurology, 47, 141-143.

Inzitari, D., Lamassa, M. & Pantoni, L. (2003). Treatment of vascular dementias.
In: Vascular Cognitive Impairment: Preventable Dementia. Oxford
University Press: New York.

Jagust, W. (2001). Untangling vascular dementia. The Lancet, 358, 2097-2098.

J este, D.V. & McClure, RS. (1997). Psychosis: Diagnosis and treatment in the elderly.

New directions for mental health services, No. 76. In Schneider, L.S. (Ed),
Developments in geriatric psychiatry, 53- 70. San Francisco, CA, US.

96

Junque, C., Pujol, J., Vendrell, P., Brurna, O., Jodar, M., Ribas, J.C., Vinas, J.,
Capevila, A., & Marti-Vilalta, IL (1990). Leukoaraiosis on magnetic
resonance imaging and speed of mental processing. Archives of Neurology,
4 7, l 5 1-156.

Ketonen, L.M. (1998). Neuroimaging of the aging brain. Clinical Neurology, 16,
581-598.

Kilander, L., Andren, B., Nyman, H., Lind, L., Boberg, M. & Lithell, H. (1998). Atrial
fibrillation is an independent determinant of low cognitive function: a cross-
sectional study in elderly men. Stroke, 19(9), 1816-1820.

Korczyn, A.D., Romarn, G.C., & Bomstein, N.M. (2002). Vascular dementia. Journal of
the Neurological Sciences 203-204 1.

 

Krishari, K.R.R., Hays, LC, 8: Blazer, D.C. (1997). MRI-defined vascular depression.
American Journal of Psychiator, I 5 4, 497-501 .

Kurnar, A., Bilker, W., Jin, 2., & Udupa, J. (2000). Atrophy and high internsity lesions:
Complementary neurobiological mechanisms in late-life major depression.
Neuropsychopharmacoloy, 22(3), 264—274.

Lafosse, J.M., Reed, B.R., Mungas, D., Sterling, S.B., Wahbeh, H., & Jagust, WJ.
(1997). Fluency and memory differences between ischaemic vascular dementia
and Alzheimer’s disease. Neuropsychology, I], 514-522.

Larrieu, S., Letenneur, L., Orgogozo, J.M., Fabrigoule, C., Amieva, H., Le Carret, N.,
Barberger, G.P., & Dartigues, J .F . (2002). Incidence and outcome of mild
cognitive impairment in a population-based prospective cohort. Neurology,
59(10), 1594-1599.

Leuchter, A.F., Newton, T.F., Cook, I.A., Walter, S.O., Rosenberg-Thompson, S.,
& Lachenbruch, RA. (1992). Changes in brain functional connectivity in
Alzheimer-type and multi-infarct dementia. Brain, 115, 1543-1561.

Libon, D.J., Bogdanoff, B., Bonavita, J., Skalina, 8., Cloud, B.S., Resh, R., et al. (1997).
Dementia associated with periventricular deep white matter alterations: a subtype
of subcortical dementia. Archives of Clinical Neuropsychology, 12, 239-250.

Libon, D.J., Bogdanoff, B., Leopold, N., Hurka, R., Bonavita, J., Skalina, S. et al.,
(2001). Neuropsychological profiles associated with subcortical white matter
alterations and Parkinson’s disease: Implications for the diagnosis of dementia.
Archives of Clinical Neuropsychology, 16, 19-32.

91

Li, Y., Meyer, J.S., Haque, M.A., Chowdlnury, M.H., Hinh, P., & Quach, M. (2002).
Feasibility of vascular dementia treatment with cholinesterase inhibitors.
International Journal of Geriatric Psychiatry, 1 7, 194-196.

Longstreth, W.T., Manolio, T.A., & Arnold. A. Clinical correlates of white matter
findings on cranial magnetic resonance imaging of 3301 elderly people. Stroke,
27, 1274-1282.

Longstreth, W.T., Arnold, A.M., & Manolio, T.A. (2000). Clinical correlates of
ventricular and sulcal size on cranial magnetic resonance imaging of 3301 elderly
people. Neuroepidemiology, 19, 30—42.

Looi, J .C.L. & Sachdev, P.S. (2000a). Diagnosis and management of vascular dementia.
Medicine Today, I, 16-24.

Masur, D.M., Sliwinski, M., Lipton, R.B., Blau, A.D., & Crystal, HA. (1994).
Neuropsychological prediction of dennentia and the absernce of dementia in healthy
elderly persons. Neurology, 44, 1427-1432.

McKhann, G., Drachman, D., Folstein, M., et al. (1984). Clinical diagnosis of
Alzheimer’s Disease report of the NINCDS-ADRDA work goup under the
auspices of Department of Health and Human Services Task Force on
Alzheimer’s Disease. Neurology, 34, 939-944.

Mega, M.S. & Cummings, J.L. (1994). Frontal-subcortical circuits and
neuropsychiatric disorders. Journal of Neuropsychiatry and Clinical
Neurosciences, 6, 358-370.

Meguro, K., Shimada, M., Yarnagushi, S., Ishizaki, J ., Ishii, H., Shimada, Y., Sato,
M., Yamadori, A., & Sekita, Y. (2001). Cognitive function and fi'ontal lobe
atrophy in normal elderly adults: Implications for dementia not as aging-related
disorders and the reserve hypothesis. Psychiatry and Clinical
Neurosciences, 55, 565-572.

Meyer, J.S., Rogers, R.L., McClinitic, K., Mortel, K.F., & Lotfi, J. (1989).
Randomized clinical trial of daily aspirin therapy in multi—infarct dementia:
a pilot study. Journal of American Geriatric Society, 27, 549-555.

Meyer, J .S., Xu, G., Thornby, J., Chowdhury, M.G., & Quach, M. (2002). Is mild
cognitive impairment prodromal for vascular dementia like Alzheimer’s Disease?
Stroke, 33, 1981-1985.

Morris, J.C., McKee], D.W., Storandt M., et al., (1991). Very mild Alzheimer’s disease:

Informant based clinical, psychometric, and pathologic distinction from normal
aging. Neurolog, 41, 469-478.

92

Morris, J.C. (1993). The Clinical Dementia Rating (CDR): Current version and scoring
rules. Neurology, 43, 2412-2414.

Munoz, D.G., Hastak, S.M., Harper, B., Lee, D., Hachinski, V. (1993). Pathologic
correlates of increased signals of the centrum ovale on magnetic resonance
imaging. Archives of Neurology, 50, 492-497.

Naarding, P., de Koning, 1., dan Kootern, F., Dippel, D., Janzing, J ., van der Mast, R. et
al. (2003). Depression in vascular dementia. International Journal of
Geriatric Psychiatry, 18(4), 325-330.

Nagahama, Y., Okina, T., Suzuki, N., Matsuzaki, S., Yarnauchi, H., Nabatame, H.,
& Matsuda, M. (2003). Factor structure of a modified version of the
Wisconsin Card Sorting Test: an analysis of executive deficit in Alzheimer’s
disease and mild cognitive impairment. Dementia and Geriatric Cognitive
Disorders, 16(2), 103-1 12.

Nolan, K.A., Lino, M., Selignann, A.W., & Blass, JP. (1998). Absence of vascular
dementia in an autopsy series form a dementia clirnic. Journal of the American
Geriatrics Society. 46(5), 597-604.

O’Brien, J .T., Erkinjuntti, T., & Reisberg, B., et al. (2003). Vascular cognitive
impairment. Lancet Neurology, 2, 89-98.

O’Sullivan, M., Summers, P.E., Jones, D.K., Jarosz, J .M., Williams, S.C., &
Markus, HS. (2001). Normal-appearing white matter in ischemic leukoaraiosis:
a diffusion tensor MR1 study. Neurology, 5 7(12), 23-7-2310.

Pantoni, L., & Garcia, J.H. (1997). Pathogenesis of leukoaraiosis: A review. Stroke,
28, 652-659.

Petersen, R.C., Smith, G.E., Evrnik, R.J., Tangalos, E.G., Schaid, D.J., Thibodeau, S.N.,
Kokmen, E., Waring, S.C., & Kurland, LT. (1995). Apolipoprotein E status as a
predictor of the development of Alzheimer’s disease in memory-impaired
individuals. Journal of the American Medical Association, 273, 1274-1278.

Petersen, R.C., Waring, S.C., Smith, G.E., Tangalos, E.G., & Thibodeau, SN. (1996).
Predictive value of APOE genotyping in incipient Alzheimer’s disease. Annals of
New York Academy of Science, 802, 58-69.

Petersenn, R.C., Smith, G.E., Waring, S.C., Ivnik, R.J., Tangalos, E.G., & Kokrnen,

E. (1999). Mild cognitive impairment: Clinical characterization and
outcome. Archives of Neurology, 56, 303-308.

93’

Petersen, R.C., Doody, R., Kurz, A., Mohs, R.C., Morris, J.C., Rabins, P.V.,
Ritchie, K., Rossor, M., Thal, L., & Winblad, B. (2001). Current concepts in
mild cognitive impairment. Archives of Neurology, 58, 1985-1992.

Petersen, R.C., Smith, G.E., & Waring, S.C. (1997). Aging memory and mild cognitive
impairment. International Psychogeriatrics, 9(1), 65-69.

Petersen, R.C., Stevens, J.C., Ganguli, M., Tangalos, E.G., (hunmings, J.L., & DeKosky,
S.T. (2001). Practice parameter: Early detection of dementia: mild cognitive
impairment (an evidence-based review). Neurology, 56(9), 1133-1142.

Petersen, R.C. (2004). MCI as a useful clinical concept. Geriatric Times, 5(1), 1- 3.

Petersen, R.C. (2004). Mild cognitive impairment as a diagnostic mtity. Journal of
Internal Medicine, 25 6, 183-194.

Pohjasvaara, T., Mantyla, R., Ylikoski, R., Kaste, M., & Erkinjuntti, T. (2000).
Comparison of different clinical criteria (DSM-III, ADDTC, ICD-l 0,
NINDS-AIREN, DSM-IV) for the diagnosis of vascular dernnentia. National
Institute of Neurological Disorders and Stroke-Association Internationale pour la
Recherche et l’Enseignement en Neurosciences. Stroke, 31 (12), 2952-2957.

Price, C., Schmalfuss, 1.., & Sistrom, C. (2005). Quantification of white matter
alterations: A reliability analysis. Abstract presented at the International
Neuropsychological Society.

Raz, N., Rodrigue, K.M., & Acker, JD. (2003). Hypertension and the brain:
Vulnerability of the prefrontal regions and executive functions. Behavioral
Neuroscience, 11 7(6), 1 169-1 180.

Rockwood, K., Dobbs, A.R., Rule, E.G., Howlett, S.E., & Black, W.R. (1992). The
impact of pacemaker implantation on cognitive functioning in elderly
patients. Journal of the American Geriatrics Society, 40(2), 142-146.

Rockwood, K., Parhad, I, & Hachinski, V. (1994). Diagnosis of vascular dementia:
Consortium of Canadian Centres for Clinical Cognitive Research Consensus
Statement. Canadian Journal of Neurological Sciences, 21, 358-364.

Rockwood, K. (1997). Lessons fi'om mixed dementia. International
Psychogeriatrics, 9, 245-249.

Rockwood, K., Bowler, J., Erkinjuntti, T., Hachinski, V., & Wallin, A. (1999).

Subtypes of vascular dementia. Alzheimer 's Disease and Associated
Disorders, 13(Suppl. 3), $59-$65.

94'

Rockwood, K., Wentzel, C., Hachinski, V., Hogan, D.B., MacKnight, C., &
McDowell, I. (2000). Prevalence and outcomes of vascular cognitive
impairment. Neurology, 54, 447-451.

Rockwood, K., Davis, H., & MacKnight, C., et al. (2003). The Consortium to
Investigate Vascular Impairmernt of Cognition: methods and first findings.
Canadian Journal of Neurological Science, 30, 237-243.

Roman, G.C., Tatennichi, T.K., Erkinjuntti, T., Cummings, J .L., Masdeu, J .C., & Garcia,
J.H. (1993). Vascular dementia: Diagnostic criteria for research studies. Report
of the NINDS-AIREN International Workshop. Neurology, 43(2), 250-260.

Roman, GO. (2002). Vascular dementia may be the most common form of
dementia in the elderly. Journal of the Neurological Sciences, 203-204, 7- 10.

Rubin, D.H., Storandt, M., & Miller, LP. (1998). A prospective study of cognitive
fimction and onset of dementia in cognitively healthy elders. Archives of
Neurology, 55, 395-401.

Sachdev, P.S. & Looi, J.C.L. (2003). Neuropsychological differentiation of
Alzheimer’s disease and vascular dementia. Irn: Vascular Cognitive
Impairment: Preventable Dementia. Oxford University Press: New York.

Scarpelli, M., Salvolini, U., Diarnanti, L., Montironi, R., Chiaromoni, L., Maricotti,
M. (1994). MR1 and pathological examination of post-mortem brains: the
problem of white matter high signal areas. Neuroradiology, 36, 393-398.

Schmidt, R., Fazekas, F., & Kapeller, P., et al., (1998). White matter changes and
cognitive decline in non-demented subjects. In: Leys, D., Pasquier, F ., &
Scheltens, P., eds. Stroke and Alzheimer 's Disease. The Hague: Holland
Academic Graphics, 174-182.

Schwamm, L.H., Van Dyke, C., Kieman, R.J., Menin, E.L., & Mueller, J. (1987).
The Neurobehavioral Cognitive Status Examination: comparison with the
Cognitive Capacity Screening examination and the Mini-Mental State
Examination in a nem'osm-gical population. Annals of Internal Medicine, 107,
486-491.

Skoog, I. (1994). Risk factors for vascular dementia: A review. Dementia, 5 (3-4),
1 37- 1 44.

Skoog, I. (1998). Status of risk factors for vascular dementia. Neuroepidemiology,
1 7, 2-9.

95,

Skoog, I. & Gustafson, D. (2003). Vascular disorders and Alzheimer’s disease. In:
Vascular Cognitive Impairment: Preventable Dementia (eds Bowler, J.V. &
Hachinski, V.). Oxford University Press: New York.

Small, G.W., Rabins, P.V., Barry, P.P., Buckholtz, N.S., et a1. (1997). Diagnosis and
treatment of Alzheimer disease and related disorders: Consensus statement of the
American Association for Geriatric Psychiatry, the Alzheimer’s Association, and
the American Geriatrics Society. Journal of the American Medical Association,
278(16), 1363-1371.

Smith, C.D., Snowdon, D.A., Wang, H., Markesbery, W.R. (2000). White matter
volumes and periventricular white matter hyperintensities in aging and
dementia. Neurology, 54(4), 838-842.

Squire, L. R. & Shimamura, A. (1996). Nenn'opsychology of memory dysfunction
and its assessment. In: Neuropsychological Assessment of Neuropsychiatric
Disorders (ed. 1. Grant and KM. Adams), pp. 242-263. New York: Oxford
University Press.

Starr, J.M., Leaper, S.A., Murray, A.D., Lernmon, H.A., Staff, R.T., Deary, 1.1., &
Whalley, L.J. Brain white matter lesions detected by magnetic resonance
imaging are associated with balance and gait speed. Journal of Neurology,
Neurosurgery, and Psychiatry, 74, 94-98.

Starkstein, S.E., Sabe, L., & Vasquez, S. et al. (1996). Neuropsychological,
psychiatric, and cerebral blood flow findings in vascular dementia and
Alzheimer’s disease. Stroke, 7, 408-414.

Stephens, S., Kenny, R.A., Rowan, E., Allarn, L., Kalaria, R.N., Bradbury, M., &
Ballard, CG. (2004). Neuropsychological characteristics of mild vascular
cognitive impairment and dementia after stroke. International Journal of
Geriatric Psychiatry, 19, 1053-1057.

Swan, G.E., DeCarli, C., Miller, B.L., Reed, T., Wolf, P.A., & Jack, L.M. (1998).
Association of midlife blood pm to late-life cognitive decline and brain
morphology. Neurology, 51(4), 986-993.

Sze, G., DeArmond, S., Brant-Zawadski, M., Davis, R.L., Norman, D., & Newton,
TH. (1986). Foci of MRI signal (pseudo lesions) anterior to the fi'ontal horns:
Histologic correlations of a normal finding. American Journal of Neurology and
Radiology, 7, 381-387.

Takao, M., Koto, A., & Tanahashi, N. et al. (1999). Pathologic findings of silent
hyperintense white matter lesions on MRI. Journal of Neurological Science,
167, 127-131.

96

Tham, W., Auchus, A.P., Thong, M., Goh, M., Chang, H., Wong, M., & Chen, C. (2002).
Progession of cognitive impairment after stroke: One year results firm a
longitudinal study of Singaporean stroke patients. Journal of the Neurological
Sciences, 203-204, 49-52.

Tierney, M.C., Szalai, J .P., & Snow, W.G. (1996). Prediction of probable Alzheimer’s
disease in memory-impaired patients: a prospective longitudinal study.
Neurology, 46, 661-665.

Tombaugh, T.N. & McIntyre, NJ. (1992). The Mini-Mental State Examination: a
comprehensive review. Journal of the American Geriatric Society, 40, 922-
935.

Ungerleider, LG. (1995). Functional brain imaging studies of cortical systems for
memory. Science, 270, 769-772.

Villardita, C. (1996). Alzheimer’s disease compared with cerebrovascular
dementia: neuropsychological similarities and differences. Acta Neurologica
Scandinavica, 87, 299-308.

Wentzel., C., Rockwood, K., MacKnight, C., Hachinski, V., Hogan, DB, 85
Feldman, H. (2001). Progession of impairment in patiernts with vascular
cognitive impairment without dementia. Neurology, 28, 5 7(4), 714-716.

Wetterling, T., Kanitz, R.D., & Borgis, KJ. (1996). Comparison of different
diagnostic criteria for vascular dementia (ADDTC, DSM-IV, ICD-lO,
NINDS-AIREN). Stroke, 27. 30—36.

Wolf, H., Ecke, G.M., Betin, S., Dietrich, J., & Gertz, H. (2000). Do white matter
changes contribute to the subsequent development of dementia in patients with
mild cognitive impairment? A longitudinal study. International Journal of
Geriatric Psychiatry. 15, 803-812.

Ylikoski, R., Ylikoski, A., Erkinjuntti, T., Sulkava, R., Raininko, R., Tilvis, R. (1993).

White matter changes in healthy elderly persons correlate with attention and
speed of mental processing. Archives of Neurology, 50, 818-824.

97

Willi“MUM “El “Willi“