Journal of the Neurological Sciences 226 (2004) 3 – 7

www.elsevier.com/locate/jns

The neuropsychology of vascular cognitive impairment:

is there a specific cognitive deficit?
David W. Desmond*
Departments of Neurology and Pathology, SUNY Downstate Medical Center, Brooklyn, NY 11203, USA

Available online 7 October 2004

Abstract

The concept of Vascular Cognitive Impairment (VCI) encompasses patients across the entire continuum of cognitive impairment resulting
from cerebrovascular disease (CVD), ranging from high-risk patients with no frank cognitive deficit (the bbrain-at-riskQ stage) through
vascular dementia (VaD).
There are accepted differences in the neuropsychological profile of patients with Alzheimer’s disease (AD) and VaD. In patients with
VaD, executive functions that tend to be disproportionately impaired include planning and sequencing, speed of mental processing,
performance on unstructured tasks, and attention. Language production may be impaired in patients with VaD but primary language functions
otherwise tend to be preserved. Patients with VaD also exhibit significantly more perseverations than patients with AD. Memory impairment
is typically evident in patients with AD+CVD but memory impairment may also occur as a primary consequence of stroke in the posterior
cerebral artery territory with involvement of the medial temporal lobe, or as a secondary consequence of a cognitive syndrome involving
inattention due to primary executive dysfunction.
Compared to VaD, patients with AD may exhibit greater deficits in functions (including memory) mediated by posterior cortical
structures, such as the temporal and parietal lobes. AD patients exhibit a faster rate of information decay, reduced ability to benefit from cues
to facilitate retrieval, and higher frequency of intrusion errors; in addition, certain aspects of language function, such as naming, may
exacerbate deficits on verbal memory tasks. AD tends to affect lexicon while VaD tends to affect syntax. When patients with AD exhibit
perseverations, they tend to be elicited by tests of semantic knowledge.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Neuropsychology; Vascular cognitive impairment; Cognitive deficit

1. Introduction diagnosis and the features of cognitive impairment

Vascular Cognitive Impairment (VCI) [1] is a broad
concept that encompasses patients across the entire
continuum of cognitive impairment resulting from cere-
brovascular disease (CVD), ranging from high-risk
patients with no frank cognitive deficit (the bbrain-at-
riskQ stage) through severe dementia (vascular dementia,
VaD). In proposing that concept, Hachinski argued that
there is little agreement regarding the methods for the
* SUNY Downstate Medical Center, 450 Clarkson Avenue, Box 25,
Brooklyn, NY 11203, USA. Tel.: +1 718 270 1291; fax: +1 718 270 3313.
E-mail address: dwdesmond@usa.net.
0022-510X/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jns.2004.09.002
resulting from CVD. He suggested that the term
bvascularQ is too generic and fails to communicate
specific etiologies that might be amenable to treatment.
Similarly, he felt that the term bdementiaQ identifies
patients who are too severely affected to be helped.
Hachinski suggested that cognitive impairment should be
described in the language of standardized neuropsycho-
logical measures and related to the specific vascular cause
among more mildly affected patients so that treatment can
be initiated. Previous studies that have investigated
patterns of cognitive deficits among patients with CVD
have tended to focus on patients with dementia, however,
and those studies will serve as the basis for this brief
review.
4 D.W. Desmond / Journal of the Neurological Sciences 226 (2004) 3–7
2. Patterns of cognitive deficits by dementia subtype
Prior studies that have administered comprehensive
neuropsychological test batteries to patients with dementia
and identified differences in performance between patients
with different dementia subtypes (fully referenced in citations
2 and 3) have suggested that patients with VaD exhibit greater
impairment of executive functions but superior performance
in memory testing relative to patients with Alzheimer’s
disease (AD), even early in the course of both disorders. In
addition, it has been suggested that VaD is characterized by
bpatchyQ deficits, or inconsistent patterns of relative strengths
and weaknesses between patients, and that patients with VaD
more frequently exhibit fluctuations in cognitive function
than patients with AD.
In patients with VaD, the executive, or frontal lobe,
functions that tend to be disproportionately impaired include
planning and sequencing, speed of mental processing,
performance on unstructured tasks, and attention. Patients
with VaD exhibit greater deficits on measures of verbal
fluency than patients with AD, most likely due to the frontal
lobe demands of those tasks. Although the motor aspects of
language production may be impaired in patients with VaD,
primary language functions otherwise tend to be preserved. A
number of VaD studies have excluded patients with
significant aphasia, however, thus reducing their ability to
characterize language functions accurately in patients with
that dementia subtype. Patients with VaD also exhibit
significantly more perseverations than patients with AD,
particularly during tasks that assess frontal lobe functions.
Although memory impairment is typically evident in patients
with CVD when concomitant AD is present, memory
impairment may also occur as a primary consequence of
CVD, such as following a posterior cerebral artery territory
infarction involving the medial temporal lobe, or as a
secondary consequence of a cognitive syndrome involving
inattention due to primary executive dysfunction.
Patients with AD may exhibit greater deficits than
patients with VaD in functions mediated by posterior
cortical structures, such as the temporal and parietal lobes,
including memory. AD patients exhibit a faster rate of
information decay, reduced ability to benefit from cues to
facilitate retrieval, and higher frequency of intrusion errors;
in addition, certain aspects of language function, such as
naming, may exacerbate deficits on verbal memory tasks. It
has also been suggested that AD tends to affect lexicon
while VaD tends to affect syntax. When patients with AD
exhibit perseverations, they tend to be elicited by tests of
semantic knowledge.
3. Patterns of cognitive deficits by vascular dementia
subtype
Stroke patients may exhibit patterns of cognitive deficits
that vary dramatically according to characteristics such as
infarct location, number, and size. In part for this reason, it
would be reasonable for studies addressing cognitive
impairment resulting from CVD to focus on more homoge-
neous subgroups of patients, particularly such as those with
subcortical, small-vessel disease. Although the leading edge
of the cognitive syndrome of subcortical VaD typically
involves executive function due to lacunar infarctions
affecting the structures (e.g., thalamus, caudate) and
connecting pathways of frontal–subcortical circuits and
may resemble the syndrome seen in other subcortical
diseases [4], it later broadens to encompass other aspects
of cognitive function in association with multiple recurrent
subcortical strokes. Separate from these cognitive deficits,
patients with subcortical VaD may exhibit disinhibited or
unusual behavior consistent with frontal lobe dysfunction as
well as other neuropsychiatric symptoms, such as apathy or
abulia.
Single subcortical gray matter infarctions involving the
thalamus and caudate can cause cognitive impairment, but it
has also been reported that certain patients who experience
single small infarcts in the deep white matter may exhibit
significant cognitive deficits. Specifically, the syndrome of
bstrategic infarct dementiaQ was proposed as a result of the
observation of a series of six patients with infarcts involving
the inferior capsular genu that resulted in an abrupt change
in behavior [5]. The acute cognitive syndrome featured
fluctuating alertness, inattention, memory loss, apathy,
abulia, and psychomotor retardation, suggesting frontal lobe
dysfunction. Neuropsychological testing in five patients
with left-sided infarcts revealed severe verbal memory loss,
while a right-sided infarct caused transient impairment in
visuospatial memory. Additional cognitive deficits consis-
tent with dementia occurred in four patients. Functional
brain imaging in three patients showed a focal reduction in
hemispheric perfusion, most prominent in the ipsilateral
inferior and medial frontal cortex. It is likely that the
capsular genu infarcts in these patients interrupted the
inferior and anterior thalamic peduncles, causing functional
deactivation of the ipsilateral frontal cortex, supporting the
idea that lacunar infarction can cause cognitive decline
through disconnection of thalamocortical white matter
tracts.
Other subtypes of VaD have also been recognized and
they are worthy of note with regard to both their similarities
to and differences from subcortical forms of VaD that may
result from one strategically located lesion or the cumulative
effects of multiple subcortical lesions. For many years,
multi-infarct dementia (MID) [6] was the predominant
subtype of vascular dementia in research studies. As
originally conceived, multiple completed thromboembolic
infarctions were thought to cause a stereotypically stepwise
course of cognitive decline and dementia. Although some
investigators have suggested that MID is associated with
multiple lacunar infarctions, most consider MID to be the
result of cortical damage. Paradoxically, a single large
cortical infarction can result in less significant clinical
 D.W. Desmond / Journal of the Neurological Sciences 226 (2004) 3–7
consequences than a single strategically located subcortical
infarction because its effects are restricted to the region of
infarction, while a strategically located subcortical infarction
can have clinical effects remote from and disproportionate
to its location and size due to the metabolic abnormalities
that may result from pathway disruption. Thus, while some
patients with multiple infarcts may meet criteria for MID,
other patients with multiple infarcts may exhibit features of
multiple focal stroke syndromes, such as aphasia and spatial
neglect, while failing to meet operationalized criteria for
dementia. In addition, unless an MID patient has a frontal
lobe lesion, executive dysfunction should not predominate
in the clinical picture. When patients with multiple infarcts
do meet dementia criteria, it is typically due to the combined
effects of a number of focal stroke syndromes that
compromise functional competence.
It should be noted that certain single cortical infarcts can
have dramatic clinical consequences, however, such as those
in the anterior cerebral artery territory affecting the medial
frontal lobe and those in the posterior cerebral artery
territory affecting the medial temporal lobe, particularly
when they are located in the left hemisphere [7]. The former
lesion can produce executive dysfunction, and, in more
severe cases, the boneroid stateQ described by Luria [8],
while the latter lesion can produce a syndrome that is often
mistakenly termed an anomic aphasia but which actually
involves a primary disorder of memory. For that reason,
memory impairment separate from executive dysfunction
can be a predominant feature in patients with CVD for
reasons unrelated to AD due to the vascular involvement of
the same structures that AD tends to affect.
Regarding other subtypes of vascular dementia, hypo-
perfusion dementia can result from the coexistence of a
variety of medical illnesses, such as congestive heart failure,
systemic hypotension, or pneumonia [9], which can cause
the insidious onset and gradual progression of generalized
cognitive deficits, thus mimicking the stereotypic course of
AD. In addition, although dementia from bilateral internal
carotid artery (ICA) occlusions is a recognized syndrome
generally thought to result from multifocal infarction in the
borderzone territory, consistent with the concept of MID and
associated with executive dysfunction when those infarc-
tions involve the anterior borderzone territory, a less
commonly recognized mechanism for intellectual decline
from ICA occlusion is chronic ischemia due to hemody-
namic insufficiency. Certain examples in the literature
support the causal role of perfusion insufficiency from
bilateral ICA occlusions by demonstrating reversal of
intellectual deficits and improved cerebral perfusion follow-
ing surgical intervention [10].
Finally, although cerebral hemorrhage is a less common
cause of dementia than ischemic CVD, hemorrhagic
dementia warrants consideration. When hemorrhagic
dementia results from multiple lobar hemorrhages in a
setting of hypertension or amyloid angiopathy, the clinical
course and consequences associated with the lesions may be
5
similar to those seen in MID. Amyloid angiopathy
frequently exists in combination with AD, however, which
would have a significant impact on the cognitive presenta-
tion of the patient. Similarly, studies suggest that it is quite
common for CVD and AD to coexist, such as when a patient
with incipient or frank AD experiences a stroke. In most of
those cases, the clinical picture is primarily determined by
AD, not CVD.
4. Cognition and white matter lesions
Most prior studies of white matter lesions and cognitive
function have focused on the risk associated with varying
severities of those lesions, while fewer studies have
investigated the importance of the location of the lesions.
Typically, studies that have administered neuropsychological
test batteries have recognized deficits in executive function
in association with more severe white matter lesions. In one
influential study, Boone et al. [11] examined a sample of 100
subjects between the ages of 45 and 83 who were free of
neurological disease. On MRI, 46% of those subjects had no
white matter lesions, 48% had minimal or moderate white
matter lesions, and 6% had severe white matter lesions.
Those three groups received virtually identical mean Mini-
Mental State Examination total scores, all of which fell
above 29 out of a possible score of 30, and they did not differ
significantly with regard to their scores on neuropsycho-
logical measures of general intelligence, verbal and non-
verbal memory, visuospatial function, or language.
Significant deficits on measures of attention and other
executive functions were detected in the group of subjects
with severe white matter lesions, however, suggesting that a
threshold for the total area of the white matter lesions must
be surpassed before cognitive deficits occur. When the extent
of the white matter lesions is large, multiple brain regions
tend to be affected, which may be of differing importance to
cognitive function. Thus, it is likely that if a threshold exists,
it varies in association with the specific locations of the white
matter lesions, with a smaller total area of white matter lesion
being required to produce cognitive deficits when certain
critical locations are involved.
Regarding lesion location, de Groot et al. [12] performed
MRI on 1077 subjects between the ages of 60 and 90
randomly selected from the general population and admin-
istered multiple measures of psychomotor speed and
memory. They found that both periventricular and sub-
cortical white matter lesions were associated with poorer
performance on all cognitive measures, particularly those
related to psychomotor speed. They further noted that this
association was maintained for periventricular white matter
lesions while adjusting for the presence of subcortical white
matter lesions, while the association was not maintained for
subcortical white matter lesions while adjusting for the
presence of periventricular white matter lesions. Other
studies have reported a similar association between peri-
6 D.W. Desmond / Journal of the Neurological Sciences 226 (2004) 3–7
ventricular white matter lesions and processing speed,
which may be due to the high density of pathways running
through periventricular regions and interconnecting distant
cortical structures.
Although prior studies have suggested that cognitive
deficits are associated with increasing severities of white
matter lesions, it is also likely that the number of
subcortical infarctions will increase and tend to be masked
by more extensive white matter lesions and serve as a
direct cause of those cognitive deficits. In addition, it is
important to note that certain concomitant neurological
disorders, such as AD, are also associated with white
matter lesions and can significantly influence the clinical
presentation of the patient.
5. Clinical course of vascular dementia
In certain cases, the nature of the course of cognitive
decline can be informative with regard to the determi-
nation of the dementia subtype. Specifically, issues such
as whether the onset of cognitive decline is abrupt or
gradual, whether it persists, and whether its course is
stepwise or gradually progressive are relevant. Many
studies have demonstrated that the typical course of
decline in patients with AD is gradual and progressive,
but it is probably not truly linear and may vary in rate
between and within patients in association with the onset
of certain clinical features, such as extrapyramidal signs
[13]. In contrast, little effort has been made to document
the presumed stepwise or fluctuating course of decline in
VaD. Consistent with the concept of MID [6] and the
results of certain prior prospective studies of risk factors
for incident VaD, that stepwise or fluctuating course has
been thought to result from multiple recurrent strokes,
each of which may cause an acute change in the patient’s
level of cognitive function followed by a period of
stability or partial recovery.
Fischer et al. [14] performed an informative study of the
course of cognitive decline on a sample of patients with
either MID or AD. As expected, they found that 94% of
patients with AD experienced an insidious onset of their
dementia syndrome and that 81% of patients with that
dementia subtype exhibited a gradually progressive course
of decline, with 81% of patients with AD exhibiting both of
those stereotypic characteristics. Surprisingly, 54% of
patients with MID experienced the insidious onset of their
symptoms and 50% of patients with that dementia subtype
experienced a gradually progressive course of decline. Only
34% of patients with MID exhibited both of the stereotypic
characteristics of an abrupt onset of dementia followed by a
stepwise course of decline.
A critical point is that the specific cause or causes of
cognitive decline can also influence the cognitive profile
and make it more specific to or less distinctive of CVD.
Obviously, when patients are experiencing the effects of
progressive CVD, particularly when subcortical regions are
involved, we would anticipate a progression of deficits in
executive function, but when patients with CVD are also
affected by AD, any deficits in executive function are likely
to be accompanied by progressive memory impairment.
6. Future studies
In the future, studies of cognitive patterns should focus on
patients with VCI who do not meet formal dementia criteria,
perhaps specifically those with subcortical CVD in order to
maximize the homogeneity of the cohort. An appropriate
reference group would be patients with Mild Cognitive
Impairment (MCI) [15] or incipient dementia that would
most likely be due to AD, but the standard criteria for MCI
that require memory impairment should be broadened to
permit any cognitive deficit to render patients eligible.
Neuropsychological testing should be performed in con-
junction with structural and functional brain imaging in order
to identify relationships between specific cognitive deficits
and specific structural and metabolic abnormalities. Studies
should include long-term follow-up in order to assess the
comparative evolution of cognitive deficits.
Why is it better to study patients with VCI who are
non-demented rather than demented patients? First, unlike
MCI and dementia, memory impairment is not required
for the diagnosis of VCI, eliminating a source of
recognition bias. Second, less severely affected patients
will be more likely to demonstrate a subtle cognitive
deficit specific to the frontal lobes. Third, VCI patients
may be more testable than patients with dementia on
complex and challenging frontal tests, permitting deficits
to be assessed and recognized. Finally, it is likely that
non-demented patients with VCI, that is, those patients
who are at risk of progression to dementia, would be the
most appropriate candidates for clinical trials, and it will
be important to fully understand their clinical character-
istics both at baseline and during long-term follow-up.
Acknowledgments
This work was supported by Grants R01-NS26179 and
K07-AG00959 from the National Institutes of Health.
References
[1] Hachinski V. Vascular dementia: a radical redefinition. Dementia
1994;5:130 – 2.
[2] Desmond DW, Erkinjuntti T, Sano M, Cummings JL, Bowler JV,
Pasquier F, et al. The cognitive syndrome of vascular dementia:
implications for clinical trials. Alzheimer Dis Assoc Disord
1999;13(Suppl 3):S21 – 9.
[3] Looi JCL, Sachdev PS. Differentiation of vascular dementia from AD
on neuropsychological tests. Neurology 1999;53:670 – 8.
 D.W. Desmond / Journal of the Neurological Sciences 226 (2004) 3–7
[4] Cummings JL. Vascular subcortical dementias: clinical aspects.
Dementia 1994;5:177 – 80.
[5] Tatemichi TK, Desmond DW, Prohovnik I, Cross DT, Gropen TI,
Mohr JP, et al. Confusion and memory loss from capsular genu
infarction: a thalamocortical disconnection syndrome? Neurology
1992;42:1966 – 79.
[6] Hachinski VC, Lassen NA, Marshall J. Multi-infarct dementia: a
cause of mental deterioration in the elderly. Lancet 1974;2:207 – 10.
[7] Desmond DW, Moroney JT, Paik MC, Sano M, Mohr JP, Aboumatar S,
et al. Frequency and clinical determinants of dementia after ischemic
stroke. Neurology 2000;54:1124 – 31.
[8] Luria AR. The working brain: an introduction to neuropsychology.
New York7 Basic Books; 1973.
[9] Desmond DW, Moroney JT, Sano M, Stern Y. Incidence of dementia
after ischemic stroke: results of a longitudinal study. Stroke 2002;
33:2254 – 60.
[10] Tatemichi TK, Desmond DW, Prohovnik I, Eidelberg D. Dementia
associated with bilateral carotid occlusions: neuropsychological and
7
haemodynamic course after extracranial to intracranial bypass surgery.
J Neurol Neurosurg Psychiatry 1995;58:633 – 6.
[11] Boone KB, Miller BL, Lesser IM, Mehringer CM, Hill-Gutierrez E,
Goldberg MA, et al. Neuropsychological correlates of white-matter
lesions in healthy elderly subjects. A threshold effect. Arch Neurol
1992;49:549 – 54.
[12] de Groot JC, de Leeuw FE, Oudkerk M, van Gijn J, Hofman A, Jolles
J, et al. Cerebral white matter lesions and cognitive function: the
Rotterdam Scan Study. Ann Neurol 2000;47:145 – 51.
[13] Stern Y, Liu X, Albert M, Brandt J, Jacobs DM, Del Castillo-
Castaneda C, et al. Modeling the influence of extrapyramidal signs on
the progression of Alzheimer disease. Arch Neurol 1996;53:1121 – 6.
[14] Fischer P, Gatterer G, Marterer A, Simanyi M, Danielczyk W. Course
characteristics in the differentiation of dementia of the Alzheimer type
and multi-infarct dementia. Acta Psychiatr Scand 1990;81:551 – 3.
[15] Petersen RC, Doody R, Kurz A, Mohs RC, Morris JC, Rabins PV, et
al. Current concepts in mild cognitive impairment. Arch Neurol
2001;58:1985 – 92.