Review

Stroke, cognitive deficits, and rehabilitation: still an incomplete picture

Toby B. Cumming1*, Randolph S. Marshall2, and Ronald M. Lazar2
Cognitive impairment after stroke is common and can cause example, remembering a list of grocery items you have been told
disability with major impacts on quality of life and indepen- to buy is not just reliant on memory but also on attention and
dence. There are also indirect effects of cognitive impairment
language. What constitutes a cognitive domain is unclear, with the
on functional recovery after stroke through reduced participa-
tion in rehabilitation and poor adherence to treatment guide- following dimensions often incorporated: neglect (inattention to
lines. In this article, we attempt to establish the following: one side of space), agnosia (failure in object recognition), apraxia
• whether there is a distinct profile of cognitive impairment (disordered motor planning), abstract thinking (ability to use
after stroke; • whether the type of cognitive deficit can be higher-order semantic information), and arithmetic. It is impor-
associated with the features of stroke-related damage; and
tant to recognize the effects that different states of physiology and
• whether interventions can improve poststroke cognitive per- mood (e.g., fatigue, apathy, depression) can have on cognition,
formance. There is not a consistent profile of cognitive deficits
in stroke, though slowed information processing and execu- though these factors will not be reviewed here as they are modi-
tive dysfunction tend to predominate. Our understanding of fiers rather than components of cognition. The accepted ‘gold
structure–function relationships has been advanced using standard’ for determining cognitive abnormality is a battery of
imaging techniques such as lesion mapping and will
neuropsychological tests covering various domains, with norma-
be further enhanced through better characterization of
damage to functional networks and identification of subtle tive information used to indicate domain-specific deficits. Our
white matter abnormalities. Effective cognitive rehabilitation focus will be on cognitive impairment, which includes the milder
approaches have been reported for focal cortical deficits end of the cognitive dysfunction spectrum and not the more
such as neglect and aphasia, but treatments for more diffu- disabling dementia. In an Alzheimer-dominated landscape, the
sely represented cognitive impairment remain elusive. In the
term dementia has become amnestically loaded, with a dementia
future, the hope is that different techniques that have been
shown to promote neural plasticity (e.g., exercise, brain stimu- diagnosis requiring memory impairment. Yet memory dysfunc-
lation, and pharmacological agents) can be applied to improve tion is often not the most pronounced feature after stroke; only
the cognitive function of stroke survivors. about half of those with vascular cognitive impairment exhibit
Key words: cerebrovascular disease, cognitive impairment, amnestic signs (1). A more useful framework for classifying post-
neuropsychology, rehabilitation, stroke
stroke cognitive deficits emphasizes cognitive impairment rather
than dementia (2). This classification, though, still entails reduc-
Introduction ing rich information gained from multiple cognitive assessments
down to a coarse dichotomy. Ideally, we want detailed informa-
To reduce the cognitive fallout from stroke, we must be able to tion on performance across various cognitive tasks, thus allowing
characterize abnormalities in cognition, understand the underly- insight into the severity of and interrelationships among deficits
ing causes of cognitive impairment, and determine the efficacy of in different domains.
different treatment and rehabilitation approaches. In this review,
we will mostly confine ourselves to clinically apparent stroke but
will discuss some evidence from studies of asymptomatic cere-
Is there a distinct profile of cognitive impairment
brovascular disease and white matter changes.
arising from stroke?

There is a substantial literature on domain-specific cognitive

Defining cognitive impairment
function after stroke. In an influential 1994 study, Tatemichi and
colleagues identified a generalized profile of cognitive problems
Cognition is not a unitary concept; it incorporates multiple
across all domains. Attention, memory, language, and orientation
domains, including attention (focusing, shifting, dividing, or
were most affected by stroke, yet patients also had marked deficits
sustaining attention on a particular stimulus or task), executive
in visuospatial skills and abstract reasoning (3). In 1998, a large
function (planning, organizing thoughts, inhibition, control),
European study yielded a similarly diffuse poststroke cognitive
visuospatial ability (visual search, drawing, construction),
profile. Tasks of attention, visuospatial ability, and verbal fluency
memory (recall and recognition of visual and verbal informa-
were the most affected, but language and memory performance
tion), and language (expressive and receptive). Classification is
was also down (4). It is now thought that stroke tends to have
far from straightforward, as domains are not independent – for
greater deleterious impact on attention and executive function
Correspondence: Toby B. Cumming*, Florey Neuroscience Institutes,
245 Burgundy St, Heidelberg, Melbourne, Vic. 3084, Australia.
than on memory, but such a profile was not evident in these early
E-mail: [email protected] studies. A possible explanation relates to experimental design. The
1
Stroke Division, Florey Neuroscience Institutes, Melbourne, Vic., above two studies compared hospital-based stroke patients with
Australia healthy volunteer controls and thus may have overestimated post-
2
Department of Neurology, Columbia University Medical Center, stroke cognitive impairments, diminishing any domain-specific
New York, NY, USA
differences. In a community-based comparison of stroke patients
DOI: 10.1111/j.1747-4949.2012.00972.x with population controls, results fit more closely with current

38 Vol 8, January 2013, 38–45 © 2012 The Authors.

International Journal of Stroke © 2012 World Stroke Organization
 T. B. Cumming et al. Review
thinking. Patients were more frequently impaired than controls prominent. Early studies on multi-infarct or vascular dementia
in spatial ability, executive function, attention, and language but focused on differentiating these entities from Alzheimer’s disease.
were not more impaired in orientation or memory (5). Other Data typically showed that vascular dementia patients had supe-
studies have reinforced the view that stroke-related cognitive rior long-term memory but more frontal executive impairment
problems are weighted more toward attention-executive dysfunc- than Alzheimer patients (16). It is important to recognize,
tion than memory dysfunction. Marked deficits in abstraction, though, that such results do not imply spared memory in the
executive function, and processing speed have been reported vascular patients. More direct evidence for this sparing is the
(6). Population-based evidence suggests that a history of stroke finding that memory performance of patients with cerebrovascu-
is associated with a higher risk of nonamnestic [odds ratio lar disease was much closer to a healthy elderly profile than an
(OR) = 2·85] than amnestic (OR = 1·77) cognitive impairment Alzheimer profile (17). Of course, memory problems can mani-
and is associated with every cognitive domain except memory (7). fest following stroke. The five-year progression of vascular cogni-
Contributing evidence comes not only from stroke but from tive impairment has been shown to include memory deficits (18).
studies of other vascular changes. In a cohort of 164 Catholic The presence of sub-cortical infarcts in older people has been
nuns, priests, and brothers, the presence of one or more silent associated with lower episodic, semantic, and working memory
infarcts on autopsy was related most strongly with perceptual performance (19). Memory deficits, though, appear to be less
speed and least strongly with episodic memory (8). Magnetic prevalent than deficits in other cognitive domains, and when they
resonance imaging (MRI) data from elderly participants in the do occur, they are likely to have a different genesis to those seen in
LADIS study indicated that the appearance of new lacunes is Alzheimer patients. Recognition memory, which tests retention of
associated with decline in executive function and processing information without effortful search and retrieval, may be less
speed but not memory (9). affected than noncued recall after stroke (4,6), suggesting that the
underlying cause may be less amnestic and more executive.
Speed of processing
Cognitive slowing is a common complaint after stroke, and a
Are stroke characteristics associated with the
majority of patients exhibit marked slowness of information pro-
profile of cognitive deficits?
cessing (4,10). Processing speed is clinically relevant, as it makes
an independent contribution to functional outcome after stroke
Many studies into this question rely on correlative evidence, asso-
(11) and is independently predictive of dependency in stroke
ciating stroke features and cognitive status in large cohorts. Early
survivors (12). While not typically considered a stand-alone
research indicated that cognitive impairment was most frequent
domain, processing speed has a major influence on cognitive
after left anterior and posterior cerebral artery infarcts, and less
performance. Importantly, this influence is not consistent across
frequent after vertebrobasilar artery infarcts (3). Others have
different cognitive domains. It is possible that attention and
found that infarcts in the middle cerebral artery (MCA) are asso-
executive deficits appear to predominate after stroke because
ciated with greater likelihood of cognitive impairment (20).
these domains are more often tested using time-sensitive tasks
Stroke in cortical regions increases the likelihood of cognitive
(e.g., Trail-Making and verbal fluency) than the domains of
dysfunction. In the acute stage of stroke, cognitive impairments
memory or language. There is evidence to support this: when the
were detected in 74% of patients with cortical stroke but in less
block design task from the Wechsler Adult Intelligence Scale
than 50% of those with sub-cortical or infratentorial stroke (13).
(WAIS) was time limited, 47% of stroke patients performed
Stroke etiology also plays a part in cognitive outcome: cortical
beyond two standard deviations below the control mean, com-
deficits such as aphasia and neglect were more common after
pared with only 24% when there was no time limit (4). We
cardioembolic stroke than either large-vessel or small-vessel
contend that the ‘typical’ poststroke cognitive profile should
disease stroke (21).
include deficits in processing speed alongside deficits in attention
In terms of more general stroke characteristics, a meta-analysis
and executive function. The prominence of these domains
identified haemorrhagic stroke, left hemisphere involvement,
appears to be unaffected by the length of time since stroke. In
and recurrence as the factors that predicted subsequent dementia
acute stroke, disorders in executive function were found to be
(22). Haemorrhagic stroke has been independently associated
particularly prevalent (13). At one year poststroke, a majority of
with cognitive deficits across multiple domains (13). Lesions tend
patients still had attention deficits, while deficits in language and
to be larger in cognitively impaired patients than in nonimpaired
memory were more likely to have resolved (14). At five years
patients (27 vs. 9 cm3) (13), but this is neither surprising nor
poststroke, deficits were more pronounced in processing speed
particularly informative as larger strokes are more likely to
(z = -2·16) and executive function (z = -1·92) than in visual
encroach upon cortical and other regions that support cognition.
memory (z = -0·43) and verbal memory (z = -0·32) (11).
Left hemisphere bias in cognitive impairment is a common
Memory function finding, though the reliance of many neuropsychological tasks on
Identifying the extent to which memory is compromised after language ability is a confounding factor. Stroke recurrence has
stroke can be difficult. Ballard et al. (15) concluded that impair- been linked to cognitive decline and was independently associated
ments of processing speed, attention, and executive function were with incident dementia at two-year follow-up (23). In terms of
the most pronounced after stroke, but the exclusion of patients cognitive prognosis, lesion location was a useful predictor of
with dementia meant memory impairment was unlikely to be domain-specific recovery, but lesion volume was an independent

© 2012 The Authors. Vol 8, January 2013, 38–45 39

International Journal of Stroke © 2012 World Stroke Organization
 Review T. B. Cumming et al.

predictor of recovery only in visual memory (24). Other studies Though the damage is diffuse, it may be possible to identify
have shown that lesion volume, while correlated with initial some regional specificity in the relationship between white matter
aphasia severity, was not associated with recovery in language abnormalities and cognition. Slowed processing and attentional
(25,26). The impact of lesion laterality on cognitive recovery is and executive impairments have been associated with the severity
unclear, with examples of right hemispheric superiority (27), left of WMHs in the internal capsule, caudate, and thalamus of stroke
hemispheric superiority (28), and no difference (24). patients (40). Lesions in these areas disrupt fronto-striato-
thalamic circuits and thus impact upon dorsolateral prefrontal
Aphasia and neglect cortex and anterior cingulate, regions known to support attention
In clinical practice, two of the most prominent focal cognitive and executive function. Cognitive impairment has been corre-
deficits after stroke are aphasia and hemispatial neglect (29). The lated with WMHs in the frontal lobes and internal capsule,
relationship between type of aphasia or neglect and location of with hyperintense lesions in the basal ganglia and thalamus
infarction is well described. The limitation to spoken output that most closely associated with neuropsychological performance
is characteristic of Broca’s aphasia is associated with damage (6). Others have found that, while WMHs were associated with
in the left posterior, inferior frontal gyrus, in addition to other reduced mental speed, executive function, memory, and visuo-
regions supplied by the upper division of the left middle cerebral spatial ability, regional correlation was relatively weak (41). The
artery (30). Wernicke’s aphasia, on the other hand, is character- impact of small-vessel disease is not restricted to stroke: incidence
ized by fluent but relatively meaningless speech alongside poor of new lacunes in an elderly population has been linked to decline
language comprehension and is linked to damage in the left in executive function and processing speed (9).
posterior, superior temporal gyrus (30). In hemispatial neglect, Advanced imaging techniques, such as diffusion tensor imaging
the visuospatial component is linked to the right inferior parietal (DTI), can identify subtle abnormalities in axonal function that
lobule, the visuomotor component to the right dorsolateral may be a marker for generalized cognitive impairment. Calculat-
prefrontal cortex, and the object-centered component to the deep ing functional anisotropy from DTI scans provides a metric for
temporal lobe regions (31). Structure–function evidence of this the integrity of white matter structures, with lower anisotropy
kind is often derived from lesion mapping, where lesions from a reflecting greater diffusivity. Lower anisotropy has been identified
group of stroke patients with a particular cognitive deficit can be in those with vascular cognitive impairment, even in regions
superimposed to determine the area of greatest overlap. This without visible white matter abnormalities on conventional MRI
powerful technique employs group data while accounting for (42). In ischemic stroke patients, cognition appears more closely
individual variability in brain structure. The problem is that while associated with anisotropy in frontal and parietal regions than in
the mapping approach can work well for focal syndromes associ- occipital and temporal areas (43). We know that predicting cog-
ated with lesions in circumscribed areas, more complex cognitive nitive outcome using the features of a focal lesion is imperfect;
deficits such as executive dysfunction involve more widely distrib- DTI offers the promise of increased explanatory power by deter-
uted injury and more likely a broad network. Neuropsychological mining the effects of altered connectivity between brain regions.
(32) and functional imaging (33) evidence indicates that execu-
tive processes are not limited to the frontal lobes but are sub- Hypoperfusion
served by a distributed network of cortical, sub-cortical, and Compromised blood flow is relevant to both focal and diffuse
infratentorial areas. Even in the case of neglect, it has been argued deficits. In ischemic stroke, focal cognitive deficits arise not just
that dysfunction of distributed cortical networks for attention from the infarction itself but also from hypoperfusion in adjacent
provides a better account than structural damage to specific tissue. In the acute setting, aphasia and neglect are more closely
regions (34). associated with the volume of peri-infarct perfusion failure than
the infarct itself (44). White matter hyperintensities, thought to
Diffuse neuronal dysfunction represent incomplete infarction, have been attributed to transient
There is a broad distinction between focal damage, which can lead decreases in cerebral blood flow (45). At the whole-brain level, the
to selective cognitive impairments, and diffuse neuronal dysfunc- dynamic nature of information processing, attention, and working
tion, which produces a more uniform profile of mental slow- memory may be uniquely susceptible to dynamic changes in blood
ing, memory problems, and executive deficits (35). Clouding the flow across widely distributed brain regions (46).
picture is the observation that certain strategic infarcts (e.g., genu Perhaps the clearest evidence for a cerebral hemodynamic
of the internal capsule) can produce what clinically appears to be effect is in large-vessel disease, where loss of perfusion pressure
a diffuse cognitive syndrome (36). Diffuse dysfunction typically to a cerebral hemisphere supplied by a blocked carotid artery
results from underlying sub-clinical cerebrovascular disease, such induces a series of hemodynamic responses, including dilation of
as white matter disease, or an accumulation of small infarcts as in cerebral arterioles and increase of oxygen extraction fraction.
small-vessel disease (37). Over the four years following stroke, TIA patients with recently symptomatic carotid artery occlusion
higher load of white matter hyperintensities (WMHs) is strongly who had single-hemisphere cerebral hypoperfusion had greater
associated with dementia and cognitive decline (38). Stroke cognitive impairment than those without this hypoperfusion
patients with white matter lesions and silent infarcts were worse (47). Hypoperfusion may also impact cognition through reduc-
on cognitive tasks at baseline and two-year follow-up than those tions in brain volume. Gray matter volume reductions have been
without this damage (39). identified in cognitively impaired ischemic stroke patients,

40 Vol 8, January 2013, 38–45 © 2012 The Authors.

International Journal of Stroke © 2012 World Stroke Organization
 T. B. Cumming et al. Review
predominantly in the thalamus (48). With abnormalities in gray tations often prevent lengthy, battery-type assessments. The fact
matter unrelated to the site of infarction, generalized hypo- that repair mechanisms, such as resolution of possible perfusion
perfusion is a plausible mechanism. deficits (62), take their course in the days following stroke should
Cardiac research also supports the idea of hypoperfusion- not necessarily preclude early assessment. In striatocapsular
related cognitive dysfunction. Patients with global hemodynamic stroke, for example, the cortical features that are diagnostic of this
compromise arising from congestive heart failure have deficits stroke sub-type are best identified in the first few days, after which
in distributed processing skills, including attention, executive they slowly resolve and become harder to detect. Timely charac-
function, and memory (49). In older patients with heart failure, terization of cognitive impairment provides valuable prognostic
memory performance was worse in those with lower left ventricu- information (24), assists in rehabilitation planning, and allows
lar ejection fraction (<30%) (50). Again, brain volume changes therapies to be targeted to specific cognitive domains.
are potentially relevant; cognitive deficits arising from ischemic
heart disease may be a result of reduced cerebral gray matter (51). Are there rehabilitation approaches that improve
cognitive performance after stroke?
Sub-cortical stroke
Historically, damage to the cerebellum has been associated
Reducing the impact of poststroke cognitive impairment is an
with disorders of motor coordination. While cerebellar stroke is
important goal. In addition to having a direct influence on the
unlikely to cause the classic cortical signs of aphasia or neglect
quality of life of patients and their caregivers, cognitive impair-
(52), this does not mean the cerebellum is uninvolved in cognition
ment after stroke is associated with higher mortality (63), greater
(53). Groups of patients with cerebellar damage, mostly with
rates of institutionalization (64), and higher health-care costs
stroke etiology, have been shown to have deficits in visuospatial
(65). Cognition is important for recovery in other neurological
ability (54), verbal working memory (55), and across multiple
domains – patients with higher cognitive status on admission to
domains including executive function and abstract reasoning
rehabilitation had better functional outcomes, even when relevant
(56). There are anatomical grounds for expecting the cerebellum
confounders were controlled (66). Cognitive impairments can
to play a role in cognition. While the cerebellum is a structurally
reduce a person’s ability to understand task instructions, to plan
distinct brain region, it is linked by neuronal circuits within the
and initiate self-directed activities, and to solve problem. The
brain stem and has many projections toward associative brain
executive function of stroke patients in inpatient rehabilitation
areas (57). Stroke in the basal ganglia can produce cognitive dys-
has been independently associated with the level of participation
function. In a group of 12 patients with stroke confined to the
in rehabilitation (67).
basal ganglia, significant abnormalities were found in all domains
Approaches to improving cognitive outcome after stroke can be
tested (memory, attention, visuospatial, and language), though
classified as either compensatory or restorative. Compensatory
individual profiles were not presented (58). The importance of the
approaches involve adapting the external environment to altered
thalamus to cognition has been well documented. In a study of 10
cognitive abilities. Characteristics of the external environment
patients with isolated thalamic lesions, deficits in long-term
that help or hinder cognitive performance were identified in the
memory, executive function, and attention were associated with
Analysis of Cognitive Environmental Support tool (68). One spe-
specific areas of the thalamus (59). The thalamus is a good illus-
cific external strategy that has been tested is an electronic paging
tration of why lesion size is not everything – small but strategically
system. While found to be effective in compensating for everyday
placed lesions here can lead to severe cognitive impairments (60).
memory and planning problems after brain injury (69), main-
Neuropsychological assessment taining these benefits over time may be difficult for stroke
Increased knowledge of structure–function relationships can survivors (70). Compensatory strategies can also be internally
inform our approach to neuropsychological assessment. There is generated. Rather than attempt to restore reaction time to a
now sufficient experience in different aspects of cerebrovascular normal speed, Winkens et al. (71) introduced a ‘time pressure
disease that a long, comprehensive battery may not be necessary; management’ strategy: stroke patients were taught to compensate
rather, a tailored approach based on the nature of the disease (e.g., for mental slowness in real-life tasks by reorganizing the execu-
branch occlusion, carotid artery stenosis, and small-vessel tion of sub-tasks that had a time pressure component. Their
disease) can be both clinically meaningful and cost effective treatment group significantly outperformed controls on speed of
(47,61). Of greatest relevance is the neuropsychological delinea- performance on everyday tasks at three-month follow-up. Other
tion of suspected cortical signs. For example, a brief, targeted cognitive rehabilitation approaches aim for the compelling goal of
neuropsychological assessment of cortical features such as direct restoration of function. The realization that neural plastic-
neglect, spatial processing, and visuoconstructional function in a ity is present throughout life and can be influenced by training
patient with a suspected right parietal stroke not only informs has generated hope in this area, but conclusive studies remain
clinical diagnosis but also aids clinical decision making. The pres- scarce. The strongest evidence of effectiveness is for treatment of
ence of cortical signs provides information about the likely stroke focal cortical deficits. A review of evidence-based cognitive reha-
territory. This relates directly to probable stroke mechanism bilitation identified substantial evidence for cognitive-linguistic
which, in turn, affects the preferred choice of treatment. therapies to treat aphasia and for visuospatial rehabilitation
A tailored assessment of this type can be particularly beneficial in to treat neglect after stroke, but effective treatments in other
an acute stroke setting where patient fatigue and attentional limi- domains were lacking (72).

© 2012 The Authors. Vol 8, January 2013, 38–45 41

International Journal of Stroke © 2012 World Stroke Organization
 Review T. B. Cumming et al.

Domain-specific interventions decline (risk ratio = 19%), most likely by preventing the advance
Originally used in the context of upper extremity and motor of additional lacunar infarcts (85). This positive result, however,
retraining after stroke, the principles of constraint-induced was not reproduced in the PRoFESS trial, where treatment with
therapy have been applied to aphasia. Constraint-induced treat- antiplatelet and antihypertensive therapy made no difference to
ment protocols can improve functional communication in cognitive outcomes (86). Blood pressure control may also have a
chronic aphasia after stroke (73). There is also evidence that low- downside, given the association between hypoperfusion and cog-
frequency repetitive transcranial magnetic stimulation (rTMS) nitive performance. A quarter of a century ago, Meyer et al. (87)
can improve language abilities in patients with chronic nonfluent followed 52 multi-infarct dementia patients over two years.
aphasia. The theory is that the stimulation modulates and inhibits Among hypertensive patients, improved cognition was correlated
overactivity in the right hemisphere homologous language sites, with systolic blood pressure control within upper normal limits
such as the inferior frontal gyrus. Significant improvements (135–150 mmHg) and cognition worsened if blood pressure was
following rTMS have been reported in the picture naming reduced below this level. As more direct evidence of this link, a
(74), expressive language and auditory comprehension (75), and pilot study showed that cortical reperfusion with temporary
semantic fluency (76) of stroke patients with chronic aphasia. blood pressure increase for patients with large-vessel stenosis in
In stroke patients with hemispatial neglect, using prism glasses the acute stage resulted in improved cognition, both in the short
to create an optical shift of the visual field to the right has been and longer terms (88).
successful. The original 1998 study included patients in the first Pharmacological agents can influence cognition. Normally
year after stroke and revealed improvements in sensorimotor used to treat depression, escitalopram has been found to benefit
and spatial function after a single prism adaptation (77). More cognitive function in stroke patients (89). Compared with
recent work has shown that prism adaptation also works in patients who received placebo or problem solving therapy, those
chronic stroke. In patients who were one to seven years post- receiving escitalopram had improved global cognition on the
stroke and had persistent neglect, an eight-week intervention RBANS (Repeatable Battery for the Assessment of Neuropsycho-
with prism glasses produced improved eye movements on the logical Status), particularly in memory. Rivastigmine, typically
neglected side and improved standing center of gravity (78). used in Alzheimer’s disease, was tested in a randomized con-
Prism adaptation may actually be best suited to the sub-acute trolled trial among stroke patients with cognitive impairment but
and chronic stages of stroke when the nature of the neglect not dementia (90). The treatment group exhibited significantly
deficit has been established. Prism adaptation in acute stroke improved performance on verbal fluency for animals relative to
patients yielded some benefits on spatial tasks (line bisection and controls. The results of these two studies and the previously cited
cancellation), but these benefits were not maintained at one study on attention process training (84), however, need to be
month posttreatment (79). It is likely that much of the improve- interpreted with caution until successfully replicated. All three
ment in both treatment and placebo control groups was due to studies featured improvement in one or two cognitive measures in
spontaneous recovery. the context of multiple other cognitive outcomes that were not
Cochrane reviews have revealed the paucity of studies in post- affected, and in each case there was an imbalance in baseline
stroke cognitive rehabilitation for both memory deficits (80) and performance between the groups.
attention deficits (81). Unlike the example of constraint-based Increasing physical activity is another potential treatment; it
therapy, the principles of task-specific training do not seem to has been shown to improve cognitive performance in the cogni-
translate from motor rehabilitation to cognitive rehabilitation. tively impaired (91), those at risk of Alzheimer’s disease (92), and
Repetition and rehearsal of the targeted cognitive task is insuffi- patients with multiple sclerosis (93). Benefits seem to be greatest
cient to produce meaningful improvement. Results from one in cognitive speed and attention (94) and executive control (95).
small stroke study (n = 12) indicated that teaching mnemonic This is important as these central cognitive processes provide the
strategies resulted in better memory performance, but improve- foundation for many other aspects of cognition. The relationship
ments did not generalize beyond the trained memory tasks (82). between physical and mental activity may also apply to stroke
In a small trial (n = 27), a computerized training program had patients, but there is scant empirical evidence to date. In a recent
significant effects on alertness and sustained attention, but the review, only 12 studies were identified that employed physical
benefits did not generalize to other attentional and cognitive activity interventions in stroke and had cognitive outcomes (96).
functions (83). In the years since the Cochrane reviews, one of the One of the few that had cognition as a primary focus found stroke
most promising studies was a randomized controlled trial of patients exposed to an eight-week exercise program had improved
‘attention process training’ in 78 stroke patients with attention information processing speed on a serial reaction time task
deficits (84). Postintervention change scores on the Integrated relative to control patients (97). Combining increased physical
Visual Auditory Continuous Performance Test were superior activity with mental challenges, sensory stimulation, and social
in the treatment group, and this superiority was maintained at interaction in an ‘enriched environment’ may contribute to
six-month follow-up. improvements in cognition after stroke. Most of the current evi-
dence for this comes from animals (98), but there are data from
Interventions for generalized cognitive impairment humans showing that listening to music early after stroke can
Hypertension is an obvious treatment target. In the PROGRESS enhance cognitive recovery (99). Epidemiological studies have
trial, active blood pressure lowering reduced the risk of cognitive demonstrated the cognitive value of interactions with other

42 Vol 8, January 2013, 38–45 © 2012 The Authors.

International Journal of Stroke © 2012 World Stroke Organization
 T. B. Cumming et al. Review
people, using markers such as marital status and frequency of 8 Schneider JA, Wilson RS, Cochran EJ et al. Relation of cerebral
social activities (100,101). infarctions to dementia and cognitive function in older persons.
Neurology 2003; 60:1082–8.
9 Jokinen H, Gouw AA, Madureira S et al. Incident lacunes influence
Conclusion cognitive decline. Neurology 2011; 76:1872–8.
10 Rasquin S, Lodder J, Verhey F. The association between psychiatric
General patterns can be detected in the profile of cognitive and cognitive symptoms after stroke: a prospective study. Cerebrovasc
impairment after stroke: focal disorders such as aphasia and Dis 2005; 19:309–16.
11 Barker-Collo S, Feigin VL, Parag V, Lawes CM, Senior H. Auckland
neglect are common, as are more diffuse abnormalities such as
Stroke Outcomes Study. Part 2: cognition and functional outcomes 5
slowed information processing and executive dysfunction. Lesion years poststroke. Neurology 2010; 75:1608–16.
mapping studies have been helpful in delineating the brain areas 12 Narasimhalu K, Ang S, De Silva DA et al. The prognostic effects of
that contribute to domain-specific cognitive processes, and new poststroke cognitive impairment no dementia and domain-specific
research into how structural damage affects functional networks cognitive impairments in nondisabled ischemic stroke patients.
Stroke 2011; 42:883–8.
will further increase our understanding. Determining the causes
13 Nys GM, van Zandvoort MJ, de Kort PL, Jansen BP, de Haan EH,
of diffusely represented cognitive deficits is difficult but may well Kappelle LJ. Cognitive disorders in acute stroke: prevalence and clini-
be related to sub-clinical vascular changes and subtle abnormali- cal determinants. Cerebrovasc Dis 2007; 23:408–16.
ties in white matter; such factors are starting to be revealed by 14 Lesniak M, Bak T, Czepiel W, Seniow J, Czlonkowska A. Frequency
new imaging techniques. Cognitive rehabilitation approaches and prognostic value of cognitive disorders in stroke patients.
Dement Geriatr Cogn Disord 2008; 26:356–63.
have been relatively successful for focal cortical deficits but less so
15 Ballard C, Stephens S, Kenny R, Kalaria R, Tovee M, O’Brien J. Profile
for more generalized cognitive impairment. Repetitive transcra- of neuropsychological deficits in older stroke survivors without
nial magnetic stimulation and prism adaptation are two very dementia. Dement Geriatr Cogn Disord 2003; 16:52–6.
different techniques to manipulate neural activity and drive plas- 16 Looi JC, Sachdev PS. Differentiation of vascular dementia from AD
ticity. Further studies are needed to test other ways of stimulating on neuropsychological tests. Neurology 1999; 53:670–8.
17 Reed BR, Mungas DM, Kramer JH et al. Profiles of neuropsycho-
functional brain changes after stroke to improve cognition.
logical impairment in autopsy-defined Alzheimer’s disease and
cerebrovascular disease. Brain 2007; 130:731–9.
Acknowledgements 18 Ingles JL, Wentzel C, Fisk JD, Rockwood K. Neuropsycho-
logical predictors of incident dementia in patients with vascular
cognitive impairment, without dementia. Stroke 2002; 33:1999–
We thank Dr Jennifer Bradshaw for her valuable contribution to
2002.
the section on neuropsychological assessment. Dr Cumming is 19 Schneider JA, Boyle PA, Arvanitakis Z, Bienias JL, Bennett DA.
funded by a National Heart Foundation Postdoctoral Research Subcortical infarcts, Alzheimer’s disease pathology, and memory
Fellowship. The Florey Neuroscience Institutes acknowledges function in older persons. Ann Neurol 2007; 62:59–66.
strong support from the Victorian Government and in particular 20 Jaillard A, Grand S, Le Bas JF, Hommel M. Predicting cognitive
dysfunctioning in nondemented patients early after stroke.
the funding from the Operational Infrastructure Support Grant.
Cerebrovasc Dis 2010; 29:415–23.
21 Hoffmann M. Higher cortical function deficits after stroke: an analy-
References sis of 1,000 patients from a dedicated cognitive stroke registry.
1 Feldman H, Levy AR, Hsiung GY et al. A Canadian cohort study Neurorehabil Neural Repair 2001; 15:113–27.
of cognitive impairment and related dementias (ACCORD): study 22 Pendlebury ST, Rothwell PM. Prevalence, incidence, and factors asso-
methods and baseline results. Neuroepidemiology 2003; 22:265– ciated with pre-stroke and post-stroke dementia: a systematic review
74. and meta-analysis. Lancet Neurol 2009; 8:1006–18.
2 Hachinski V, Iadecola C, Petersen RC et al. National Institute of 23 Srikanth VK, Quinn SJ, Donnan GA, Saling MM, Thrift AG.
Neurological Disorders and Stroke-Canadian Stroke Network vascu- Long-term cognitive transitions, rates of cognitive change, and pre-
lar cognitive impairment harmonization standards. Stroke 2006; dictors of incident dementia in a population-based first-ever stroke
37:2220–41. cohort. Stroke 2006; 37:2479–83.
3 Tatemichi TK, Desmond DW, Stern Y, Paik M, Sano M, Bagiella E. 24 Nys GM, Van Zandvoort MJ, De Kort PL et al. Domain-specific cog-
Cognitive impairment after stroke: frequency, patterns, and relation- nitive recovery after first-ever stroke: a follow-up study of 111 cases.
ship to functional abilities. J Neurol Neurosurg Psychiatry 1994; J Int Neuropsychol Soc 2005; 11:795–806.
57:202–7. 25 Lazar RM, Minzer B, Antoniello D, Festa JR, Krakauer JW, Marshall
4 Hochstenbach J, Mulder T, van Limbeek J, Donders R, Schoonder- RS. Improvement in aphasia scores after stroke is well predicted by
waldt H. Cognitive decline following stroke: a comprehensive study initial severity. Stroke 2010; 41:1485–8.
of cognitive decline following stroke. J Clin Exp Neuropsychol 1998; 26 Lazar RM, Antoniello D. Variability in recovery from aphasia. Curr
20:503–17. Neurol Neurosci Rep 2008; 8:497–502.
5 Srikanth VK, Thrift AG, Saling MM et al. Increased risk of cognitive 27 Hochstenbach JB, den Otter R, Mulder TW. Cognitive recovery
impairment 3 months after mild to moderate first-ever stroke: a after stroke: a 2-year follow-up. Arch Phys Med Rehabil 2003; 84:
community-based prospective study of nonaphasic English-speaking 1499–504.
survivors. Stroke 2003; 34:1136–43. 28 Desmond DW, Moroney JT, Sano M, Stern Y. Recovery of cognitive
6 Sachdev PS, Brodaty H, Valenzuela MJ et al. The neuropsychological function after stroke. Stroke 1996; 27:1798–803.
profile of vascular cognitive impairment in stroke and TIA patients. 29 Gottesman RF, Hillis AE. Predictors and assessment of cognitive
Neurology 2004; 62:912–9. dysfunction resulting from ischaemic stroke. Lancet Neurol 2010;
7 Knopman DS, Roberts RO, Geda YE et al. Association of prior stroke 9:895–905.
with cognitive function and cognitive impairment. Arch Neurol 2009; 30 Hillis AE. Aphasia: progress in the last quarter of a century. Neurology
66:614–9. 2007; 69:200–13.

© 2012 The Authors. Vol 8, January 2013, 38–45 43

International Journal of Stroke © 2012 World Stroke Organization
 Review T. B. Cumming et al.

31 Verdon V, Schwartz S, Lovblad KO, Hauert CA, Vuilleumier P. 52 Frank B, Maschke M, Groetschel H et al. Aphasia and neglect are
Neuroanatomy of hemispatial neglect and its functional compo- uncommon in cerebellar disease: negative findings in a prospective
nents: a study using voxel-based lesion-symptom mapping. Brain study in acute cerebellar stroke. Cerebellum 2010; 9:556–66.
2010; 133(Pt 3):880–94. 53 O’Halloran CJ, Kinsella GJ, Storey E. The cerebellum and neuro-
32 Vataja R, Pohjasvaara T, Mantyla R et al. MRI correlates of executive psychological functioning: a critical review. J Clin Exp Neuropsychol
dysfunction in patients with ischaemic stroke. Eur J Neurol 2003; 2012; 34:35–56.
10:625–31. 54 Molinari M, Petrosini L, Misciagna S, Leggio MG. Visuospatial abili-
33 Fassbender C, Murphy K, Foxe JJ et al. A topography of executive ties in cerebellar disorders. J Neurol Neurosurg Psychiatry 2004;
functions and their interactions revealed by functional magnetic 75:235–40.
resonance imaging. Brain Res Cogn Brain Res 2004; 20:132–43. 55 Ravizza SM, McCormick CA, Schlerf JE, Justus T, Ivry RB, Fiez JA.
34 Corbetta M, Shulman GL. Spatial neglect and attention networks. Cerebellar damage produces selective deficits in verbal working
Annu Rev Neurosci 2011; 34:569–99. memory. Brain 2006; 129(Pt 2):306–20.
35 de Haan EH, Nys GM, Van Zandvoort MJ. Cognitive function fol- 56 Schmahmann JD, Sherman JC. The cerebellar cognitive affective syn-
lowing stroke and vascular cognitive impairment. Curr Opin Neurol drome. Brain 1998; 121(Pt 4):561–79.
2006; 19:559–64. 57 Lagarde J, Hantkie O, Hajjioui A, Yelnik A. Neuropsychological dis-
36 Tatemichi TK, Desmond DW, Prohovnik I. Strategic infarcts in vas- orders induced by cerebellar damage. Ann Phys Rehabil Med 2009;
cular dementia. A clinical and brain imaging experience. Arzneimit- 52:360–70.
telforschung 1995; 45:371–85. 58 Hochstenbach J, van Spaendonck KP, Cools AR, Horstink MW,
37 Pantoni L. Cerebral small vessel disease: from pathogenesis and Mulder T. Cognitive deficits following stroke in the basal ganglia.
clinical characteristics to therapeutic challenges. Lancet Neurol 2010; Clin Rehabil 1998; 12:514–20.
9:689–701. 59 Van der Werf YD, Scheltens P, Lindeboom J, Witter MP, Uylings
38 Dufouil C, Godin O, Chalmers J et al. Severe cerebral white matter HBM, Jolles J. Deficits of memory, executive functioning and atten-
hyperintensities predict severe cognitive decline in patients with tion following infarction in the thalamus: a study of 22 cases with
cerebrovascular disease history. Stroke 2009; 40:2219–21. localised lesions. Neuropsychologia 2003; 41:1330–44.
39 Rasquin SM, Verhey FR, Lousberg R, Lodder J. Cognitive perfor- 60 Szirmai I, Vastagh I, Szombathelyi E, Kamondi A. Strategic infarcts
mance after first ever stroke related to progression of vascular brain of the thalamus in vascular dementia. J Neurol Sci 2002; 204:91–7.
damage: a 2 year follow up CT scan study. J Neurol Neurosurg 61 Williamson JD, Miller ME, Bryan RN et al. The Action to Control
Psychiatry 2005; 76:1075–9. Cardiovascular Risk in Diabetes Memory in Diabetes Study
40 Burton E, Ballard C, Stephens S et al. Hyperintensities and (ACCORD-MIND): rationale, design, and methods. Am J Cardiol
fronto-subcortical atrophy on MRI are substrates of mild cogni- 2007; 99:112i–22i.
tive deficits after stroke. Dement Geriatr Cogn Disord 2003; 16: 62 Hillis AE, Barker PB, Beauchamp NJ, Gordon B, Wityk RJ. MR per-
113–8. fusion imaging reveals regions of hypoperfusion associated with
41 Jokinen H, Kalska H, Mantyla R et al. White matter hyperintensities aphasia and neglect. Neurology 2000; 55:782–8.
as a predictor of neuropsychological deficits post-stroke. J Neurol 63 Tatemichi TK, Paik M, Bagiella E, Desmond DW, Pirro M, Hanzawa
Neurosurg Psychiatry 2005; 76:1229–33. LK. Dementia after stroke is a predictor of long-term survival. Stroke
42 Medina D, DeToledo-Morrell L, Urresta F et al. White matter changes 1994; 25:1915–9.
in mild cognitive impairment and AD: a diffusion tensor imaging 64 Pasquini M, Leys D, Rousseaux M, Pasquier F, Henon H. Influence of
study. Neurobiol Aging 2006; 27:663–72. cognitive impairment on the institutionalisation rate 3 years after a
43 Williamson J, Nyenhuis D, Stebbins GT et al. Regional differences in stroke. J Neurol Neurosurg Psychiatry 2007; 78:56–9.
relationships between apparent white matter integrity, cognition and 65 Claesson L, Linden T, Skoog I, Blomstrand C. Cognitive impairment
mood in patients with ischemic stroke. J Clin Exp Neuropsychol 2010; after stroke – impact on activities of daily living and costs of care for
32:673–81. elderly people. The Goteborg 70+ Stroke Study. Cerebrovasc Dis 2005;
44 Hillis AE, Wityk RJ, Barker PB et al. Subcortical aphasia and neglect 19:102–9.
in acute stroke: the role of cortical hypoperfusion. Brain 2002; 125 66 Heruti RJ, Lusky A, Dankner R et al. Rehabilitation outcome of
(Pt 5):1094–104. elderly patients after a first stroke: effect of cognitive status at admis-
45 Pantoni L, Garcia JH. Pathogenesis of leukoaraiosis: a review. Stroke sion on the functional outcome. Arch Phys Med Rehabil 2002; 83:
1997; 28:652–9. 742–9.
46 Marshall RS, Lazar RM. Pumps, aqueducts, and drought manage- 67 Skidmore ER, Whyte EM, Holm MB et al. Cognitive and affective
ment: vascular physiology in vascular cognitive impairment. Stroke predictors of rehabilitation participation after stroke. Arch Phys Med
2011; 42:221–6. Rehabil 2010; 91:203–7.
47 Marshall RS, Festa JR, Cheung YK et al. Cerebral hemodynamics and 68 Ryan JD, Polatajko HJ, McEwen S et al. Analysis of Cognitive
cognitive impairment: baseline data from the RECON trial. Neurol- Environmental Support (ACES): preliminary testing. Neuropsychol
ogy 2012; 78:250–5. Rehabil 2011; 21:401–27.
48 Stebbins GT, Nyenhuis DL, Wang C et al. Gray matter atrophy in 69 Wilson BA, Emslie HC, Quirk K, Evans JJ. Reducing everyday
patients with ischemic stroke with cognitive impairment. Stroke memory and planning problems by means of a paging system: a
2008; 39:785–93. randomised control crossover study. J Neurol Neurosurg Psychiatry
49 Vogels RL, Oosterman JM, Laman DM et al. Transcranial Doppler 2001; 70:477–82.
blood flow assessment in patients with mild heart failure: correlates 70 Fish J, Manly T, Emslie H, Evans JJ, Wilson BA. Compensatory strat-
with neuroimaging and cognitive performance. Congest Heart Fail egies for acquired disorders of memory and planning: differential
2008; 14:61–5. effects of a paging system for patients with brain injury of traumatic
50 Festa JR, Jia X, Cheung K et al. Association of low ejection fraction versus cerebrovascular aetiology. J Neurol Neurosurg Psychiatry 2008;
with impaired verbal memory in older patients with heart failure. 79:930–5.
Arch Neurol 2011; 68:1021–6. 71 Winkens I, Van Heugten CM, Wade DT, Habets EJ, Fasotti L. Efficacy
51 Almeida OP, Garrido GJ, Beer C, Lautenschlager NT, Arnolda L, of time pressure management in stroke patients with slowed infor-
Flicker L. Cognitive and brain changes associated with ischaemic mation processing: a randomized controlled trial. Arch Phys Med
heart disease and heart failure. Eur Heart J 2012; 33:1769–76. Rehabil 2009; 90:1672–9.

44 Vol 8, January 2013, 38–45 © 2012 The Authors.

International Journal of Stroke © 2012 World Stroke Organization
 T. B. Cumming et al. Review
72 Cicerone KD, Dahlberg C, Malec JF et al. Evidence-based cognitive Second Strokes (PRoFESS) trial: a double-blind, active and placebo-
rehabilitation: updated review of the literature from 1998 through controlled study. Lancet Neurol 2008; 7:875–84.
2002. Arch Phys Med Rehabil 2005; 86:1681–92. 87 Meyer JS, Judd BW, Tawaklna T, Rogers RL, Mortel KF. Improved
73 Meinzer M, Rodriguez AD, Gonzalez Rothi LJ. First decade of cognition after control of risk factors for multi-infarct dementia.
research on constrained-induced treatment approaches for aphasia JAMA 1986; 256:2203–9.
rehabilitation. Arch Phys Med Rehabil 2012; 93:S35–45. 88 Hillis AE, Ulatowski JA, Barker PB et al. A pilot randomized trial of
74 Naeser MA, Martin PI, Nicholas M et al. Improved picture naming induced blood pressure elevation: effects on function and focal per-
in chronic aphasia after TMS to part of right Broca’s area: an open- fusion in acute and subacute stroke. Cerebrovasc Dis 2003; 16:236–46.
protocol study. Brain Lang 2005; 93:95–105. 89 Jorge RE, Acion L, Moser D, Adams HP Jr, Robinson RG. Escitalo-
75 Barwood CH, Murdoch BE, Whelan BM et al. Improved language pram and enhancement of cognitive recovery following stroke. Arch
performance subsequent to low-frequency rTMS in patients with Gen Psychiatry 2010; 67:187–96.
chronic non-fluent aphasia post-stroke. Eur J Neurol 2011; 18: 90 Narasimhalu K, Effendy S, Sim CH et al. A randomized controlled
935–43. trial of rivastigmine in patients with cognitive impairment no
76 Szaflarski JP, Vannest J, Wu SW, DiFrancesco MW, Banks C, Gilbert dementia because of cerebrovascular disease. Acta Neurol Scand 2010;
DL. Excitatory repetitive transcranial magnetic stimulation induces 121:217–24.
improvements in chronic post-stroke aphasia. Med Sci Monit 2011; 91 Heyn P, Abreu BC, Ottenbacher KJ. The effects of exercise training
17:CR132–9. on elderly persons with cognitive impairment and dementia: a meta-
77 Rossetti Y, Rode G, Pisella L et al. Prism adaptation to a rightward analysis. Arch Phys Med Rehabil 2004; 85:1694–704.
optical deviation rehabilitates left hemispatial neglect. Nature 1998; 92 Lautenschlager NT, Cox KL, Flicker L et al. Effect of physical activity
395:166–9. on cognitive function in older adults at risk for Alzheimer disease:
78 Shiraishi H, Yamakawa Y, Itou A, Muraki T, Asada T. Long-term a randomized trial. JAMA 2008; 300:1027–37.
effects of prism adaptation on chronic neglect after stroke. Neuro- 93 Prakash RS, Snook EM, Erickson KI et al. Cardiorespiratory fitness: a
Rehabilitation 2008; 23:137–51. predictor of cortical plasticity in multiple sclerosis. Neuroimage 2007;
79 Nys GM, de Haan EH, Kunneman A, de Kort PL, Dijkerman HC. 34:1238–44.
Acute neglect rehabilitation using repetitive prism adaptation: a 94 Angevaren M, Aufdemkampe G, Verhaar HJ, Aleman A, Vanhees L.
randomized placebo-controlled trial. Restor Neurol Neurosci 2008; Physical activity and enhanced fitness to improve cognitive function
26:1–12. in older people without known cognitive impairment. Cochrane
80 das Nair RD, Lincoln NB. Cognitive rehabilitation for memory Database Syst Rev 2008; (2):CD005381.
deficits following stroke. Cochrane Database Syst Rev 2007; (3): 95 Colcombe S, Kramer AF. Fitness effects on the cognitive function of
CD002293. older adults: a meta-analytic study. Psychol Sci 2003; 14:125–30.
81 Lincoln NB, Majid MJ, Weyman N. Cognitive rehabilitation for 96 Cumming TB, Tyedin K, Churilov L, Morris ME, Bernhardt J. The
attention deficits following stroke. Cochrane Database Syst Rev 2000; effect of physical activity on cognitive function after stroke: a system-
(4):CD002842. atic review. Int Psychogeriatr 2012; 24:557–67.
82 Doornhein K, De Haan EHF. Cognitive training for memory deficits 97 Quaney BM, Boyd LA, McDowd JM et al. Aerobic exercise improves
in stroke patients. Neuropsychol Rehabil 1998; 8:393–400. cognition and motor function poststroke. Neurorehabil Neural Repair
83 Sturm W, Willmes K. Efficacy of a reaction training on various atten- 2009; 23:879–85.
tional and cognitive functions in stroke patients. Neuropsychol 98 Janssen H, Bernhardt J, Collier JM et al. An enriched environment
Rehabil 1991; 1:259–80. improves sensorimotor function post-ischemic stroke. Neurorehabil
84 Barker-Collo SL, Feigin VL, Lawes CM, Parag V, Senior H, Rodgers A. Neural Repair 2010; 24:802–13.
Reducing attention deficits after stroke using attention process train- 99 Sarkamo T, Tervaniemi M, Laitinen S et al. Music listening enhances
ing: a randomized controlled trial. Stroke 2009; 40:3293–8. cognitive recovery and mood after middle cerebral artery stroke.
85 Tzourio C, Anderson C, Chapman N et al. Effects of blood pressure Brain 2008; 131(Pt 3):866–76.
lowering with perindopril and indapamide therapy on dementia and 100 Glei DA, Landau DA, Goldman N, Chuang YL, Rodriguez G,
cognitive decline in patients with cerebrovascular disease. Arch Intern Weinstein M. Participating in social activities helps preserve cogni-
Med 2003; 163:1069–75. tive function: an analysis of a longitudinal, population-based study of
86 Diener HC, Sacco RL, Yusuf S et al. Effects of aspirin plus extended- the elderly. Int J Epidemiol 2005; 34:864–71.
release dipyridamole versus clopidogrel and telmisartan on disability 101 Hakansson K, Rovio S, Helkala EL et al. Association between mid-life
and cognitive function after recurrent stroke in patients with marital status and cognitive function in later life: population based
ischaemic stroke in the Prevention Regimen for Effectively Avoiding cohort study. BMJ 2009; 339:b2462.

© 2012 The Authors. Vol 8, January 2013, 38–45 45

International Journal of Stroke © 2012 World Stroke Organization