Functional magnetic resonance imaging in chronic ischaemic stroke

Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
Functional magnetic resonance imaging
in chronic ischaemic stroke
rstb.royalsocietypublishing.org
Evelyn M. R. Lake1, Paolo Bazzigaluppi1,4 and Bojana Stefanovic1,2,3
1
Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada
Physical Sciences Platform, Sunnybrook Research Institute, Toronto, Ontario, Canada
3
Heart and Stroke Foundation Centre for Stroke Recovery, Ottawa, Canada
4
Fundamental Neurobiology, Krembil Research Institute, University Health Network, Toronto, Ontario, Canada
2
Review
Cite this article: Lake EMR, Bazzigaluppi P,
Stefanovic B. 2016 Functional magnetic
resonance imaging in chronic ischaemic stroke.
Phil. Trans. R. Soc. B 371: 20150353.
http://dx.doi.org/10.1098/rstb.2015.0353
Accepted: 1 June 2016
One contribution of 15 to a Theo Murphy
meeting issue ‘Interpreting BOLD: a dialogue
between cognitive and cellular neuroscience’.
Subject Areas:
neuroscience
Keywords:
stroke recovery, functional magnetic resonance
imaging, blood oxygen level dependent,
arterial spin labelling, electrophysiology
Author for correspondence:
Bojana Stefanovic
e-mail: [email protected]
Ischaemic stroke is the leading cause of adult disability worldwide. Effective
rehabilitation is hindered by uncertainty surrounding the underlying mechanisms that govern long-term ischaemic injury progression. Despite its potential
as a sensitive non-invasive in vivo marker of brain function that may aid in
the development of new treatments, blood oxygenation level-dependent
(BOLD) functional magnetic resonance imaging (fMRI) has found limited
application in the clinical research on chronic stage stroke progression.
Stroke affects each of the physiological parameters underlying the BOLD
contrast, markedly complicating the interpretation of BOLD fMRI data.
This review summarizes current progress on application of BOLD fMRI in
the chronic stage of ischaemic injury progression and discusses means by
which more information may be gained from such BOLD fMRI measurements. Concomitant measurements of vascular reactivity, neuronal activity
and metabolism in preclinical models of stroke are reviewed along with
illustrative examples of post-ischaemic evolution in neuronal, glial and
vascular function. The realization of the BOLD fMRI potential to propel
stroke research is predicated on the carefully designed preclinical research
establishing an ischaemia-specific quantitative model of BOLD signal contrast to provide the framework for interpretation of fMRI findings in clinical
populations.
This article is part of the themed issue ‘Interpreting BOLD: a dialogue
between cognitive and cellular neuroscience’.
1. Introduction
Although widely used in neuroscience for the study of healthy human brain
function, blood oxygenation level-dependent (BOLD) functional magnetic
resonance imaging (fMRI) has found limited application in clinical research
on brain disorders and diseases. Indeed, the perceived clinical potential of
BOLD fMRI remains unrealized. In this review, we focus on the application
of fMRI to studies of brain function in the chronic stage of stroke progression.
The large societal and personal burden of stroke, the mechanistic basis of stroke
in vascular injury and the strong dependence of BOLD signal on blood flow
make fMRI application to the chronic stage of stroke a particularly significant
research direction. Indeed, BOLD neuroimaging has been proposed to be a
robust in vivo marker of ischaemia-induced brain dysfunction/stroke recovery
that could aid in identification of the determinants of prognosis, and guide
the development of new ischaemic stroke treatments [1– 4].
The positive BOLD response is robustly observed in various brain regions under
a variety of input tasks/stimuli across many physiological conditions and in many
species. It reflects a focal decrease in deoxyhaemoglobin (dHb) content, which due to
the paramagnetism of dHb, its sequestration to red blood cells, and the geometry of
the focal vascular architecture, results in the decrease of local microscopic magnetic
field heterogeneities and thus lengthening of the apparent tissue spin–spin relaxation time during periods of increased neuronal activity. The BOLD fMRI signal is
hence a function of dHb concentration, which is in turn driven by adjustments in cerebral blood flow (CBF), cerebral blood volume (CBV) and cerebral metabolic rate of
& 2016 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
In clinical practice, stroke recovery is evaluated using one or
more structured neuropsychological tests, which incorporate
simple physical tasks and quality of life assessments [29].
Although neuropsychological tests are carefully designed to
evaluate whether a patient is improving, they are confounded
by self-reporting bias and high incidence of very gradual
gains (recovery over a long period often results in a flawed
perception of progress). Further, they offer little insight into
the underlying processes driving or impeding recovery [30].
To address this gap, several studies have tested whether
fMRI can provide a more objective metric of behavioural
impairment and/or functional recovery while still correlating
with performance on neuropsychological assessments.
3. Behavioural correlates of blood oxygenation
level-dependent functional magnetic
resonance imaging response
On the whole, evoked BOLD fMRI response measurements
following ischaemic stroke have shown pronounced alterations in amplitude and kinetics of the BOLD signal both in
4. Mechanistic basis of blood oxygenation leveldependent functional magnetic resonance
imaging response in stroke recovery
Notwithstanding the lack of correlation between BOLD fMRI
alterations and behavioural performance, researchers have
also attempted to use BOLD fMRI measurements to gain
insight into the neurobiological changes in the chronic stage
of ischaemic injury progression. Delayed, low amplitude,
absent or negative BOLD responses have been purported to
indicate dysfunction. BOLD response attenuation was thus proposed to result from compromised local perfusion due to
structural damage to the vasculature (caused by either ischaemia itself or pre-existing comorbidities) [33], exhausted
vasomotor reactivity due to chronically maximally dilated vasculature [34,35] or absent neuronal activity [31,36,37]. To test
these hypotheses, some fMRI studies probed neuronal activity
while others assessed cerebrovascular reactivity [34,35,38].
Employing tactile exploration of objects to elicit BOLD
signal changes and motor evoked potentials (assessed with
transcranial magnetic stimulation; TMS), Binkofski et al. [38]
conducted a longitudinal multimodal neuroimaging study
in patients imaged at one, two to four weeks and one
month following injury [38]. One week after stroke, BOLD
signal change and TMS revealed the representation of the
paretic hand fingers to be remapped to the adjacent cortical
2
Phil. Trans. R. Soc. B 371: 20150353
2. Clinical functional magnetic resonance
imaging research on ischaemic stroke recovery
the ipsi- and in the contralesional cortex months after the
ischaemic insult. In the majority of patients studied to date,
however, BOLD response was altered whether or not the
patient had made a good recovery, suggesting BOLD fMRI
may not be a sensitive marker of functional recovery. For
example, in patients with residual paralysis 9 –57 months
after stroke, Blicher et al. [31] used unilateral isolated wrist
extension –flexion and observed a positive BOLD response
in 55%, negative BOLD response in 9% and no BOLD
response in 36% of patients [31]. In patients with poor outcome, imaged more than 270 days after stroke, Newton
et al. [32] employed near-isometric wrist extension movements and found the peak BOLD response in ipsilesional
M1 to be delayed relative to contralesional M1 responses
during movement of either wrist. The patients exhibited
abnormally delayed M1 responses in the ipsilesional hemisphere when moving the paretic wrist. In mildly impaired
patients performing a finger- or hand-tapping task (4 –660
days after stroke), Pineriro et al. [33] observed the rate of
BOLD signal increase and maximum BOLD signal change
in the sensorimotor cortex of both hemispheres to be
decreased by approximately 30% in patients relative to controls. Rossini et al. [34] and Altamura et al. [35] studied
individuals who experienced a good recovery: 12 months following injury, direct median nerve stimulation elicited no
BOLD activation in either hemisphere in 25% of these
patients, while in 45% of the patients, activation was present
only in the contralesional hemisphere [34,35]. In another
study of patients with good recovery 1–2 years after stroke,
a visually guided bilateral handball squeeze task produced
a biphasic BOLD response, with patients showing earlier
onset and extended duration of the initial, negative BOLD
response followed by a positive BOLD response with a
delayed time-to-peak but normal peak magnitude [36].
rstb.royalsocietypublishing.org
O2 consumption (CMRO2); as well as the distribution of dHb in
the tissue, as determined by the microvascular architecture [5,6].
Stroke affects each of the physiological parameters underlying
BOLD contrast: ischaemia triggers perilesional angiogenesis
(preceded acutely by vessel loss) thus changing microvascular
architecture and affecting CBF and CBV; and modulates
CMRO2 due to neuronal loss [7], glial recruitment and activation
[8], generation of new vessels [9–12], axonal sprouting [13], and
synaptogenesis [14–16]. Consequently, BOLD may be a sensitive
and widely accessible marker of ischaemic sequelae. On the other
hand, because all the parameters dictating the BOLD signal are
differentially affected by ischaemic injury progression, the
interpretation of BOLD signal change following ischaemic
injury is complex [17] and BOLD neuroimaging in isolation has
been challenged to impact clinical research on stroke recovery.
This review summarizes progress on clinical and preclinical application of BOLD fMRI in the chronic stage of
ischaemic injury progression, and discusses means by which
more information may be gained from such BOLD fMRI
measurements. The focus on the chronic stage is motivated
by an increasingly recognized need for interventions in the
weeks to months following ischaemic injury as the majority
of ischaemic stroke patients (as many as 97% in some centres)
arrive at a care facility beyond the therapeutic window for safe
application of currently available treatments [18]. This research
direction is further supported by the incidence and temporal
evolution of spontaneous recovery: as many as 25% of ischaemic stroke patients show improvement with physical therapy
and continue to recover function more than 12 months following stroke [4,19 –26]. The therapeutic window for novel
treatments, aimed at enhancement of endogenous healing,
may thus be much longer than is often assumed [1– 3,26–28].
BOLD fMRI may thus also be valuable for elucidation of the
mechanisms which govern recovery in the long-term and for
identification of new, chronic stage treatment targets to
improve recovery in a much wider population of patients.
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
vasculature in the perilesional tissue could modulate blood
flow and volume and alter vascular reserve capacity and/or
neurovascular coupling [31,37]. Further, BOLD attenuation
may be secondary to increases in basal CBF of the peri-infarct
zone due to persistent collateral perfusion [37].
In addition to BOLD response attenuation, BOLD fMRI
studies of ischaemic injury progression have also shown
evidence of an altered spatial pattern of BOLD responses.
In particular, focal signal enhancement within the primary
motor cortex, increased activity in non-primary sensorimotor
areas, and/or increased activity in the contralateral
hemisphere have all been reported [32,39–43]. However,
whether such remapping of the BOLD response provides
salient insight and if so, whether it signifies recovery or is a
manifestation of dysfunction remains controversial [32,43,44].
Some studies have provided evidence of BOLD response
remapping, but such remapping was not found to correlate
with outcome. For instance, employing the finger opposition
task in patients 5–43 months after injury, with varied NIHSS
and Rankin scores, Cao et al. [43] examined event related
BOLD responses [43]. Seventy five per cent of patients
showed extended activation of the contralesional sensorimotor
cortex; 50% showed bilateral activation of the primary sensorimotor cortex; 50% showed contralesional activation alone
during paretic hand movement; 25% showed activation in contralesional premotor and dorsolateral prefrontal cortex, and
25% of patients exhibited activation in the contralesional supramarginal gyrus and premotor cortex during paretic hand
movement. Despite such extensive BOLD response reorganization in the patients, no correlation between degree of
remapping and neuropsychological test scores was observed.
In contrast, some authors have postulated that fMRI
response remapping, namely activation in the contralesional
motor cortex, increased sensorimotor activation volume and
activation within the perilesional tissue indicate remapping
of neuronal activity patterns and/or recruitment of
additional neurons and are manifestations of recovery.
In particular, Cramer et al. [39] employed a finger- or handtapping task to measure BOLD responses in recovered
stroke patients between 11 days and 15 months following
ischaemia [39]. Seventy per cent of the patients were found
to have a greater activation volume in the contralesional sensorimotor cortex than that observed in control subjects. Most
patients with greater activation volumes in the contralesional
sensorimotor cortex also showed increased activation in the
contralesional cerebellum, premotor cortex and both the ipsilesional and contralesional supplementary motor areas [39].
In addition, degree of peri-infarct activation correlated with
clinical improvement. Similarly, Marshall et al. [40] used the
finger –thumb opposition task to measure BOLD responses
in patients within the first few days of stroke onset and
again 3–6 months thereafter [40]. In the first few days of
injury and in the chronic phase after stroke, patients
showed greater activation in the ipsilesional sensorimotor
cortex, ipsilesional posterior parietal and bilateral prefrontal
regions than did the control subjects. Moreover, ipsilesional
Phil. Trans. R. Soc. B 371: 20150353
5. Blood oxygenation level-dependent functional
magnetic resonance imaging response
remapping
3
rstb.royalsocietypublishing.org
areas. At 2–4 weeks, there was a transient lack of BOLD
responses despite clinical improvement of hand function
and preserved motor evoked potentials on TMS. At one
month after ischaemia, BOLD responses were again detected;
motor evoked potentials were still present, and further clinical improvement was made. Similarly, Rossini et al. [34]
and Altamura et al. [35] observed absent or reduced BOLD
responses in both hemispheres, yet measured stereotypical
somatosensory evoked fields from ipsi- and contralesional
hemispheres via magnetoencephalography (MEG) [34,35].
Combined, these data provide evidence that at least in the
well-recovering patients BOLD fMRI attenuation was not
due to an absence of neuronal activity.
With respect to impairment in cerebrovascular reactivity,
some studies did not, while others did, find it correlated with
BOLD attenuation. Rossini et al. [34] measured cerebrovascular
reactivity to hypercapnia (inhalation of 7% CO2) with transcranial doppler colour-coded duplex sonography [34]. In 25% of
the patients, missing BOLD fMRI activation was accompanied
by impaired vascular reactivity to hypercapnia. In the remaining
75% of patients, who exhibited some BOLD response, no correlation was found between the amplitude of the BOLD response
and the degree of hypercapnic reactivity of the major arterial
vessels, though impaired microvascular reactivity was thus not
excluded. Moreover, Blicher et al. [31] observed no BOLD
signal change, on average, in M1 in response to unilateral isolated
wrist extension–flexion in patients with residual paralysis, yet
arterial spin labelling (ASL) and VASO-FLAIR (vascular-spaceoccupancy fluid attenuated inversion recovery) showed this
task elicited CBF and CBV increases that were indistinguishable
from those observed in control subjects [31].
Krainik et al. [37] used fMRI to estimate cerebrovascular
reactivity by comparing BOLD signal between normal breathing and hyperventilation-elicited hypocapnia (which
increases oxygen extraction, induces vasoconstriction and
reduces CBF) in fully recovered patients with frontal lobe
stroke 11–85 days after injury [37]. Hyperventilation resulted
in a smaller BOLD signal decrease in the ipsilesional primary
sensorimotor cortex and supplementary motor areas relative
to contralesional hemispheres and relative to control subjects
(bilaterally). Hyperventilation-elicited BOLD signal decrease
in the contralesional supplementary motor areas was larger
than that in the control subjects. In a separate experiment in
the same study, BOLD response during a repeated hand grip
task was decreased in the ipsilesional primary sensorimotor
cortex and ipsilesional supplementary motor areas despite a
full clinical recovery and no evidence of lesions in the primary
sensorimotor cortex, supplementary motor areas, cerebellum,
basal ganglia or thalamus on three-dimensional T1-weighted
MRI [37]. The authors suggested that impaired cerebrovascular
reactivity was a predictor of impaired motor-related BOLD
response in the sensorimotorcortex during contralateral
movements [37]. The discrepancy between this study and
those by Rossini [34] and Blicher [31] may be attributed to
methodological differences.
Preceding reports of persistent neuronal activity, vascular
reactivity impairment and alterations in BOLD response
amplitude and temporal profile have prompted further
mechanistic hypotheses of modified neurovascular coupling
due to abnormal vascular anatomy and/or changes in tissue
metabolic needs. Ultrastructural changes in cerebral vessels
caused by atherosclerosis, perilesional gliosis, and/or disruption of aminergic and cholinergic fibres that innervate the
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
Preclinical studies offer well-controlled experimental conditions
including tightly regulated ischaemic injury parameters as well
4
Phil. Trans. R. Soc. B 371: 20150353
6. Preclinical functional magnetic resonance
imaging research on ischaemic stroke recovery
as careful control of comorbidities, and thus provide a powerful
means for examining the neurobiological determinants of stroke
recovery and establishing a quantitative model of BOLD signal
contrast post-ischaemic injury. However, translation potential
of preclinical findings critically depends on careful experimental design and thoughtful consideration of salient interspecies
differences. Of note, there are structural, functional and hence
metabolic differences between humans and animals in addition
to differences in preclinical versus clinical BOLD fMRI
implementation. The higher nominal spatial resolution used in
preclinical fMRI (e.g. 0.32 mm3 as in Weber et al. [46]) versus clinical fMRI (e.g. 27 mm3) research is to be evaluated in the light of
three orders of magnitude smaller rodent versus human total
brain volume [46,47]. A preclinical BOLD fMRI voxel thus represents a 100 times larger relative brain volume than that
comprising a clinical BOLD fMRI voxel. Additionally, cortical
neuronal density in rats is approximately twice that of
humans, whereas the ratio of glia to neurons is lower in rats
(approx. 1 : 1) than in humans [48,49]. Interestingly, fundamental
cortical communication processes (e.g. action potentials, pre- and
post-synaptic potentials, glial neurotransmitter and Kþ clearance) were estimated to have the same relative energy costs in
rats and in humans [50]. On a final methodological note, the sensitivity of BOLD fMRI to magnetic susceptibility gradients limits
the invasiveness of manipulations that can be performed on the
animals and hence the spectrum of complementary techniques
that can be done in combination with fMRI.
As combining fMRI with other methodologies represents
a major technical challenge, there are only a few multimodal
preclinical fMRI reports which monitor ischaemia-induced
neurovascular remodelling during the chronic period (more
than 7 days after the injury): further, their conclusions with
respect to neurovascular coupling post-ischaemia have not
been consistent. In particular, Shih et al. [51] measured
intra-cortical local field potentials (LFPs) and CBF and CBV
fMRI signal changes with forepaw stimulation up to four
weeks following a 20-min middle carotid artery (MCA)
occlusion [51]. Noxious forepaw stimulation-elicited negative
fMRI responses were associated with LFP recovery 28 days
following ischaemic insult. Negative fMRI responses were
interpreted as evidence of uncoupling between neuronal
activity and local vascular reactivity. However, when
Weber et al. [46] performed the adhesive tape removal test,
BOLD fMRI and electroencephalography (EEG; through
skull) up to seven weeks following MCA occlusion [46]
they found that at 2–7 days after ischaemia, rats were
impaired and showed attenuated BOLD responses to forepaw stimulation, but within 1–4 weeks they exhibited
recovery on the adhesive tape removal test and showed reestablishment of normal BOLD responses, suggesting that
BOLD fMRI could be used as an indicator of functional
loss and restoration. Moreover, they observed a tight coupling of electrical brain activity and haemodynamic
responses in early and late phases after stroke, suggesting
preservation of neurovascular coupling in the somatosensory
cortex in this model. The complexity of the perilesional neurovascular changes induced by focal ischaemia is further
exemplified by two multimodal preclinical fMRI examinations from our own work. Using intra-cortical microinjection of endothelin-1 into the right sensorimotor cortex
of adult Sprague Dawley rats, we induced ischaemic lesions
of 50 + 10 mm3 volume on T2-weighted MRI. Three weeks
following stroke, BOLD and continuous ASL readings were
rstb.royalsocietypublishing.org
(relative to contralesional) activity in the sensorimotor
cortex increased progressively as the paretic hand regained
function [40]. Combined, these studies suggest that a dynamic
bihemispheric reorganization of motor networks probably
contributes to recovery following focal ischaemia [39,40].
Some authors suggested that remapping of the BOLD
response correlates with a poor prognosis. For instance, Ward
et al. [41] employed a dynamic visually paced handgrip task
at least three months after stroke in patients with documented
deficits lasting a minimum of 48 h following stroke, and with
a varied degree of recovery at the time of imaging [41]. Patients
with poor recovery were more likely to recruit a number of
motor-related brain regions over and above those activated in
the control group, whereas patients showing recovery were
more likely to show an fMRI activation pattern similar to control subjects [41]. Across the entire patient group, Ward et al.
[41] were able to demonstrate a negative correlation between
outcome and the degree of task-related activation in the supplementary motor areas, cingulate motor areas, premotor
cortex, posterior parietal cortex and cerebellum [41]. In another
example, Feydy et al. [42] used repetitive first clench (n ¼ 10), or
a modified movement for those unable to fist clench (n ¼ 4), to
measure evoked BOLD responses at three imaging sessions
over a period of 1–6 months following ischaemic stroke [42].
At earlier time points, patients showed contralateral and ipsilateral recruitment of sensorimotor cortex, supplementary
motor areas, frontal premotor cortex, frontal premotor areas
and, less commonly, posterior parietal areas. By latter imaging
time points, this recruitment subsided in 66% of patients; the
remaining 34% of the patients showed a higher incidence
of M1 lesions [42]. The authors suggested that ipsilateral
recruitment after stroke corresponds to a compensatory corticocortical process, which may be due to a decrease in
inhibition caused by an M1 lesion and damage to the underlying reciprocal corticocortical connections from M1 to the
secondary motor areas, leading to more extensive ipsilateral
and contralateral activation in worse-off patients.
In summary, clinical research has provided ample evidence
of BOLD modulation post-ischaemic injury, but findings on the
correlation between BOLD fMRI response features (magnitude,
kinetics or spatial pattern) and clinical outcome have been inconsistent [44]. Mechanistically, there has been some, though still
varied, support for BOLD modulation resulting from alterations
in impairment in cerebrovascular reactivity and alterations in
neurovascular coupling, though details of these mechanisms
remain uncertain. Concomitant measurements of metabolism,
haemodynamics and BOLD in ischaemic stroke patients
during the weeks to months following injury need to be more
widely employed [45] to improve our understanding of the natural progression of the disease in patients and facilitate
interpretation of BOLD response signals in studies of novel therapeutic agents. At the same time, a detailed mechanistic
understanding of the determinants of BOLD contrast postischaemic injury will necessitate measurements of changes in
morphology and function of specific cellular populations and
hence calls for studies in preclinical models of ischaemic stroke.
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
5
(i) DBOLD
(a)
(b)
(i) glial cells
×103
7
6
6
4
5
2
4
0
3
–2
2
1
–4
0
–6
L
ET-1 injection site
R
R
L
(ii) Dperfusion
×103
7
6
5
4
3
2
1
0
100
L
(ii) astrocytes
R
R
L
80
60
40
20
0
SI (arb. units) Dperfusion (arb. units)
(iii) perfusion
×103
7
6
5
4
3
2
1
0
6
L
(iii) blood vessels
R
L
R
5
4
3
2
1
0
SI (arb. units) ASL (%)
Figure 1. Functional MRI and immunohistochemistry in a rat model of focal sensorimotor ischaemia. One week following right sensorimotor cortex injection of
endothelin-1, BOLD fMRI responses (a(i)) and blood flow changes (a(ii)) elicited by hypercapnia (10% CO2), as well as resting perfusion (a(iii)) are shown overlaid on
the corresponding structural T2-weighted images. Perilesional M1/M2 BOLD response (1.09 + 0.07%) was attenuated by approximately 60% relative to its level in
the contralesional cortex (2.5 + 0.1%), with an ipsi- to contralesional ratio of cortical DBOLD of 0.44 + 0.02. Perilesionally, blood flow increase in the M1/M2 ROI
in response to hypercapnia, of 39 + 5%, was strongly potentiated (ratio: 3.25 + 0.04) relative to the contralateral M1/M2 ROI (12 + 2%). Similarly, perilesional
resting perfusion (1.66 + 0.09%) was doubled (ratio: 1.95 + 0.01) relative to that of the contralesional cortex. Post-mortem immunohistochemistry was performed
in the same animals immediately following imaging, with Iba-1 staining for microglia (b(i)), GFAP for astrocytic activation (b(ii)), and RECA-1 for blood vessel
endothelium (b(iii)). Microglial recruitment (b(i)) as well as a pronounced astrocytic activation (b(ii)) were evident in the peri-infarct zone. Endogenous angiogenesis
was also observed in the perilesional tissue (b(iii)). Scale bar in (a) 2 mm; scale bar in (b) 500 mm.
acquired to measure BOLD and CBF signal changes elicited
by hypercapnia (10% CO2), while skilled reaching ability
was evaluated using the Montoya Staircase skilled reaching
task [52]. In addition, we recorded intra-cortical spontaneous
LFP responses simultaneously from both hemispheres using
a pair of multi-electrode arrays composed of 16 (2 8) Pl/Ir
electrodes ( placed within the cortex such that the anterior
pair of electrodes was within the core of the lesion) or
performed immunohistochemical analysis post-mortem.
These data allowed the investigation of BOLD signal
changes, cerebrovascular reactivity, motor recovery and
neuronal functioning, as well as post-mortem changes in
select cellular populations.
In the first example, we considered the subacute period
(7 days following ischaemic injury). Representative fMRI and
immunohistochemical results are shown in figure 1. At this
time point, hypercapnia elicited an attenuated BOLD response
in the perilesional M1/M2 region of interest (ROI) relative to
the contralesional M1/M2 ROI. The perilesional BOLD
response was attenuated by approximately 60%: 1.09 + 0.07%
in the perilesional cortex versus 2.5 + 0.1% in the contralesional
cortex. However, perfusion increases with hypercapnia in the
perilesional M1/M2 cortex were triple those in the contralesional M1/M2 (39 + 5% perilesionally versus 12 + 2%
contralesionally). Perilesional resting perfusion was twice the
contralesional perfusion: 1.66 + 0.09% versus 0.85 + 0.04%.
Phil. Trans. R. Soc. B 371: 20150353
SI (arb. units) DBOLD (%)
rstb.royalsocietypublishing.org
ET-1 injection site
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
7. Conclusion and future work
Notwithstanding the complexities, BOLD fMRI still has
significant potential to advance the research of neurophysiological changes in the chronic stage of stroke recovery, due to
its easy implementation in patient populations and its robustness. Longitudinal BOLD fMRI measurements based on
signal intensity changes have consistently proven more
reliable than activation volumes in ischaemia patients [61]
and in healthy subjects [62–65]. For the potential of fMRI
to be fully realized, however, a much more comprehensive
understanding of the neurovascular coupling over this
period is needed. Such understanding necessitates that
BOLD fMRI data be widely acquired and combined with
other, more specific cellular measures as infarct affects all
cell types (neurons, glia and vascular endothelial cells) both
6
Phil. Trans. R. Soc. B 371: 20150353
time, ongoing perilesional vascular remodelling catches up
with the metabolic demands of the tissue.
To examine the attenuation of the BOLD response perilesionally in the FR122047-treated animal, despite the
increase in vascular reactivity, we performed intracranial
array electrode recordings in the same animals (figure 3).
During rest, the brain generates oscillations in different
frequency bands, which are believed to be the neuronal
correlate of different behaviours and cognitive states (for
review, see Womelsdorf et al. [59]). We found high coherence
in the contralesional cortex of both animals (green traces in
figure 3a(iii) and b(iii)). In the perilesional cortices (black
traces in figure 3a(iii) and b(iii)), the coherence was lower,
indicating stroke-induced impairment in neuronal functioning; the treated animal, however, exhibited a smaller degree
of attenuation across all frequency bands (figure 3b(iii)) than
did the placebo administered animal (figure 3a(iii)). Notwithstanding the attenuated BOLD response, these data suggest
COX-1 inhibition in the chronic phase after ischaemia may
exert a beneficial effect on neuronal communication and putative cortical reorganization. To examine underlying neuronal
metabolic activity, we estimated neuronal power as a correlate
of neuronal metabolic intake. In both animals, the average
neuronal power was attenuated in the perilesional relative to
that in the contralesional cortex in each frequency band. However, in all frequency bands this attenuation was smaller in the
FR122047-treated animal than in the placebo administered
animal (control versus treated): u, 73% versus 90%; a, 37%
versus 58%; b, 21% versus 31%; low g, 12% versus 13%; and
g, 24% versus 28% (cf. Figure 3a(iv) and b(iv)). In summary,
the treated animal exhibited higher relative ( perilesional
versus contralesional) coherence and higher relative neuronal
power. The perilesional BOLD response attenuation in the
FR122047-treated animal was thus associated with improved
neuronal function and may be attributable to high oxygen
extraction accompanying elevated neuronal activity, spurred
on by inhibition of microglial invasion by COX-1. These
examples are in line with the acute phase observations of
depressed CBF and metabolic responses despite the recovery
of somatosensory evoked potentials [60]. Combined, these
findings emphasize the need for multimodal examination of
neuronal, vascular and glial function as well as metabolism
over the course of ischaemic injury progression to establish a
framework for interpreting BOLD signal changes in the
highly dynamic peri-infarct region.
rstb.royalsocietypublishing.org
Perilesional BOLD fMRI response attenuation thus
accompanied vascular hyper-reactivity. Immunohistochemical
analysis of these brains demonstrated perilesional neuroinflammation, with dense microglial invasion (figure 1b(i)) and
pronounced astrocytic reactivity (figure 1b(ii)). Proximal to
the site of injury, RECA-1 staining (figure 1b(iii)) showed an
increase in vascular endothelial density indicating endogenous
angiogenesis, as previously reported [9–12]. In parallel
experiments, Montoya staircase testing demonstrated approximately 70–80% impairment in skilled reaching ability of the left
forepaw. Intra-cortical LFP recordings collected 7 days after
injury showed a 10–30% lower ipsi- to contralesional power
ratio (across all frequency bands), suggesting silencing of the
neuronal network in the peri-infarct zone.
BOLD response attenuation and motor impairment were
thus accompanied by a hyper-reactive and proliferating vasculature on the one hand, and neuronal hypoactivity on the
other. We hypothesized that the pronounced increase in perilesional vascular reactivity and local baseline perfusion were
triggered by increased metabolic needs (O2 consumption
and glutamate/glutamine cycling) of the densely packed
neuroinflammatory cells. Direct measurements of tissue
metabolism are needed to examine this hypothesis.
In the second example, we considered the neurovascular state three weeks following the endothelin-1 elicited
ischaemic insult. Representative fMRI and electrophysiological results are shown in figures 2 and 3, respectively.
In these experiments, we investigated the effects of selective
cyclooxygenase-1 (COX-1) inhibition through the chronic
administration of FR122047 (Cedarlane Laboratories,
Burlington, ON, Canada) from 7 to 19 days following stroke,
followed by fMRI or electrophysiological recordings 20 or 21
days after stroke. In clinical and preclinical ischaemic stroke
studies, COX-1 metabolites have been shown to contribute to
injury during the chronic phase of stroke [53,54]. In preclinical
studies, COX-1 inhibition has been shown to prevent neuronal
loss [55], reduce oxidative stress (by disrupting free radical
production) [56,57], attenuate neuroinflammation [53] and
increase resting vascular tone [58]. By sustained inhibition of
COX-1 in the chronic stage of stroke, we expected to
ameliorate neurovascular function through reduction in microglial invasion. We therefore measured BOLD and flow
responses to hypercapnia (figure 2) in vehicle-administered
and FR122047-treated animals, and subsequently recorded
neuronal activity via intracranial electrophysiology (figure 3).
Twenty-one days after stroke, the BOLD response to hypercapnia in the vehicle-administered animal was not lateralized
(2.6 + 0.1% versus 2.8 + 0.1%). However, perilesional blood
flow increase to hypercapnia was double that of the contralesional cortex (ratio of 2.3 + 0.1). In the FR122047-treated
animal, BOLD response to hypercapnia was attenuated by
40% perilesionally (3.7 + 0.3% in the peri- versus 6.1 + 0.3%
in the contralesional cortex). Nonetheless, its perilesional perfusion response to hypercapnia was elevated by 20% over the
contralesional level (ratio of 1.2 + 0.1). Although perilesional
vascular reactivity was elevated in both animals, the control
animal exhibited no BOLD response lateralization whereas
the FR122047 rat showed attenuated BOLD response in the
perilesional tissue.
Moreover, the non-lateralized BOLD response of the control animal is in contrast with the attenuated BOLD response
in the subacute study, though both are accompanied by elevated perilesional vascular reactivity. It is possible that with
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
(a)
7
×103
10
7
L
(ii) FR122047
L
R
R
8
5
6
3
4
2
0
0
Phil. Trans. R. Soc. B 371: 20150353
1
SI (arb. units) BOLD (%)
(iv) .
.
normalized SI (arb. units)
(iii)
0.8
0.8
0.4
0.4
0
0
5
9
13
5
0.8
0.8
0.4
0.4
0
9
13
0
5
(b)
(i) control
9
time (min)
13
5
ET-1 injection site
9
time (min)
13
(ii) FR122047
×103
7
5
100 L
R
L
R
80
60
3
40
1
20
0
0
SI (arb. units) Dperfusion (arb. units)
Figure 2. Functional MRI following chronic COX-1 inhibition in a rat model of focal sensorimotor ischaemia. Maps of BOLD signal changes elicited by hypercapnia
(10% CO2) in two representative rats after ALZET-mini pump intracerebroventricular administration of vehicle (a(i)) or FR122047 (a(ii)) are shown overlaid on the
corresponding T2-weighted structural images (SI, signal intensity). The treatment was initiated 7 days following injection of endothelin-1 into the right sensorimotor
cortex and terminated after 10 days, with imaging performed 24 h following treatment cessation. The average normalized SI in the M1/M2 ROIs (top, black trace:
contralateral ROI; bottom, blue trace: perilesional ROI) for the vehicle-administered control subject are plotted in (a(iii)) and for the FR122047-treated animal are
displayed in (a(iv)). These traces were normalized, animal-wise, to the maximum signal in that animal’s contralateral M1/M2 ROI. Periods during which 10% CO2
was delivered are indicated by grey rectangles. Concomitant blood flow changes are shown in (b(i)) for a vehicle-administered rat; and in (b(ii)) for a FR122047treated animal. The control animal showed no lateralization in DBOLD (2.6 + 0.1% versus 2.8 + 0.1%), with a peri- to contralesional ratio of 0.9 + 0.1; and
doubled peri- versus contralesional DCBF, with a peri- to contralesional ratio of 2.3 + 0.1. In the FR122047-treated animal, BOLD response was attenuated
peri- (3.7 + 0.3%) versus contralesionally (6.1 + 0.3%); whereas perilesional Dperfusion was increased by 20% relative to that of the contralesional cortex
(ratio of 1.2 + 0.1). Scale bar (a(i) and b(i)) 2 mm. Although the across-brain average BOLD responses were higher in the treated rat than in the vehicle-administered rat, this inter-subject variability was not significant cohort-wise; but it did motivate the interhemispheric normalization within each subject.
acutely and chronically. As was the case with BOLD fMRI
applications to neuroscientific research in physiological conditions, multimodal preclinical studies can be used to
quantitatively characterize neurovascular state in ischaemic
progression and yield an ischaemia-specific model of BOLD
signal contrast to support the interpretation of fMRI findings
in clinical research. In particular, combining BOLD fMRI with
in vivo measurements of glucose and oxygen metabolism in
rstb.royalsocietypublishing.org
ET-1 injection site
(i) control
addition to direct measures of vascular state will provide
important insight into the determinants of BOLD contrast
in ischaemic injury progression. In addition, combining
BOLD with CBF measurements in calibrated fMRI experiments may provide quantitative estimates of the relative
oxygen consumption during the execution of a task
[66 –69]. This approach relies on CMRO2 measurements
during iso-metabolic perturbation and the conservation of
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
(a)
8
B 1.75
1.75
1.50
1.50
1.25
1.25
1.00
1.00
0.75
0.75
0.4
0.50
0.50
0.2
0.25
0.25
0
20
(iii)
|coherence|2
1.0
(b)
40
60
frequency (Hz)
(iv) 1.0
0.85 0.99
± 0.02 ± 0.02
0.95
± 0.02
0.89
± 0.02
0.80
± 0.02
0.8
0.6
0.56
0.4 0.66 0.42
± 0.01 ± 0.01
± 0.01
q
a
b
(2–8 Hz) (9–13 Hz) (15–30 Hz)
0.47
± 0.02
low g
(31–50 Hz)
0.46
± 0.03
g
(51–80 Hz)
80
0
100
0.8
g
100
low g
80
b
40
60
frequency (Hz)
a
20
0.6
q
0
0.8
R/L ratio power
distance from Bregma (mm)
1.0
0.6
0.4
0.2
(ii)
(i)
1.00
1.00
0.75
0.75
0.50
0.50
0.25
0.25
0
20
40
60
frequency (Hz)
80
100
0.6
0.4
0.2
0
20
(iii)
|coherence|2
1.0
40
60
frequency (Hz)
(iv)
0.82 0.99
± 0.02 ± 0.02
0.94
± 0.03
0.89
± 0.02
0.78
± 0.03
0.8
0.6
0.87 0.75
0.4 ± 0.01 ± 0.01
0.73
± 0.01
q
a
b
(2–8 Hz) (9–13 Hz) (15–30 Hz)
0.63
± 0.02
low g
(31–50 Hz)
0.61
± 0.05
g
(51–80 Hz)
1.0
0.8
80
0
100
g
1.25
low g
1.25
0.8
b
1.50
a
1.50
q
1.75
R/L ratio power
distance from Bregma (mm)
1.0
1.75
0.6
0.4
0.2
Figure 3. (Caption opposite.)
the assumed blood flow and volume coupling between
iso-metabolic stimulation and the functional stimulus/task
under investigation. The identification of clinically translatable iso-metabolic perturbers remains an active area of
research [70], as does the relationship between flow and
volume changes under various ( patho)physiological
conditions. Further insight into the metabolic changes in
the peri-infarct tissue may be attained by combining BOLD
fMRI with complementary MRI approaches such as amide
proton transfer imaging to assess tissue pH. Its preclinical
applications have shown that MCAO-induced pH deficits
extend beyond the region of diffusion changes, but are
Phil. Trans. R. Soc. B 371: 20150353
(ii)
rstb.royalsocietypublishing.org
(i)
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
comprehensive characterization of treatment effects. Another
important aspect in treatment development for stroke is elucidating the contributors to subacute reperfusion damage/
oedema. Recent research suggests that reperfusion-induced
injury in the subacute stage may be prevented by inhibiting
microglial and astrocytic production of matrix metalloproteinases responsible for delayed blood–brain barrier breakdown
(reviewed in [79]). fMRI is particularly beneficial for these investigations due to its robustness, sensitivity and easy repeatability,
all the while providing whole brain coverage and thus affording
the interpretation of signal changes in affected areas in relation
to those in distal (notionally intact) brain regions. fMRI use in
the development of such interventions may yield indispensable
information on spatio-temporal changes in neurovascular functioning (e.g. oxygen extraction with calibrated BOLD) by at-risk
tissue, furthering our understanding of the mechanisms of subacute reperfusion injury, which may allow the extension of the
time-window for safe and effective application of thrombolytics.
Ethics. All experimental procedures discussed in this paper were
approved by the Animal Care Committee of the Sunnybrook
Research Institute, which adheres to the Policies and Guidelines of
the Canadian Council on Animal Care and meets all the requirements
of the Provincial Statute of Ontario, Animals for Research Act as well
as those of the Canadian Federal Health of Animals Act.
Authors’ contributions. All authors conceived and wrote the review.
Competing interests. The authors have no competing interests to declare.
Funding. The authors would like to acknowledge the funding sources:
CIHR, CCNA, Heart and Stroke Canadian Partnership for Stroke
Recovery, Heart and Stroke Foundation of Canada.
References
1.
2.
3.
Hallett M. 2001 Plasticity of the human motor
cortex and recovery from stroke. Brain Res. Rev. 36,
169–174. (doi:10.1016/S0165-0173(01)00092-3)
Seil FJ. 1997 Recovery and repair issues after stroke
from the scientific perspective. Curr. Opin. Neurol.
10, 49 –51. (doi:10.1097/00019052-19970200000010)
Steinberg BA, Augustine JR. 1997 Behavioral,
anatomical, and physiological aspects of recovery
of motor function following stroke. Brain Res.
Rev. 25, 125–132. (doi:10.1016/S0165-0173
(97)00013-1)
4.
5.
6.
Lee RG, van Donkelaar P. 1995 Mechanisms
underlying functional recovery following stroke.
Can. J. Neurol. Sci. 22, 257– 263. (doi:10.1017/
S0317167100039445)
Buxton RB, Wong EC, Frank LR. 1998 Dynamics of
blood flow and oxygenation changes during brain
activation: the balloon model. Magn. Reson. Med.
39, 855– 864. (doi:10.1002/mrm.1910390602)
Davis TL, Kwong KK, Weisskoff RM, Rosen BR. 1998
Calibrated functional MRI: mapping the dynamics of
oxidative metabolism. Proc. Natl Acad. Sci. USA 95,
1834 –1839. (doi:10.1073/pnas.95.4.1834)
7.
8.
9.
De Girolami U, Crowell RM, Marcoux FW. 1984
Selective necrosis and total necrosis in focal cerebral
ischaemia: neuropathologic observations on
experimental middle cerebral artery occlusion in the
macaque monkey. J. Neuropathol. Exp. Neurol. 43,
57– 71. (doi:10.1097/00005072-198401000-00005)
Jin R, Yang G, Li G. 2010 Inflammatory mechanisms
in ischemic stroke: role of inflammatory cells.
J. Leukoc. Biol. 87, 779– 789. (doi:10.1189/jlb.
1109766)
Hayward NM, Yanev P, Haapasalo A, Miettinen R,
Hiltunen M, Gröhn O, Jolkkonen J. 2011 Chronic
Phil. Trans. R. Soc. B 371: 20150353
more circumscribed than hypoperfusion [71], and promising
preliminary results have also been obtained in stroke patients
[72,73]. Owing to the spatial heterogeneity of the metastable
peri-infarct zone, these studies will profit from improved
spatial resolution [74]. Moreover, such assessments will
need to be undertaken longitudinally, at multiple time
points following injury due to the highly dynamic neurovascular changes in the chronic stage of ischaemic injury
progression. Finally, the generalizability of the preclinical
findings will be greatly extended by careful consideration of
the appropriateness of the preclinical model of ischaemia
[75], modelling of relevant comorbidities that are known to
exert a profound impact on outcomes in patients [76], as
well as standardization of these protocols to enable the integration of data across different studies and centres [77]. The
mechanistic understanding of ischaemic injury and recovery
ensuing from the careful application of BOLD fMRI in
combination with other modalities in preclinical models
will be invaluable in the development of new therapeutic
treatments for stroke.
In particular, fMRI may prove a valuable tool in the preclinical development of novel pleiotropic therapeutic approaches,
called for by the evidence of perilesional plasticity and thus
potential for functional recovery and the hitherto poor efficacy
of exclusively neuroprotective approaches [78]. In addition to
affecting neuronal populations, these treatments will target
inflammatory cells and cerebral vasculature in the chronic
stage of stroke [58], making multimodal fMRI (combining
BOLD fMRI with MRI-based measurements of blood flow, glucose metabolism, pH and inflammation) a useful tool for the
9
rstb.royalsocietypublishing.org
Figure 3. (Opposite.) Neuronal activity following chronic COX-1 inhibition in a rat model of focal sensorimotor ischaemia. Intracranial array electrophysiological
recordings were done in the same animals shown in figure 2. Coherence and power estimates were evaluated in the control animal (a) and in the FR122047treated animal (b). Normalized heat-plots of the multi-electrode array recordings from contralesional (left (i)), and ipsilesional (right (ii)) hemispheres. We recorded
from 0 to 1.75 mm from Bregma and plotted distance along the x-axis, and from 0.3 Hz to 100 Hz and plotted frequency along the y-axis. The contralesional
hemispheres (a(i) and b(i)) magnitude squared coherence (msc) estimates between neighbouring electrodes showed a high degree of coherence across the
cortex throughout the frequency spectrum in the right (contralesional) hemispheres. In the placebo administered animal, the ipsilesional hemisphere
(a(ii)) msc was reduced, especially immediately proximal to the site of injury (within 0 – 0.25 mm of Bregma). In the FR122047-treated animal, the msc reduction
in the perilesional tissue was less pronounced. The cross-electrode averaged msc is plotted for the contralesional hemisphere (in green) and for the ipsilesional
hemisphere (in black) for the control animal (a(iii)) and for the FR122047-treated animal (b(iii)). The contralesional hemispheres (green in a(iii) and b(iii)) showed a
higher coherence than the ipsilesional hemispheres (black in a(iii) and b(iii)) across all frequency bands. Contra- versus ipsilesional msc ratios in each band are
quoted below and above the corresponding traces in panels (a(iii)) and (b(iii)). The ipsi- to contralesional ratio of the average neuronal power within each frequency
band is plotted in (a(iv)) and (b(iv)). In both animals, the average neuronal power was attenuated in the ipsilesional hemisphere relative to contralesional
hemisphere in each frequency band. However, in all frequency bands, this attenuation was greater in the vehicle-only administered animal relative to the
FR122047-treated animal: u, 73% versus 90%; a, 37% versus 58%; b, 21% versus 31%; low g, 12% versus 13%; g, 24% versus 28%.
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
local haemodynamics? Brain 127, 99 –110. (doi:10.
1093/brain/awh012)
Altamura C et al. 2007 fMRI-vs-MEG evaluation
of post-stroke interhemispheric asymmetries in
primary sensorimotor hand areas. Exp. Neurol. 204,
631–639. (doi:10.1016/j.expneurol.2006.12.017)
Roc AC, Wang J, Ances BM, Liebeskind DS, Kasner
SE, Detre JA. 2006 Altered hemodynamics and
regional cerebral blood flow in patients with
hemodynamically significant stenoses. Stroke 37,
382–387. (doi:10.1161/01.STR.0000198807.
31299.43)
Krainik A, Hund-Georgiadis M, Zysset S, von Cramon
DY. 2005 Regional impairment of cerebrovascular
reactivity and BOLD signal in adults after stroke.
Stroke 36, 1146 –1152. (doi:10.1161/01.STR.
0000166178.40973.a7)
Binkofski F, Seitz RJ. 2004 Modulation of the BOLDresponse in early recovery from sensorimotor stroke.
Neurology 63, 1223 –1229. (doi:10.1212/01.WNL.
0000140468.92212.BE)
Cramer SC, Nelles G, Benson RR, Kaplan JD, Parker
RA, Kwong KK, Kennedy DN, Finklestein SP, Rosen
BR. 1997 A functional MRI study of subjects
recovered from hemiparetic stroke. Stroke 28,
2518– 2527. (doi: 10.1161/01.STR.28.12.2518)
Marshall RS, Perera GM, Lazar RM, Krakauer JW,
Constantine RC, DeLaPaz RL. 2000 Evolution of
cortical activation during recovery from corticospinal
tract infarction. Stroke 31, 656 –661. (doi:10.1161/
01.STR.31.3.656)
Ward NS, Brown MM, Thompson AJ, Frackowiak RS.
2003 Neural correlates of outcome after stroke: a
cross-sectional fMRI study. Brain 126, 1430–1448.
(doi:10.1093/brain/awg145)
Feydy A, Carlier R, Roby-Brami A, Bussel B, Cazalis
F, Pierot L, Burnod Y, Maier MA. 2002 Longitudinal
study of motor recovery after stroke: recruitment
and focusing of brain activation. Stroke 33, 1610–
1617. (doi:10.1161/01.STR.0000017100.68294.52)
Cao Y, D’Olhaberriague L, Vikingstad EM, Levine SR,
Welch KM. 1998 Pilot study of functional MRI to
assess cerebral activation of motor function after
poststroke hemiparesis. Stroke 29, 112–122.
(doi:10.1161/01.STR.29.1.112)
Buma FE, Lindeman E, Ramsey NF, Kwakkel G. 2010
Functional neuroimaging studies of early upper
limb recovery after stroke: a systematic review of the
literature. Neurorehabil. Neural Repair 24, 589 –608.
(doi:10.1177/1545968310364058)
Astrakas LG, Naqvi SHA, Kateb B, Tzika AA. 2012
Functional MRI using robotic MRI compatible
devices for monitoring rehabilitation from chronic
stroke in the molecular medicine era (review).
Int. J. Mol. Med. 29, 963– 973. (doi:10.3892/ijmm.
2012.942)
Weber R, Ramos-Cabrer P, Justicia C, Wiedermann
D, Strecker C, Sprenger C, Hoehn M. 2008 Early
prediction of functional recovery after experimental
stroke: functional magnetic resonance imaging,
electrophysiology, and behavioral testing in rats.
J. Neurosci. 28, 1022–1029. (doi:10.1523/
JNEUROSCI.4147-07.2008)
10
Phil. Trans. R. Soc. B 371: 20150353
11.
22. Lum PS, Burgar CG, Shor PC, Majmundar M, van der
Loos M. 2002 Robot-assisted movement training
compared with conventional therapy techniques for
the rehabilitation of upper-limb motor function
after stroke. Arch. Phys. Med. Rehabil. 83, 952–959.
(doi:10.1053/apmr.2001.33101)
23. Fasoli SE, Krebs HI, Stein J, Frontera WR, Hogan N.
2003 Effects of robotic therapy on motor
impairment and recovery in chronic stroke. Arch.
Phys. Med. Rehabil. 84, 477– 482. (doi:10.1682/
JRRD.2010.10.0210)
24. Fasoli SE, Krebs HI, Stein J, Frontera WR, Hughes R,
Hogan N. 2004 Robotic therapy for chronic motor
impairments after stroke: follow-up results. Arch.
Phys. Med. Rehabil. 85, 1106 –1111. (doi:10.1016/j.
apmr.2003.11.028)
25. Taub E, Uswatte G, King DK, Morris D, Crago JE,
Chatterjee A. 2006 A placebo-controlled trial of
constraint-induced movement therapy for upper
extremity after stroke. Stroke 37, 1045– 1049.
(doi:10.1161/01.STR.0000206463.66461.97)
26. Lai SM, Studenski S, Duncan PW, Perera S. 2002
Persisting consequences of stroke measured by the
stroke impact scale. Stroke 33, 1840– 1844. (doi:10.
1161/01.STR.0000019289.15440.F2)
27. Duncan PW, Lai SM, Keighley J. 2000 Defining poststroke recovery: implications for design and
interpretation of drug trials. Neuropharmacology 39,
835 –841. (doi:10.1016/S0028-3908(00)00003-4)
28. Krishnamurthi RV et al. 2013 Global and regional
burden of first-ever ischaemic and haemorrhagic
stroke during 1990–2010: findings from the Global
Burden of Disease Study 2010. Lancet Glob. Health.
1, 259–281. (doi:10.1016/S2214-109X(13)70089-5)
29. Salter K et al. 2013 Chapter 21: Outcome measures
in stroke rehabilitation 2013. In Evidence-based
review of stroke rehabilitation, pp. 1–144. Ontario,
Canada: Heart and Stroke Foundation Canadian
Partnership for Stroke Recovery.
30. Donovan NJ, Kendall DL, Heaton SC, Kwon S, Velozo
CA, Duncan PW. 2008 Conceptualizing functional
cognition in stroke. Neurorehabil. Neural Repair 22,
122 –135. (doi:10.1177/1545968307306239)
31. Blicher JU, Stagg CJ, O’Shea J, Østergaard L,
MacIntosh BJ, Johansen-Berg H, Jezzard P, Donahue
MJ. 2012 Visualization of altered neurovascular
coupling in chronic stroke patients using
multimodal functional MRI. J. Cereb. Blood Flow
Metab. 32, 2044–2054. (doi:10.1038/jcbfm.
2012.105)
32. Newton J, Sunderland A, Butterworth SE, Peters
AM, Peck KK, Gowland PA. 2002 A pilot study of
event-related functional magnetic resonance
imaging of monitored wrist movements in patients
with partial recovery. Stroke 33, 2881–2887.
(doi:10.1161/01.STR.0000042660.38883.56)
33. Pineiro R, Pendlebury S, Johansen-Berg H,
Matthews PM. 2002 Altered hemodynamic
responses in patients after subcortical stroke
measured by functional MRI. Stroke 33, 103–109.
(doi:10.1161/hs0102.100482)
34. Rossini C et al. 2004 Does cerebrovascular disease
affect the coupling between neuronal activity and
rstb.royalsocietypublishing.org
10.
hyper-perfusion and angiogenesis follow subacute
hypoperfusion in the thalamus of rats with focal
cerebral ischemia. J. Cereb. Blood Flow Metab. 31,
1119–1132. (doi:10.1038/jcbfm.2010.202)
Lin CY, Chang C, Cheung WM, Lin MH, Chen JJ, Hsu
CY, Chen JH, Lin TN. 2008 Dynamic changes in
vascular permeability, cerebral blood volume,
vascular density, and size after transient focal
cerebral ischemia in rats: evaluation with contrastenhanced magnetic resonance imaging. J. Cereb.
Blood Flow Metab. 28, 1491– 1501. (doi:10.1038/
jcbfm.2008.42)
Dijkhuizen RM, Singhal AB, Mandeville JB, Wu O,
Halpern EF, Finklestein SP, Rosen BR, Lo EH. 2003
Correlation between brain reorganization, ischemic
damage, and neurologic status after transient focal
cerebral ischemia in rats: a functional magnetic
resonance imaging study. J. Neurosci. 23, 510–517.
Lin TN, Sun SW, Cheng WM, Li F, Chang C. 2002
Dynamic changes in cerebral blood flow and
angiogenesis after transient focal cerebral ischemia
in rats: evaluation with serial magnetic resonance
imaging. Stroke 33, 2985 –2991. (doi:10.1161/01.
STR.0000037675.97888.9D)
Carmichael ST, Wei L, Rovainen CM, Woolsey TA.
2001 New patterns of intracortical projections after
focal cortical stroke. Neurobiol. Dis. 8, 910– 922.
(doi:10.1006/nbdi.2001.0425)
Stroemer RP, Kent TA, Hulsebosch CE. 1992 Increase
in synaptophysin immunoreactivity following
cortical infarction. Neurosci. Lett. 147, 21 –24.
(doi:10.1016/0304-3940(92)90765-Y)
Stroemer RP, Kent TA, Hulsebosch CE. 1993 Acute
increase in expression of growth associated protein
GAP-43 following cortical ischemia in rat. Neurosci.
Lett. 162, 51 –54. (doi:10.1016/0304-3940(93)905
57-2)
Stroemer RP, Kent TA, Hulsebosch CE. 1998
Neocortical neural sprouting, synaptogenesis and
behavioral recovery following neocortical infarction
in rats. Stroke 29, 2381 –2393. (doi:10.1161/01.STR.
29.11.2381)
Logothetis NK. 2008 What we can do and what we
cannot do with fMRI. Nature 453, 869 –878.
(doi:10.1038/nature06976)
Harsany M, Tsivgoulis G, Alexandrov AV. 2014
Intravenous thrombolysis in acute ischemic stroke:
standard and potential future applications. Expert
Rev. Neurother. 14, 879 –892. (doi:10.1586/
14737175.2014.934676)
Taub E, Miller NE, Novack TA, Cook III EW, Fleming
WC, Nepomuceno CS, Connell JS, Crago JE. 1993
Technique to improve chronic motor deficit after
stroke. Arch. Phys. Med. Rehabil. 74, 347–354.
Miltner WH, Bauder H, Sommer M, Dettmers C,
Taub E. 1999 Effects of constraint-induced
movement therapy on patients with chronic motor
deficits after stroke: a replication. Stroke 30,
586–592. (doi:10.1161/01.STR.30.3.586)
Liepert J, Bauder H, Wolfgang HR, Miltner WH,
Taub E, Weiller C. 2000 Treatment-induced cortical
reorganization after stroke in humans. Stroke 31,
1210–1216. (doi:10.1161/01.STR.31.6.1210)
Downloaded from http://rstb.royalsocietypublishing.org/ on October 26, 2016
68. Simon AB, Buxton RB. 2015 Understanding the
dynamic relationship between cerebral blood flow
and the BOLD signal: implications for quantitative
functional fMRI. NeuroImage 116, 158–161.
(doi:10.1016/j.neuroimage.2015.03.080)
69. Chen JJ, Pike GB. 2010 Global cerebral oxidative
metabolism during hypercapnia and hypocapnia in
humans: implications for BOLD fMRI. J. Cereb. Blood
Flow Metab. 6, 1094 –1099. (doi:10.1038/jcbfm.
2010.42)
70. Peng SL, Ravi H, Sheng M, Thomas BP, Lu H.
In press. Searching for a truly ‘iso-metabolic’ gas
challenge in physiological fMRI. J. Cereb. Blood
Flow Metabol. (doi:10.1177/0271678X16638103)
71. Zhe-Sun P, Zhou J, Sun W, HuangJ, van Zijl P. 2007
Detection of the ischemic penumbra using pH
weighted fMRI. J. Cereb. Blood Flow Metabol. 27,
1129– 1136. (doi:10.1038/sj.jcbfm.9600424)
72. Zhe-Sun P, Benner T, Copen WA, Sorensen G. 2010
Early experience of translating pH-weighted MRI
to image human subject at 3 Tesla. Stroke 41,
S147–S155. (doi:10.1161/STROKEAHA.110.595777)
73. Harston G et al. 2015 Identifying the ischaemic
penumbra using pH-weighted magnetic resonance
imaging. Brain 138, 36 –42. (doi:10.1093/brain/
awu374)
74. Shen Q, Huang S, Duong TQ. 2015 Ultra-high spatial
resolution basal and evoked cerebral blood flow MRI
of the rat brain. Brain Res. 1599, 126 –136. (doi:10.
1016/j.brainres.2014.12.049)
75. Murphy TH, Corbett D. 2009 Plasticity during stroke
recovery: from synapse to behaviour. Nat. Rev.
Neurosci. 10, 861–872. (doi:10.1038/nrn2735)
76. Buga AM, Di Napoli M, Popa-Wagner A. 2013
Preclinical models of stroke in aged animals with or
without comorbidities: role of neuroinflammation.
Biogerontology 14, 651 –662. (doi:10.1007/s10522013-9465-0)
77. Milidonis X, Marshall I, Macleod MR, Sena ES. 2015
Magnetic resonance imaging in experimental stroke
and comparison with histology: systematic review
and meta-analysis. Stroke 46, 843– 851. (doi:10.
1161/STROKEAHA.114.007560)
78. Woodruff TM, Thundyil J, Tang SC, Sobey CG,
Taylor SM, Arumugam TV. 2011 Pathophysiology,
treatment, and animal and cellular models of
human ischemic stroke. Mol. Degeneration 6, 11.
(doi:10.1186/1750-1326-6-11)
79. Da Fonseca ACC, Matias D, Garcia C, Amaral R,
Henrique Geraldo L, Freitas C, Lima FRS. 2014
The impact of microglial activation on blood-brain
barrier in brain diseases. Front. Cell. Neurosci. 8,
362. (doi:10.3389/fncel.2014.00362)
11
Phil. Trans. R. Soc. B 371: 20150353
57. Candelario-Jalil E, Pinheiro de Oliveira AC, Gräf S,
Bhatia HS, Hüll M, Muñoz E, Fiebich BL. 2007
Resveratrol potently reduces prostaglandin E2
production and free radical formation in
lipopolysaccharide-activated primary rat microglia.
J. Neuroinflammation 4, 25. (doi:10.1186/17422094-4-25)
58. Niwa K, Haensel C, Ross ME, Iadecola C. 2001
Cyclooxygenase-1 participates in selected
vasodilator responses of the cerebral circulation.
Circ. Res. 88, 600–608. (doi:10.1161/01.RES.88.
6.600)
59. Womelsdorf T, Valiante TA, Sahin NT, Miller KJ,
Tiesinga P. 2014 Dynamic circuit motifs underlying
rhythmic gain control, gating and integration. Nat.
Neurosci. 17, 1031 –1039. (doi:10.1038/nn.3764)
60. Ueki M, Linn F, Hossmann KA. 1988 Functional
activation of cerebral blood flow and metabolism
before and after global ischemia of rat brain.
J. Cereb. Blood Flow Metab. 8, 486 –494. (doi:10.
1038/jcbfm.1988.89)
61. Kimberley TJ, Khandekar G, Borig M. 2008 fMRI
reliability in subjects with stroke. Exp. Brain Res.
186, 183 –190. (doi:10.1007/s00221-007-1221-8)
62. Chung S, You J, Tack G, Yi J, Lee B. 2005 The mean
intensity of activation than the number of activated
voxels is a better index to reflect the levels of
visuospatial performance. Conf. Proc. IEEE Eng. Med.
Biol. Soc. 5, 5298–5301. (doi:10.1109/iembs.2005.
1615676)
63. Chung SC, Sohn JH, Lee B, Tack GR, Yi JH, You JH,
Kwon JH, Kim HJ, Lee SY. 2007 A comparison of the
mean signal change method and the voxel count
method to evaluate the sensitivity of individual
variability in visuospatial performance. Neurosci.
Lett. 418, 138 –142. (doi:10.1016/j.neulet.2007.
03.014)
64. Carey JR, Kimberley TJ, Lewis SM, Auerbach E,
Dorsey L, Rundquist P, Ugurbil K. 2002 Analysis of
fMRI and finger tracking training in subjects with
chronic stroke. Brain 125, 773– 788. (doi:10.1093/
brain/awf091)
65. Kimberley TJ, Lewis SM. 2007 Understanding
neuroimaging. Phys. Ther. 87, 670– 683. (doi:10.
2522/ptj.20060149)
66. Hoge RD. 2012 Calibrated fMRI. NeuroImage 62,
930 –937. (doi:10.1016/j.neuroimage.2012.02.022)
67. Shu CY, Herman P, Coman D, Sanganahalli BG,
Wang H, Juchem W, Rothman DL, de Graaf RA,
Hyder F. 2016 Brain region and activity-dependent
properties of M for calibrated fMRI. NeuroImage
125, 846 –856. (doi:10.1016/j.neuroimage.2015.
10.083)
rstb.royalsocietypublishing.org
47. Kirilina E, Lutti A, Poser BA, Blankenburg F, Weiskopf
N. 2015 The quest for the best: the impact of different
EPI sequences on the sensitivity of random effect fMRI
group analyses. NeuroImage 126, 49–59. (doi:10.
1016/j.neuroimage.2015.10.071)
48. DeFelipe L, Alonso-Nanclares L, Arellano JL. 2002
Microstructure of the neocortex: comparative
aspects. J. Neurocytol. 31, 299 –316. (doi:10.1023/
A:1024130211265)
49. Nedergaard M, Ransom B, Goldman SA. 2003
New roles for astrocytes: redefining the functional
architecture of the brain. Trends Neurosci. 26,
523–530. (doi:10.1016/j.tins.2003.08.008)
50. Hyder F, Rothman DL, Bennett MR. 2013 Cortical
energy demands of signaling and nonsignaling
components in brain are conserved across
mammalian species and activity levels. Proc. Natl
Acad. Sci. USA 110, 3549 –3554. (doi:10.1073/pnas.
1214912110)
51. Shih YY, Huang S, Chen YY, Lai HY, Kao YC, Du F,
Hui ES, Duong TQ. 2014 Imaging neurovascular
function and functional recovery after stroke in the
rat striatum using forepaw stimulation. J. Cereb.
Blood Flow Metab. 34, 1483– 1492. (doi:10.1038/
jcbfm.2014.103)
52. Montoya CP, Campbell-Hope LJ, Pemberton KD,
Dunnett SB. 1991 The ‘staircase test’: a measure of
independent forelimb reaching and grasping
abilities in rats. J. Neurosci. Methods 36, 219–228.
(doi:10.1016/0165-0270(91)90048-5)
53. Perrone MG, Scilimati A, Simone L, Vitale P. 2010
Selective COX-1 inhibition: a therapeutic target to
be reconsidered. Curr. Med. Chem. 17, 3769 –3805.
(doi:10.2174/092986710793205408)
54. Schwab JM, Nguyen TD, Postler E, Meyermann R,
Schluesener HJ. 2000 Selective accumulation of
cyclooxygenase-1-expressing microglial cells/
macrophages in lesions of human focal cerebral
ischemia. Acta Neuropathol. 99, 609–614. (doi:10.
1007/s004010051170)
55. Choi SH, Aid S, Bosetti F. 2009 The distinct roles of
cyclooxygenase-1 and -2 in neuroinflammation:
implications for translational research. Trends
Pharmacol. Sci. 30, 174 –181. (doi:10.1016/j.tips.
2009.01.002)
56. Candelario-Jalil E, González-Falcón A, Garcı́a-Cabrera
M, Álvarez D, Al-Dalain S, Martı́nez-Sánchez G, León
OS, Sprinter JE. 2003 Assessment of the relative
contribution of COX-1 and COX-2 isoforms to
ischemia-induced oxidative damage and
eurodegeneration following transient global cerebral
ischemia. J. Neurochem. 86, 545 –555. (doi:10.
1046/j.1471-4159.2003.01812.x)