Functional magnetic resonance imaging study on
dysphagia after unilateral hemispheric stroke: a
S Li,1,2C Luo,3B Yu,4B Yan,1Q Gong,5,6C He,7L He,1X Huang,5D Yao,3S Lui,5
H Tang,5Q Chen,1Y Zeng,1D Zhou1
1Department of Neurology,
West China Hospital of Sichuan
University, Chengdu, Sichuan,
PR China;2Department of
Rehabilitation Medicine, 2nd
Affiliated Hospital of Wenzhou
Medical College, Wenzhou,
Zhejiang, PR China;3School of
Science and Technology,
University of Electronic Science
and Technology of China,
Chengdu, Sichuan, PR China;
4Department of Neonatology,
2nd Affiliated Hospital of
Wenzhou Medical College,
Wenzhou, Zhejiang, PR China;
5Huaxi MR Research Centre
(HMRRC), Department of
Radiology, West China Hospital
of Sichuan University, PR China;
6Division of Medical Imaging,
Faculty of Medicine, University
of Liverpool, Liverpool, UK;
7Department of Rehabilitation
Medicine, West China Hospital
of Sichuan University, Chengdu,
Sichuan, PR China
Professor D Zhou, Department of
Neurology, West China Hospital
of Sichuan University, No 37
Chengdu, Sichuan, PR China;
Received 23 February 2009
Revised 25 May 2009
Accepted 27 May 2009
Published Online First
21 June 2009
Background: Swallowing dysfunction is common and
disabling after acute stroke; however, the mechanism of
dysphagia or recovery of swallowing from dysphagia
remains uncertain. The purpose of this study was to
explore cerebral activation of swallowing in dysphagia
using functional MRI (fMRI) to compare the functional
anatomy of swallowing in unilateral hemispheric stroke
patients and healthy adults.
Methods: In total, five left hemispheric stroke patients
with dysphagia, five right hemispheric stroke patients
with dysphagia and 10 healthy controls were examined
with event related fMRI while laryngeal swallow related
movements were recorded. Data were processed using
the general linear model.
Results: A multifocal cerebral representation of swal-
lowing was identified predominantly in the left hemi-
sphere, in a bilateral and asymmetrical manner. Cerebral
activation during swallowing tasks was localised to the
precentral, postcentral and anterior cingulate gyri, insula
and thalamus in all groups. Activation of volitional
swallowing in dysphagic unilateral hemispheric stroke
patients might require reorganisation of the dominant
hemispheric motor cortex, or a compensatory shift in
activation to unaffected areas of the hemisphere.
Conclusions: The results indicate that unilateral stroke of
either cerebral hemisphere can produce dysphagia.
Effective recovery is associated with cerebral activation
related to cortical swallowing representation in the
compensating or recruited areas of the intact hemisphere.
Functional MRI is a useful method for exploring the spatial
localisation of changes in neuronal activity during tasks
that may be related to recovery. Therefore, the
subsequent information gleaned from changes in neural
plasticity could be useful for assessing the prognosis of
Dysphagia is one of the most vexing clinical
problems encountered in stroke patients and is
associated with aspiration pneumonia. Depending
on the definition and the diagnostic tools used,
dysphagia occurs in approximately 25–50% of
stroke patients.1–3It is now well established that
the cerebral cortex plays an important functional
role in the regulation of swallowing.4While the
reflexive component of swallowing depends on
swallowing centres in the brainstem, the initiation
of swallowing is a voluntary action that relies on
the integrity of motor areas of the cerebral cortex.5
Indeed, according to a recent review of the
neurophysiology of swallowing, damage to the
cerebral cortex can have a significant effect on the
peripheral swallowing mechanism operating at the
level of the brainstem.6Several clinical studies have
reported that damage to the relevant hemisphere in
stroke patients results in increased difficulty in
swallowing, as well as corticospinal output that
appears to be associated with swallowing prob-
lems.7–10Malnutrition predicts a poor functional
outcome and increased mortality.11Although most
recovery of normal swallowing function occurs
within a few weeks after stroke,12the neural
mechanisms involved are unclear. In many cases,
improvement in swallowing is not matched by an
improvement in the degree of hemiparesis.9
Within the last few years, an increasing number
of brain imaging studies in healthy adults have
implicated multiple cortical regions in the control
of swallowing.13–21However, predictions of the
functional contributions made by each of these
brain areas have not been fully tested. Although it
is now clearly recognised that unilateral stroke of
either cerebral hemisphere can produce dysphagia,
it is unclear whether one hemisphere is dominant
for swallowing and whether damage to a specific
hemisphere affects the recovery of swallowing.22–25
Functional brain imaging studies may help to
elucidate the relevant neural mechanisms under-
lying swallowing impairment and also provide
evidence for functional changes to the cortex
resulting from a stroke. Regarding the latter, the
functional state of the cerebral cortex of stroke
patients with dysphagia has thus far not been fully
investigated, particularly using functional MRI
The purpose of this study was to investigate the
neurorehabilitation mechanisms responsible for
cerebral activation of swallowing following dys-
phagia. To this end, we examined the functional
anatomy of unilateral hemispheric stroke patients
and compared it with that of healthy adults. Using
blood oxygenation level dependent (BOLD) fMRI,
we studied functional neuroimaging features dur-
ing voluntary swallowing tasks in dysphagic
patients, at least 72 h after stroke onset. We
focused on both cortical and subcortical activation,
so as to gain a greater understanding of the
relevant processes that, in turn, might lead to
new and effective treatment strategies for dyspha-
gic stroke patients.
The study protocol was approved by the institu-
tional ethics committee at Sichuan University, and
1320J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.176214
written informed consent was obtained from all subjects prior
to commencing the research. Patients consecutively admitted to
the Stroke Centre of West China Hospital from October 2007 to
October 2008 were included in the study, based on the criteria
that they had only experienced a single cerebral hemispheric
ischaemic stroke that was accompanied by a history of mild to
severe dysphagia, lasting for up to 3 days. Of these patients,
there were five females and five males, with a mean age of 70.9
(3.4) years (age range 62–78). During the initial consultation,
we obtained the following information from each patient: sex,
age at which the stroke occurred, time since stroke, type of
stroke, location and side of the lesion (left or right hemisphere).
The 10 stroke patients were enrolled and diagnosed with their
first onset of ischaemic stroke, as defined according to World
Health Organization criteria (the sudden onset of neurological
deficit persisting for .24 h).26All patients were assessed for
swallowing dysfunction within 24 h of stroke onset by a
qualified speech and language therapist. The therapist used the
Logemann clinical indicators of dysphagia27(ie, coughing, oral
(observed by placing one finger on the hyoid and one on the
thyroid), throat clearing and choking).28To confirm the
diagnosis, patients were subjected to videofluoroscopic swal-
lowing examination (VFSS Imager, Model IA-12LD/HG12;
Shimadzu Corporation, Kyoto, Japan). Those who displayed
at least one of the following symptoms were considered to have
dysphagia29: (1) food residue occupying more than 50% of the
vallecula or piriform sinus space after swallowing; (2) subglottic
aspiration; (3) a pharyngeal transit time .2 s (operationally
defined as the time taken for the bolus to move from the point
at which the pharyngeal swallow is triggered to the cricophar-
yngeal sphincter); and (4) impaired cricopharyngeal muscle
relaxation. Exclusion criteria included: (1) prior cerebrovascular
disease; (2) pre-existing neurological or psychiatric disorders
(including a history of seizures, global cognitive impairment,
aphasia, neglect, substantial
depression or claustrophobia); (3) use of an electrically sensitive
biomedical device (eg, cardiac pacemaker or cochlear implant);
(4) metal clips in the brain; or (5) pneumonia at the time of
An age and sex matched group of healthy older volunteers
(n=10; five females; mean age 70.3 (4.2) years; range 65–75)
served as the control group. All control subjects had a normal
neurological examination, no history of a stroke and no
significant active neurological problems. They were also free
of systemic diseases and neurological disorders. The same
exclusion criteria listed above were applied to the control group.
All subjects were strongly right handed, according to the
Edinburgh Handedness Inventory. The study adhered to the
MRI safety depositional guidelines established by the US Food
and Drug Administration for clinical scanners. Before scanning,
each subject was trained to perform the voluntary swallowing
task according to instructions.
Swallowing task paradigm
Voluntary saliva swallowing task
Each participant was subjected to functional imaging runs of
3 min 20 s duration, during each event related experimental
session. Each functional run consisted of two randomly
ordered tasks performed in response to visual cues. The visual
cues were back projected onto a mirror positioned above the
subject’s eyes. For the activation task, the ‘‘green light’’
condition was a single voluntary saliva swallow performed in
response to the visual cue ‘‘do swallow’’. The subject was
instructed to swallow his/her saliva once, without making
exaggerated oral movements to produce extra saliva. For the
non-activation task, deemed the ‘‘red light’’ condition, no
overt response was required of the subject following presenta-
tion of the visual cue ‘‘don’t swallow’’. The frequency of the
visual cue was pseudo-randomly displayed during the scanning
time. Between the activation task cues, the resting task cue
was displayed. Thus with a mean random interstimulus
interval of 10 s (range 8–14), each activation task was
performed 15 times during each functional run. To ensure
that the subject understood the experimental procedures, each
subject practiced the activation task prior to participating in
the functional runs.
Identification of swallowing
To verify that the subjects swallowed during the activation
periods and remained motionless during rest periods, surface
electromyography (EMG) was performed, using a pair of bipolar
Ag/AgCl electrodes, on the submental and infrahyoid muscle
The EMG was recorded using a 10/20 system with two Ag/
AgCl electrodes soldered to 12 kV current limiting resistors that
were positioned comfortably over the subject’s thyroid cartilage.
The EMG device was a Mizar 40 amplifier (EBNeuro, Florence,
Italy), with two channels adapted for magnetic resonance. The
sampling rate was set at 4096 Hz, which allowed enough time
for resolving the switching effect from the readout gradient
under the high slew rate condition. The EMG dynamic range
was ¡65.5 mV, to prevent MRI artefact waveforms saturating
the EMG. The EMG device, placed inside a shielded box,
amplified the signal and performed analogue to digital conver-
sion. The amplifier was connected to a recording computer
located outside the scanner room via a fibreoptic cable. The MR
artefact was filtered out using the BE-MRI Toolbox software
(Galileo New Technology, Florence, Italy). The time at which
the swallow related laryngeal elevation began was recorded for
all individual swallows and all fMRI scans.
MRI data were acquired using a 3.0 T MRI system (Excite;
General Electric Company, Milwaukee, USA) with a standard
8 channel, phased array, head coil at the Department of
Radiology in West China Hospital. EMG data were acquired
simultaneously. Structural images were acquired in axial
orientation using a three-dimensional spoiled gradient recalled
sequence (repetition/echo time (TR/TE)=8.5 ms/3.4 ms; flip
angle=12u), with a voxel size of 0.9460.9461.00 mm3. The
MR images that were sensitised to changes in BOLD signal
levels (TR/TE=2000 ms/30 ms;
obtained by a gradient echo, echo planar imaging sequence.
Slice thickness was 5 mm (no slice gap) with a matrix size of
64664 and a field of view of 2406240 mm2, which resulted in
a voxel size of 3.7563.7565.00 mm3. Each brain volume
consisted of 30 axial slices and each functional run contained
100 image volumes.
flip angle=90u) were
fMRI data analysis
Image preprocessing and statistical analyses were performed
using statistical parametric mapping software (SPM2, Welcome
Department of Imaging Neuroscience, http://www.fil.ion.ucl.
ac.uk). For each subject, all echo planar images were corrected
according to slice time, realigned with the first image of the first
J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.1762141321
series and then unwarped to correct for susceptibility by
movement interaction. If a patient’s motion and rotation
parameters exceeded 0.5 mm and 0.5u, respectively, this run of
data was excluded from future analyses. There was no
significant difference in the magnitude of the motion correction
parameters between the control and study groups.
To allow for magnetisation equilibrium, the first five images
were discarded. The remaining 95 images were first corrected
to account for the delay in acquisition time among different
slices, after which the images were realigned the first volume
for head motion correction. The images were subsequently
spatially normalised to the MNI (Montreal Neurological
Institute) template brain, and spatially smoothed with a
three-dimensional Gaussian kernel of 8 mm full width at half
maximum.The volumes were
The fMRI data were first preprocessed according to the steps
mentioned above. Based on the EMG results, the time point at
which swallowing activity coincided with the first stimulation–
time signal was acquired. Then, the canonical haemodynamic
response function was modelled by two gamma variant
functions convolved with the stimulation–time pulse signal.
Finally, the canonical haemodynamic response function was
specified as an interested regressor in the SPM design matrix.
Motion correction parameters for each run were included in the
design matrix of six regressive parameters as covariates with no
interest. The data were then modelled based on voxels, using a
general linear model.30
The statistical threshold was set at p,0.001 at the cluster
level (contiguous voxels .10; lowest threshold; t value .4.29
for controls; t value .7.17 for patients). For patients and
healthy controls, contrast images were created for each subject
and entered separately into a voxel based one sample Student’s t
test (df=4 for patients; df=9 for controls), to analyse random
effects. In addition, the two sample t test was used to compare
results (1) between each unilateral hemispheric stroke patient
group and the control group; and (2) between all stroke patients
and controls, by reversing the maps from the right hemisphere
stroke group and combining them with the maps from the left
hemisphere stroke group. Each cluster showing an uncorrected
p,0.001 for its spatial extent was considered statistically
significant (contiguous voxels .10; lowest threshold; t value
.3.85 for left or right hemispheric stroke patients with
dysphagia vs controls; t value .3.61 for 10 stroke patients with
dysphagia vs controls).
Hemispheric dominance was quantified using a laterality
index (LI), defined by the ratio:
LI = [Ss left 2 Ss right]/[Ss left + Ss right] 6100
where s = (percentage of activation) 6(number of activated
pixels). A positive LI indicates left hemispheric dominance
whereas a negative LI demonstrates right hemispheric dom-
inance. Ratios at or close to 0 are thought to represent an
Clinical features and dysphagia examination
The patients’ clinical features and lesions are summarised in
table 1. There was no significant difference in the distribution of
brain infarction between patients with left and right lesions.
They had each experienced their first and only cerebral
hemispheric stroke up to 3 days prior to admission. The time
from stroke onset to the start of the fMRI study ranged from 3
to 5 days. Based on the results of the initial VFSS, four patients
were diagnosed with both oral and pharyngeal dysfunction,
three with only oral dysfunction and three with only
pharyngeal dysfunction. Of these patients, two presented with
mild initial Penetration Aspiration (P-A) Scale scores32(score of
2), six presented with moderate P-A Scale scores (score of 3–5)
and two presented with severe P-A Scale scores (score of 6),
indicating at least one aspiration episode.
According to the P-A Scale scores, there was no significant
difference in the degree of severity of dysphagia between left
and right hemispheric stroke patients (p.0.05).
Clinical data of dysphagic stroke study patients
(days) Dysphagia type*
1F 78Left Next to central
part of lateral
18/12/123.63103 Oral dysfunctionMild2
Oral and pharyngeal
Oral and pharyngeal
Oral and pharyngeal
10F 70RightCorona radiate10/7/9 2.2753 Mild2
*Features of oral dysfunction: anterior bolus loss, tongue pumping, delayed initiation of movement and uncoordinated initiation of oral transfer. Characteristics of pharyngeal
dysfunction: delayed pharyngeal swallow, reduced laryngeal elevation, penetration, aspiration and stasis.
Dysphagia Severity Rating Scale34: mild, oropharyngeal dysphagia present but can be managed with specific swallowing suggestions; moderate, significant potential for aspiration
exists. Trace aspiration of one or more consistencies may be seen under videofluoroscopy; severe, more than 10% aspiration for all consistencies.
A-P, anterior–posterior; L-R, left–right; NIHSS Score, National Institutes of Health Stroke Scale33; P-A Scale, Penetration Aspiration Scale; S-I, superior–inferior.
1322J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.176214
Performance of functional MRI tasks
Verification of swallowing
All subjects tolerated the scanning procedure well, without
displaying excess body movements, and all succeeded in
swallowing at a consistent rate during fMRI. Head movements
were restricted to avoid contamination of the signal by motion
artefacts. No gross head motions occurred during data collection
in this investigation. The EMG recordings allowed identification
of both the beginning and end of task specific muscle activity
for each individual swallowing event in all participants (fig 1).
The laryngeal movement of the EMG indicated that all subjects
swallowed once in response to each of the ‘‘do swallow’’ cues.
Furthermore, there were no instances of swallowing in response
to ‘‘don’t swallow’’ cues in any of the subjects.
According to the EMG results, there was no significant
difference in the swallowing response latencies between
patients and controls responding to the ‘‘do swallow’’ cues
(one sample Student t tests: p.0.05).
Group analysis for controls
The MNI coordinates of our results were transformed to
Talairach coordinates using mni2tal.m (http://imaging.mrc-cbu.
cam.ac.uk/downloads/MNI2tal/) The location of activated areas
in each subject was assigned to Brodmann areas based on
established neuroanatomical landmarks.35 36
Table 2 shows the distribution of activation in cortical and
subcortical sites (designated in Brodmann areas, BA) during the
volitional swallowing paradigm. The results showed a signifi-
cant increase (p,0.001, uncorrected) in fMRI signal intensity in
the primary motor cortex (BA 4), primary somatosensory cortex
(BA 3 and 2), supplementary motor cortex (BA 6), middle
frontal cortex (BA 10), transverse temporal gyrus, superior
temporal gyrus and middle temporal gyrus (BA 42, 22, 38, and
21), anterior cingulate gyrus and the insula (BA 32 and 13), as
well as in the areas of the putamen and thalamus.
The voluntary saliva swallow evoked significant activation
(p,0.001, uncorrected) in a number of discrete brain regions
(fig 2A). The total volume of the activated brain region in the
group map was 11 718 mm3(p,0.001, uncorrected), and the
spatial patterns of activation within the left and right hemi-
spheres were similar, but clearly asymmetric. The largest
activation focus was located within the left pericentral cortex,
corresponding to the primary motor (MI) and primary
somatosensory (SI) associated cortices. Group analysis of the
LI value for the 10 control subjects during performance of the
voluntary saliva swallowing task produced a positive value,
indicative of left hemispheric dominance (LI=21).
Group analysis for left hemispheric stroke patients
During the voluntary saliva swallowing task, there was
activation of the stronger right side, which included the primary
motor cortex (MI, the precentral gyrus, BA 4), the primary
somatosensory cortex (SI, postcentral gyrus, BA 3), the superior
and middle temporal gyri (BA 22 and 21), insula and thalamus.
The right primary motor and primary somatosensory cortices,
together with the medial frontal gyrus, and the superior
temporal and middle temporal gyri, underwent stronger
activation than in the left hemisphere (fig 2B). The total
volume of the sensorimotor (SM) cortex activated was
11.556 mm3. This was 1.3 times the mean value of the control
Group analysis for right hemispheric stroke patients
Prominent activation foci corresponded to the SM cortex,
middle frontal (BA 10), superior temporal and middle temporal
gyri (BA 22 and 21) and the insula (BA 13). However, activation
within the undamaged hemisphere was stronger than the
infarcted hemisphere (fig 2C). The regions of maximal activa-
tion were observed in the bilateral SM cortex, anterior cingulate,
insula and right anterior brain regions, including the medial and
inferior frontal gyri. Within the left hemisphere, activation was
only observed in the thalamus.
Intergroup analysis for each stroke patient group and control group
When a comparison was made with the control group, the left
hemispheric stroke patients showed greater activation bilater-
ally in the precentral gyrus, insula and cingulate gyrus (fig 3A).
The only structures in the left hemisphere of this group of
stroke patients to be activated more strongly than those in the
control group were the postcentral gyrus and middle frontal
gyrus. When the right hemispheric stroke patients were
compared with the control subjects, the former showed greater
activation of the precentral and postcentral gyri, superior
temporal gyrus and cingulate gyrus in both cerebral hemi-
spheres. Additional activity was also observed in the right
middle frontal gyrus and left insula (fig 3B). By reversing the
maps of the right hemispheric stroke group, and then combining
them with those of the left hemispheric stroke group, all stroke
patients showed a multifocal, large cluster of increased
activation in the left precental and postcentral gyri, left middle
frontal gyrus and insula compared with the control subjects
during the task (fig 3C). Additional activity was also observed in
the bilateral cingulate gyrus. The Talairach coordinates of the
most activated voxel of these clusters are given in table 3.
In this report, BOLD fMRI was used to determine the cortical
representations of swallowing in dysphagic stroke patients, and
to compare them to those in healthy adults. To our knowledge,
this is the first study to use fMRI to investigate cerebral cortical
activation during voluntary swallowing in dysphagic stroke
patients. Our findings indicate that swallowing in dysphagic
stroke patients, similar to healthy older adults, is processed
within multiple regions of the cerebral cortex and subcortex.
The prominent swallow related activation of the precentral,
episode of swallowing in one subject. The beginning and end of task
specific muscle activity were identified. The visual cue ‘‘do swallow’’
was presented at time=0.
EMG signals from the laryngeal movement during a typical
J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.1762141323
postcentral and anterior cingulate gyri, together with the insula,
in a bilateral and asymmetrical manner is consistent with
previous event related task paradigm studies in healthy
subjects.14–16 21 37–39Interestingly, this study also discovered the
following features that were distinct to dysphagic stroke
Firstly, group analysis of the healthy control subjects revealed
that hemispheric dominance appeared to be associated with the
left hemisphere. Our results indicate that although swallowing
involves both hemispheres, there is greater and more intense
activity in the left hemisphere. The strongest activations were
found in the sensorimotor cortices, insula and cingulate gyrus.
The concept of hemispheric dominance in swallowing derives
from clinical observation of dysphagic patients,4 23 40–43neuro-
imaging13 14 16 20 21 38
(TMS) studies44 45in healthy individuals and stroke patients.
In our study, swallowing recruited multiple cerebral regions,
often in an asymmetrical manner, and showed left hemispheric
dominance according to the LI of group analysis for controls.
The data presented here are mostly descriptive as this is the
first attempt to functionally characterise the brains of dysphagic
stroke patients using fMRI. Our findings confirm previous
swallowing tasks in Brodmann areas
Group analysis for controls and patients: distribution of activation during voluntary saliva
(mm3) t value p Value
Local maxima of cluster
Distribution of activation of group analysis for controls
Middle frontal gyrus
Superior temporal gyrus
Middle temporal gyrus3
Transverse temporal gyrus
Distribution of activation of group analysis for left hemispheric stroke patients with dysphagia
Superior temporal gyrusL
Middle temporal gyrusL
Distribution of activation of group analysis for right hemispheric stroke patients with dysphagia
Middle frontal gyrus 10
Inferior frontal gyrus
Superior temporal gyrus6
***p,0.0001. The statistical threshold was set at p,0.001 at the cluster level (10 contiguous voxels), at t value .4.29 (n=10)
for the control group and t value .7.17 for the patient group (n=5), uncorrected.
1324J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.176214
reports40–43that unilateral hemispheric lesions may produce
dysphagia in stroke patients. This study provides supporting
evidence for the bilateral redistribution of swallowing networks
after occurrence of a stroke. The key finding was that dysphagic
stroke patients who suffered ischaemic infarct in the left
hemisphere showed overactivation in their right cortical maps,
over that of the infarcted hemisphere. Conversely, dysphagic
stroke patients who suffered ischaemic infarcts in the right
hemisphere showed overactivation in their left cortical maps
and a shifting from the affected hemisphere. The results
demonstrate that swallowing may be prominently lateralised
to one specific hemisphere while recovery from dysphagia after
a stroke might require reorganisation of the motor cortex in the
dominant hemisphere. However, hemispheric dominance in
swallowing was still under debate. Our knowledge of the
mechanisms that contribute to the dysphagia in unilateral
hemispheric stroke is limited, particularly in view of the fact
that the stroke patients in this study all presented with mild to
displayed on normalised axial brain slices using the Talairach–Tournoux coordinate system. The spatial patterns of activation within the left and right
hemispheres were similar, but clearly asymmetric. The results showed a significant increase (p,0.001, uncorrected) in the primary motor cortex
(Brodmann area (BA) 4), primary somatosensory cortex (BA 3 and 2), supplementary motor cortex (BA 6), middle frontal cortex (BA 10), transverse
temporal gyrus, superior temporal gyrus and middle temporal gyrus (BA 42, 22, 38 and 21), anterior cingulate gyrus and the insula (BA 32 and 13) as
well as in the areas of the putamen and thalamus. The largest activation focus was located within the left pericentral cortex, corresponding to the
primary motor and primary somatosensory associated cortices. Colour bar represents t value (p,0.001, uncorrected). (B) Group analysis of brain
activation associated with the voluntary saliva swallow task in patients who suffered ischaemic infarct in the left hemisphere. The primary motor cortex
(BA 4), the primary somatosensory cortex (BA 3), middle temporal gyrus (BA 21), insula (BA 13), thalamus and basal ganglia were activated. The
maximal activation regions were observed in the right hemisphere. Colour bar represents t value (p,0.001, uncorrected). (C) Group analysis of brain
activation associated with the voluntary saliva swallow task in patients who suffered ischaemic infarct in the right hemisphere. Prominent activation
foci corresponded to the sensorimotor cortex, middle frontal gyrus (BA 10), superior temporal gyrus (BA 22), middle temporal gyrus (BA 21) and insula
(BA13). Extensively activated tissue volumes were observed in the left hemisphere. Colour bar represents t value (p,0.001, uncorrected).
(A) Brain activation associated with voluntary saliva swallow task for group analysis of control. Regions of significant activation are
versus control group. The clusters of increased activation were bilaterally centred on the primary motor cortex (Brodmann area (BA) 4), supplementary
motor cortex (BA 6), insula (BA 13) and cingulate gyrus (BA 24 and 32). Additional activity was also observed in the right primary somatosensory
cortex (BA 3 and 2). (B) Right hemispheric stroke patients versus control subjects. A large multifocal cluster of increased activation is seen in the
bilateral primary motor cortex (BA 4), primary somatosensory cortex (BA 3 and 2), superior temporal gyrus (BA 22) and cingulate gyrus (BA 24),
compared with control subjects. Additional activity was also observed in the right middle frontal gyrus (BA 10) and left insula (BA 13). (C) Stroke
patients with dysphagia versus controls. The increased activation was located in the left primary motor and somatosensory cortex (BA 4 and 3), left
supplementary motor cortex (BA 6) and insula (BA 13). Additional activity was also observed in the bilateral cingulate gyrus (BA 32, 24). Colour bar
represents t value (p,0.001, uncorrected).
Statistical parametric maps of intergroup comparisons between stroke patients and 10 control subjects. (A) Left hemispheric stroke patients
J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.1762141325
moderate lesions. In order to confirm these preliminary findings,
a prospective, longitudinal study is needed that contains a larger
sample of unilateral stroke patients and includes patients both
with and without dysphagia.
Secondly, we found that, compared with controls, a number
of cortical regions in dysphagic stroke patients were over-
activated by volitional swallowing where the most consistent
of these areas included the lateral pericentral and postcentral
gyri and multiple subcortical sites. Based on the theory that
the brain undergoes neurofunctional adaptation after a
stroke,46the presence of a lesion in the dominant hemisphere
would cause cortical activation to shift to the unaffected
hemisphere. Compared with control subjects, left hemispheric
stroke patients with dysphagia showed overactivation bilat-
erally in the precentral gyrus, insula and cingulate gyrus.
Moreover, right hemispheric stroke patients with dysphagia
showed increased activation compared with healthy controls,
particularly in the precentral and postcentral gyri, superior
temporal gyrus and cingulate gyrus. Interestingly, during
volitional swallowing, there was increased activation in the
left precentral and postcentral gyri, left medial frontal gyrus
and insula of all of the stroke patients, compared with control
subjects. Due to the presence of significantly different
activation modes from intergroup comparisons, it might be
postulated that the cerebral cortex is involved in the execution
of, or sensory feedback processing for, voluntary swallowing.
Our study corroborates previous reports of cortical activity in
the motor areas during swallowing. These areas control
oropharyngeal deglutitive muscle activity. Multiple sites of
overactivation, including those associated with motor proces-
sing, suggest that the motor control of swallowing may involve
several cortical sites that initiate, process and execute the
necessary output for swallowing.
Furthermore, lesion studies have demonstrated that swallow-
ing behaviours may differ between left hemispheric and right
hemispheric damage strokes. We also found that patients with
left and right hemispheric strokes may have different dysphagic
characteristics. In agreement with Robbins and Levine,23 24our
results showed that pharyngeal dysmotility was more promin-
ent in patients with right hemispheric damage strokes, and
reduced oral coordination was more prominent in patients with
left hemispheric damage. This finding was consistent with
previously published work.8
Thirdly, this is our initial effort to show the relationship
between the cerebral activation of swallowing following
unilateral ischaemic stroke and the severity of dysphagia.
These findings indicate that patients with left hemisphere
lesions may have more severe dysphagia, and that their cerebral
activation was therefore smaller during the swallowing task
than patients with right hemispheric damage. Our results
support the hypothesis that lesions to the left hemisphere may
result in more severe impairment of swallowing due to the
greater relative importance of the left hemisphere in mediating
this function. The cortical representation asymmetry may
explain why the development of dysphagia following a stroke
is variable in degree and duration. This suggests that compen-
satory contralateral hemisphere reorganisation might be respon-
sible for the improvement in swallowing seen in patients after
Cerebral regions exhibiting changes in activation during voluntary saliva swallowing
(mm3) t valuep Value
Local maxima of cluster
Intergroup comparison: increased activation (left hemispheric stroke patients with dysphagia versus controls)
Precentral gyrusL 4.6
Middle frontal gyrusL6
Cingulate gyrusL 24
Intergroup comparison: increased activation (right hemispheric stroke patients with dysphagia versus controls)
Middle frontal gyrusR 10
Superior temporal gyrusL 22
Intergroup comparison: increased activation (stroke patients with dysphagia versus controls)
Middle frontal gyrusL6
***p,0.0001. The statistical threshold was set at p,0.001 (uncorrected) at the cluster level (10 contiguous voxels), at t value
.3.85 for left or right hemispheric stroke patients with dysphagia versus controls or t value .3.61 for stroke patients with
dysphagia versus controls.
1326 J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.176214
occurrence of a unilateral stroke. Although the number of
patients studied is too small for firm conclusions to be made,
these observations may have implications for the diagnosis,
Considering the brain’s plasticity, and our observations of the
changes to these functional maps, a novel insight from our work
is that increased bilateral activation may indicate a good
prognosis for acute mild or moderate strokes. Furthermore,
improvement in the dysphagia of stroke patients may be
associated with a compensatory recruitment and bilateral
activation of areas that are involved in the cortical representa-
tion of swallowing.
The present study has several limitations. For dysphagic
stroke patients, it may be difficult to complete the water bolus
swallowing task in the supine position. Taking the safety of
patients and older controls into consideration, we only used the
dry swallowing task in this pilot study. Furthermore, these
experimental tests do not accurately represent ‘‘normal’’
swallowing conditions. Subjects were required to swallow
repeatedly while lying supine, for several seconds. Due to the
strict training before study, all subjects showed swallow related
movements, as determined by EMG. Although possible con-
founding variables such as age, lesion location and disease
duration were carefully matched between groups, we acknowl-
edge the modest sample size of this study and recognise that a
large cohort study would be required to verify the current
stroke related dysphagia.
In conclusion, the fMRI scanning results indicate that the
primary motor and somatosensory areas are consistently active
across healthy adult subjects when swallowing. The anterior
cingulate and insular cortices are also highly active during
swallowing. To further our understanding of the function of
these anatomical regions and systems in swallowing, we
demonstrated that swallowing function in dysphagic stroke
patients is associated with a compensatory recruitment and
activation of regions of the cerebral cortex in the intact
hemisphere. Given that the intact hemisphere plays an
important role in the recovery of swallowing after a stroke,
we are provided with an interesting opportunity to study the
plasticity of an intact pathway. Thus any future rehabilitation
therapies aimed at enhancing the recovery of swallowing should
target reorganisation of the intact side.
Finally, functional MRI is among the fastest growing brain
imaging technologies. It is advantageous because it is minimally
invasive, compared with some other brain imaging systems, and
it is also becoming increasingly accessible to researchers. As a
valuable method for studying swallowing control, this approach
allows assessment of the cerebral activity associated with
functional swallowing, and could serve as a useful prognos-
ticator for dysphagic stroke conditions.
Acknowledgements: The authors acknowledge support from the State Key Lab of
Biotherapy, West China Hospital/West China Medical School of Sichuan University.
We thank all of the participants for their cooperation during the period of the ‘‘5.12’’
earthquake. We are also most grateful to Wei Liao, PhD, for his collaboration and
Funding: This research was supported by the National Natural Science Foundation of
China (grant Nos 30625024, 30728017, 30525030 and 60736029), Key research
project of science and technology of MOE (107097), National Basic Research
Programme of China (973 Programme No 2007CB512305), National High Technology
Programme of China (863 Programme No 2008AA02Z408 and 2007AA02Z482), the
Programme of State Administration of Traditional Chinese Medicine of Zhejiang
Province (No 2006Y016), the Construction Project of Medical Key Subject in Zhejiang
Province of China (Rehabilitation Medicine, 2007.7–2010.7).
Competing interests: None.
Ethics approval: The study protocol was approved by the institutional ethics
committee at Sichuan University.
Provenance and peer review: Not commissioned; externally peer reviewed.
Daniels SK, Ballo LA, Mahoney MC, et al. Clinical predictors of dysphagia and
aspiration risk: outcome measures in acute stroke patients. Arch Phys Med Rehabil
Daniels SK, Brailey K, Priestly DH, et al. Aspiration in patients with acute stroke.
Arch Phys Med Rehabil 1998;79:14–19.
Johnson ER, McKenzie SW, Sievers A. Aspiration pneumonia in stroke. Arch Phys
Med Rehabil 1993;74:973–6.
Martin RE, Sessle BJ. The role of the cerebral cortex in swallowing. Dysphagia
Hamdy S, Rothwell JC, Aziz Q, et al. Organization and reorganization of human
swallowing motor cortex: implications for recovery after stroke. Clin Sci
Ertekina C, Aydogdu I. Neurophysiology of swallowing. Clin Neurophysiol
Hamdy S, Aziz Q, Rothwell JC, et al. The cortical topography of human swallowing
musculature in health and disease. Nat Med 1996;2:1217–24.
Hamdy S, Aziz Q, Rothwell JC, et al. Explaining oropharyngeal dysphagia after
unilateral hemispheric stroke. Lancet 1997;350:686–92.
Hamdy S, Rothwell JC, Aziz Q, et al. Long-term reorganisation of human motor
cortex driven by short term sensory stimulation. Nat Neurosci 1998;1:64–8.
Hamdy S, Aziz Q, Rothwell JC, et al. Recovery of swallowing after dysphagic stroke
relates to functional reorganization in the intact motor cortex. Gastroenterology
The European Stroke Organisation (ESO) Executive Committee and the ESO
Writing Committee. Guidelines for Management of Ischaemic Stroke and Transient
Ischaemic Attack 2008. Cerebrovasc Dis 2008;25:457–507.
Finestone HM, Greene-Finestone LS, Wilson ES, et al. Malnutrition in stroke patients
on the rehabilitation service and at follow-up: prevalence and predictors. Arch Phys
Med Rehabil 1995;76:310–16.
Hamdy S, Rothwell JC, Brooks DJ, et al. Identification of the cerebral loci processing
human swallowing with H2
Hamdy S, Mikulis DJ, Crawley A, et al. Cortical activation during human volitional
swallowing: an event related fMRI study. Am J Physio Gastrointest Liver Physiol
Kern MK, Birn R, Jaradeh S, et al. Swallow-related cerebral cortical activity maps
are not specific to deglutition. Am J Physiol Gastrointest Liver Physiol
Martin RE, Goodyear BG, Gati JS, et al. Cerebral cortical representation of automatic
and volitional swallowing in humans. J Neurophysiol 2001;85:938–50.
Martin RE, Macintosh BJ, Smith RC, et al. Cerebral areas processing swallowing and
tongue movement are overlapping but distinct: a functional magnetic resonance
imaging study. J Neurophysiol 2004;92:2428–43.
Mosier K, Liu WC, Maldjian JA, et al. Lateralization of cortical function in
swallowing: a functional MR imaging study. Am J Neuroradiol 1999;20:1520–6.
Narita N, Yamamura K, Yao D, et al. Effects of functional disruption of the lateral
pericentral cerebral cortex on primate swallowing. Brain Res 1999;824:140–5.
Mosier K, Bereznaya I. Parallel cortical networks for volitional control of swallowing
in humans. Exp Brain Res 2001;140:280–9.
Zald DH, Pardo JV. The functional neuroanatomy of voluntary swallowing. Ann
Daniels SK, Foundas AL. Lesion location in acute stroke patients with risk of
aspiration. J Neuroimaging 1999;9:91–8.
Robbins JA, Levin RL. Swallowing after unilateral stroke of the cerebral cortex:
preliminary experience. Dysphagia 1988;3:11–17.
Robbins JA, Levin RL, Maser A, et al. Swallowing after unilateral stroke of the
cerebral cortex. Arch Phys Med Rehabil 1993;74:1295–300.
Schroeder MF, Daniels SK, McClain M, et al. Clinical and cognitive predictors of
swallowing recovery in stroke. J Rehabil Res Dev 2006;43:301–10.
26.Cerebrovascular diseases: prevention, treatment, and rehabilitation. Report of a WHO
meeting. World Health Organ Tech Rep Ser 1971;469:1–57.
Logemann JA, Veis S, Colangelo L. A screening procedure for oropharyngeal
dysphagia. Dysphagi 1999;14:44–51.
Logemann JA. Manual for the videofluorographic study of swallowing, 2nd Edn.
Austin, TX: Pro-Ed, 1993.
Kuhlemeier KV, Yates P, Palmer JB. Intra- and interrater variation in the evaluation
of videofluorographic swallowing studies. Dysphagia 1998;13:142–7.
Friston KJ, Holmes AP, Poline JB, et al. Analysis of fMRI time-series revisited.
Mosier KM, Liu WC, Maldjian JA, et al. Lateralization of cortical function in
swallowing: A functional MR imaging study. Am J Neuroradiol 1999;20:1520–6.
Rosenbek JC, Robbins JA, Roecker EB, et al. A penetration-aspiration scale.
Weimar C, Ko ¨nig IR, Kraywinkel K, et al. Age and the National Institutes of Health
Stroke Scale within 6 hours after onset are accurate predictors of outcome after
15O PET activation. J Neurophysiol 1999;81:1917–26.
J Neurol Neurosurg Psychiatry 2009;80:1320–1329. doi:10.1136/jnnp.2009.176214 1327
cerebral ischemia: development and external validation of prognostic models. Stroke
Barbiera F, Condello S, De Palo A, et al. Role of videofluorography swallow study in
management of dysphagia in neurologically compromised patients. Radiol Med
Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human brain. New York:
Thieme Medical, 1988:37–110.
Truwit CL, Lempert TE. High resolution atlas of cranial neuroanatomy. Baltimore:
Williams & Wilkins, 1994:2–302.
Mosier K, Patel R, Liu WC, et al. Cortical representation of swallowing in normal
adults: functional implications. Laryngoscope 1999;109:1417–23.
Kern MK, Jaradeh S, Arndorfer RC, et al. Cerebral cortical representation of reflexive and
Hartnick CJ, Rudolph C, Willging JP, et al. Functional magnetic resonance imaging
of the pediatric swallow: imaging the cortex and the brainstem. Laryngoscope
Barer DH. The natural history and functional consequences of dysphagia after
hemispheric stroke. J Neurol Neurosurg Psychiatry 1989;52:236–41.
Daniels SK, Foundas AL, Iglesia GC, et al. Lesion site in unilateral stroke patients
with dysphagia. J Stroke Cerebrovas Dis 1996;6:30–4.
Gordon G, Hewer RL, Wade DT. Dysphagia in acute stroke. BMJ (Clin Res Ed)
Meadows JC. Dysphagia in unilateral cerebral lesions. J Neurol Neurosurg
Gow D, Rothwell J, Hobson A, et al. Induction of long-term plasticity in human
swallowing motor cortex following repetitive cortical stimulation. Clin Neurophysiol
Hamdy S, Aziz Q, Rothwell JC, et al. Sensormotor modulation of human cortical
swallowing pathways. J Physiol 1998;506:857–66.
Kreisel SH, Hennerici MG, Ba ¨zner H. Pathophysiology of stroke rehabilitation: the
natural course of clinical recovery, use-dependent plasticity and rehabilitative
outcome. Cerebrovasc Dis 2007;23:243–55.
Double depressor palsy caused by
bilateral paramedian thalamic infarcts
A 45-year-old woman with Coffin–Lowry syndrome, but
without risk factors for cerebrovascular disease, awoke with
double vision and unsteadiness on her feet. Examination
revealed a skew deviation of the eyes and diplopia on downgaze
due to a ‘‘double depressor’’ palsy of the inferior rectus and
superior oblique muscles. Her gait was unsteady with a
tendency to veer left. Brain computed tomography (CT) and
MRI confirmed bilateral paramedian thalamic infarcts (figs 1,
2). CT angiogram of the aortic arch and extracranial and
intracranial carotid arteries, thrombophilia screen and 24 h
ambulatory electrocardiogram were normal. An echocardiogram
lesions in the thalami (arrows).
Brain computed tomogram, showing bilateral low-intensity
(arrows) on (A) T2-weighted and (B) diffusion-weighted imaging.
Brain MRI, showing bilateral paramedian thalamic infarcts
1328 J Neurol Neurosurg Psychiatry December 2009 Vol 80 No 12
with saline contrast bubble demonstrated a right-to-left shunt Download full-text
across a large patent foramen ovale (PFO) without a valsalva
Almost one-third of people may have a common pedicle from
the proximal segment of one posterior cerebral artery,1whence
thalamoperforating branches supply the upper midbrain and
both posteromedial thalami, including the nucleus of the medial
longitudinal fasciculus, the posterior dorsomedial nucleus and
the intralaminar nuclei. Bilateral thalamic infarcts are likely to
be due to occlusion of a single thalamoperforating branch,
supplying the paramedian portions of both thalami, known as
the ‘‘artery of Percheron.’’2Manifestations include ophthalmo-
plegia, gait ataxia, hypersomnia and deficits of attention,
learning, memory or behaviour, from which recovery is usually
incomplete.3Double depressor palsy is very rare in itself and is
usually congenital rather than acquired; without impaired
consciousness or cognition, double depressor palsy is an
extremely rare manifestation of bilateral thalamic infarction.
Coffin–Lowry syndrome is a rare X linked disorder, in which
males are more severely affected than females by craniofacial
and skeletal abnormalities, learning difficulties, short stature
and hypotonia.4Mitral valve dysfunction and cardiomyopathy
have been found in a few affected males.5Although PFO has not
been described as part of the Coffin–Lowry syndrome, it may
simply have been a chance association in this case.
The patient was treated with aspirin 75 mg daily. Due to the
uncertainty about whether to close PFOs in cryptogenic stroke,
she was randomised in the Patent Foramen Ovale and
Cryptogenic Embolism (PC) trial (NCT00166257), investigating
the benefits of PFO closure versus conservative management.
Four months after symptom onset, she reported almost
complete resolution of both diplopia and gait unsteadiness.
S Pal, E Ferguson, S A Madill, R Al-Shahi Salman
Division of Clinical Neurosciences, University of Edinburgh, Western General Hospital,
Crewe Road, Edinburgh EH4 2XU, UK
Correspondence to: Dr R Al-Shahi Salman, Division of Clinical Neurosciences,
University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU,
Competing interests: None.
Patient consent: Obtained.
Provenance and peer review: Not commissioned; externally peer reviewed.
Accepted 4 February 2009
J Neurol Neurosurg Psychiatry 2009;80:1328–1329. doi:10.1136/jnnp.2008.167643
Castaigne P, Lhermitte F, Buge A, et al. Paramedian thalamic and midbrain infarct:
clinical and neuropathological study. Ann Neurol 1981;10:127–48.
Percheron G. Arteries of the human thalamus: II. Arteries and paramedian thalamic
territory of the communicating basilar artery. Rev Neurol (Paris) 1976;132:309–24.
Hermann DM, Siccoli M, Brugger P, et al. Evolution of neurological,
neuropsychological and sleep–wake disturbances after paramedian thalamic stroke.
Coffin GS, Siris E, Wegienka LC. Mental retardation with osteocartilaginous
anomalies. Am J Dis Child 1966;112:205–13.
Hanauer A, Young ID. Coffin–Lowry syndrome: Clinical and molecular features. J Med
BMJ Masterclasses are educational meetings designed specifically to meet the learning needs of
doctors. They help doctors keep up to date with the latest evidence and recent guidelines in major
clinical areas, enabling them to use the latest evidence to make better decisions. The latest evidence,
recent guidelines and best practice are delivered in an interactive and informative manner by leading
experts. The speakers are specifically chosen as highly-skilled communicators who can authoritatively
enthuse the audience and interpret the latest research and guidelines into practical tips for busy
doctors. BMJ Masterclasses have proved a huge hit with clinicians, with many saying they have
influenced their clinical practice.
J Neurol Neurosurg Psychiatry December 2009 Vol 80 No 121329