Pilot Study of Functional MRI to Assess Cerebral
Activation of Motor Function After
Poststroke Hemiparesis
Y. Cao, PhD; L. D’Olhaberriague, MD, PhD; E.M. Vikingstad, BS;
S.R. Levine, MD; K.M.A. Welch, MD
Background and Purpose—Studies of cerebral activation of motor function after ischemic stroke may enhance our
understanding of the underlying mechanisms of motor functional recovery, including the role of the
noninfarcted hemisphere.
Methods—Eight right-handed recovering hemiparetic or hemiplegic patients were studied using functional MRI. Results
were evaluated for each patient to consider individual variability in original functional organization, neuroanatomy, infarct
size and extent, treatment, age, and sex. The results were also pooled as a group for comparison with a control group of
eight right-handed normal subjects.
Results—In six of eight stroke patients, extended activation in ipsilateral sensorimotor cortex was observed during paretic
hand movements. Bilateral activation of the primary sensorimotor cortex was recorded in three of these six patients;
ipsilateral activation alone was recorded in the remaining three patients. Only two patients had mild synkinesia.
Furthermore, in two male patients, the paretic hand movements activated extended areas of ipsilateral premotor and
dorsolateral prefrontal cortex, when compared with normal subjects. In two patients with left frontal infarction, profound
activation in the right supramarginal gyrus and in the right premotor cortex was observed during the ipsilateral paretic hand
movements.
Conclusions—Synkinesia alone cannot explain the extent of ipsilateral activation in primary sensorimotor cortex. The
explanation offered for our findings is that preexisting uncrossed motor neural pathways may be accessed or recruited to
compensate for damage to the crossed motor pathways after ischemic stroke. (Stroke. 1998;29:112-122.)
Downloaded from http://ahajournals.org by on April 11, 2024

Key Words: functional reorganization n magnetic resonance imaging n motor activity n stroke, ischemic

M

otor function of limbs is predominantly represented in the
contralateral primary motor cortex in right-handed
healthy subjects. Therefore, damage of the primary motor
cortex and other motor pathway components (eg, premotor
cortex, supplementary motor areas, parietal cortex, and subcortical or brain stem) can cause contralateral hemiparesis or
hemiplegia, a common neurological ailment in stroke.1 Although poststroke hemiparesis is not a static phenomena, and
partial or complete recovery of motor function after stroke is
the rule more than the exception,2 mechanisms and determinants of this recovery are poorly understood. The concept of
brain function reorganization (plasticity)3 is useful to develop a
conceptual approach to understand motor recovery after
stroke. In recent years, new techniques (PET,4 transcranial
magnetic stimulation,5 and fMRI6) have been developed that
allow us to study the physiology and pathophysiology of the
motor pathways. Until now, no complete study of motor
recovery after stroke with fMRI has been reported. An fMRI

study of sensorimotor function in a group of young patients
with perinatal unilateral brain injury showed approximately
equally activated volumes in the uninjured hemisphere during
contralateral and ipsilateral finger movements.6 In adults with
motor loss due to striatocapsular infarction, PET identified
increased ipsilateral motor cortex activation and recruitment of
inferior parietal cortex and the anterior aspects of the insula.4,7
There are now a number of preliminary reports (in conference
presentations) in patients with ischemic stroke that have
documented activation of alternative motor patterns to normal
using fMRI.8 They suggest that cortical motor control of
function and pathways can be reorganized after focal ischemic
stroke and that the nondamaged hemisphere probably plays a
crucial role in recovery. We hypothesized that in the recovering hemiplegic or hemiparetic stroke patients the ipsilateral
activation in the primary sensorimotor cortex of the noninfarcted hemisphere would be increased, in comparison with
normal (control) subjects. To test this hypothesis, we have used

Received July 10, 1997; final revision received October 10, 1997; accepted October 31, 1997.
From the NMR Research Center, Department of Neurology, Henry Ford Health Sciences Center, Detroit, Mich, and the Department of Radiology
(E.M.V.), The University of Chicago (Illinois).
Correspondence to Yue Cao, PhD, NMR Research Center, Department of Neurology, NMR Laboratory, Henry Ford Hospital & Health Sciences
Center, 2799 W Grand Blvd, Detroit, MI 48202-2689.
E-mail yue@neurnis.neuro.hfh.edu
© 1998 American Heart Association, Inc.

112

Cao et al

Selected Abbreviations and Acronyms
AC 5 anterior cingulate
AG 5 angular gyrus
BOLD 5 blood oxygenation level– dependent
fMRI 5 functional magnetic resonance imaging
GRE 5 gradient recalled echo
LED 5 light-emitting diodes
MCA 5 middle cerebral artery
NIHSS 5 NIH Stroke Scale
paraCL 5 paracentral lobule
PET 5 positron emission tomography
SM 5 sensorimotor
SMG 5 supramarginal gyrus
SPL 5 superior parietal lobe
TPA 5 tissue plasminogen activator

an fMRI technique that is based on measurements of cerebral
BOLD contrast to investigate activation of the cerebral motor
cortex of hemiparetic patients recovering from focal ischemic
stroke. We explored the individual differences in lesion size
and location, and motor activation of cortical regions, as well
as grouped results from these patients.

Materials and Methods
Subjects

Downloaded from http://ahajournals.org by on April 11, 2024

We studied motor function of eight right-handed recovering hemiparetic or hemiplegic patients (mean age, 46 years; 2 men and 6
women) who suffered a single unilateral ischemic stroke and eight
right-handed normal control subjects (mean age, 42 years; 4 men and
4 women). The protocol was approved by the Human Rights
Committee of Henry Ford Hospital. Written, informed consent was
obtained from all subjects before the study. The handedness of the
subjects was assessed using the Edinburgh inventory.9
The criteria used to select patients were (1) Patients suffered a single
unilateral ischemic stroke determined from CT and/or MRI images,
causing hemiparesis or hemiplegia; (2) patients lost individual finger
movement at the onset of the stroke; and (3) patients must have
recovered to at least the point where they were capable of finger
opposition of thumb to the index and middle finger. There was no
time restriction on when the study was performed after the onset of
stroke symptoms. Patients were excluded from the study because of (1)
inability to obtain informed consent, (2) transient ischemic attack, (3)
brain stem stroke, (4) prior cerebrovascular disease, (5) preexisting
neurological or psychiatric disorders (ie, amyotrophic lateral sclerosis,
multiple sclerosis, Parkinsonism, AIDS, and dementia), (6) severe to
profound deafness and/or blindness, or (7) being nonambulatory.
The clinical presentation of the patients at the onset of the stroke
and at the time of the study was evaluated by a team of neurologists
(L.D., S.R.L., K.M.A.W.). The neurological abnormalities and motor
disabilities of the patients were assessed by the total and the motor
scores of the NIHSS.10 Overall recovery was also measured by the
modified Rankin score.11 Motor impairment of the paretic hands was
further assessed by a specific finger opposition task in which the time
required to perform 20 finger oppositions of each hand was recorded.
The prolonged time for the paretic hand performance was used as an
indicator of the motor impairment. Arterial territories and locations of
the infarcts were determined from T2-weighted MR images; dimensions of the infarcts were measured on the cross-sectional MR images.

Imaging Protocol
All MRI experiments were performed with a 3.0 T scanner with a
resonator head coil. A fast localizer scan was used to locate the motor
cortex and to guide adjustments of the patient’s position. A radio
frequency spoiled gradient echo–pulse sequence was used to produce
T2*-weighted images (128364 matrix over a 24-cm field of view).
Three axial sections of 10-mm thickness were acquired during 7

113

seconds with TR/TE538/30 milliseconds, flip angle 25°, and one
excitation, followed by a 0.4-second delay. The most superior slice
was placed at 10 mm below the vertex. The insula gyrus, thalamus,
and cerebellum were not imaged. Each functional study lasted
approximately 4 minutes, in which a total of 90 images were acquired.
The fMRI protocol consisted of 10 alternating periods of task
performance and rest. Axial T2-weighted images of cranial anatomy of
the whole head of each patient were acquired and used to identify the
location and size of the infarct. Also, T1-weighted images were
acquired at the same locations as the fMRI to overlap anatomically
with the voxels studied for functional activity.

Motor Protocol
During the functional MRI scan, all stroke patients and normal
control subjects performed a sequential finger opposition task in which
the thumb repeatedly touched the associated four fingers in a
sequential order. Patient 2, who could not touch all four fingers,
performed a three-finger opposition task. This task performance
occurred in periods of 24 seconds, interspaced with 24-second “rest”
periods. The cycle of task performance and rest was repeated 5 times
during each experiment. Two methods were used to instruct subjects
to start and stop the finger movement. In the first, a series of LEDs was
mounted about 1.5 feet anterior and inferior to the level of the
subject’s eyes as they lay in the magnet. The subject was instructed to
perform finger opposition while the LEDs were lit. When the LEDs
were off, the subject immediately ceased finger movement. In the
second method, an MRI-compatible video display system was used.
The subject wore mirrored goggles directed toward a display screen at
one end of the magnet. Simple instructions, such as “stop moving and
lie still” and “tap right fingers” were projected on the screen from a
computer-controlled video projector. To ensure that the patients
followed the start and stop signals, and that their finger-tapping speeds
were similar to those recorded outside the magnet, their task performance relative to prompting was monitored by a staff member who
was physically present with the patient throughout the study. If the
patient failed to carry out the task with acceptable response times and
speed, the scan was aborted. The staff explained the procedure to the
patient again, and the experiment was repeated. If the patient still had
trouble following instructions, the study was aborted again and the
patient took no further part in the study. The experiments were
repeated twice for each hand.

Data Analysis
Data analysis consisted of motion correction, statistical analysis of
activated pixels, and calculation of activated volumes in specific
anatomic regions. Motion correction was carried out by two-dimensional cross-correlation12 to correct for possible in-plane translation
and rotation of the head between serial images. Each image in the
series was registered with the first baseline image by applying a range
of test transformations consisting of planar translations and rotations.
The first image was subtracted on a pixel-by-pixel basis from the
transformed image. The transformation that yielded the minimum
difference in magnitude between the paired images was selected as the
optimal correction for the subject movement. A 0.125-degree rotation
and 0.12-mm translation were used to test for minimum rotation and
translation, respectively. If large movement of the head occurred,
motion artifact could not be corrected (usually .1.5 mm translation)
and the data were discarded. Such patients were later called back for
another study. Three studies were repeated due to motion artifact.
Statistical analysis of activated pixels was based on combinations of
temporal cross-correlation of the time course of the MR signals with
a sine function13 and cluster size thresholding14 to justify multiple
comparisons in an image. The pixels were thresholded at a level of
cross-correlation coefficients $0.513, which had one-tailed type I
error of 0.0025 per pixel. A second threshold of a cluster size at $5
pixels was further applied, resulting in the estimated probability of
false-positive of P,.0006 per pixel. Finally, activation images were
overlaid on T1-weighted anatomic images for further analysis.
Localization of the activated pixels in specific anatomic regions was
achieved by using the T1 weighted images and Human Brain Anatomy
in Computerized Images.15 Specifically, the number of pixels in the

Motor Function After Poststroke Hemiparesis

114
TABLE 1.

Motor Activation of Right-Handed Normal Subjects (n58)
Right Hand

Left Hand

Pooled Data

Contra

Ipsi

Contra

Ipsi

Contra

Ipsi

Sensorimotor

88.6620.2

12.365.7

85.4620.7

17.869.2

87.0617.8

15.065.7

Frontal

23.3611.6

9.665.8

8.464.0

4.963.5

15.866.6

7.363.1

AC, paraCL

5.362.7

7.862.7

2.361.1

0.860.8

3.861.6

4.361.6

SPL, SMG, AG

3.461.3

13.1612.0

1.961.9

5.063.4

2.660.9

9.167.5

Total

163.3637.9

126.3621.0

Contra (Ipsi); indicates activation was observed in the hemisphere contralateral (ipsilateral) to the
hand. Values are number of pixels, mean6SEM.

regions of SM cortex, the remaining parts of frontal lobe (including
superior, middle, and inferior frontal gyri), AC gyrus, paraCL, SPL,
SMG, and AG, was computed over three MR slices and tabulated.
Correlation between the age and time required for the unaffected
hand to perform 20 finger oppositions was examined using linear
regression. Furthermore, the two methods used to assess motor deficits
of the patients (the prolonged time for the paretic hand to perform 20
finger oppositions, and the summed score of NIHSS of the upper limb
and hand) were tested for correlation (linear regression).

Results
Summary of Normal Subjects

Downloaded from http://ahajournals.org by on April 11, 2024

Table 1 summarizes the fMRI results obtained from eight
right-handed control subjects. In control subjects, activation
obtained from sequential finger opposition primarily occurred
in the contralateral SM cortex with minor ipsilateral activation
for both hands. The mean6SEMs of the contralateral activation in the SM cortex were 88.6620.2 and 85.4620.7 pixels
for the right and left hands, respectively, and for ipsilateral
activation of SM cortex, 12.365.7 and 17.869.2 pixels for the
right and left hands, respectively. Activation was also observed
in contralateral and ipsilateral premotor cortex, AC gyrus,
paraCL, SPL, SMG, and AG. The mean and standard error
values of activated pixels in each specific region are given in
Table 1. No significant differences in ipsilateral SM cortex
activation (which might indicate use of uncrossed motor
pathway) were found between the dominant and nondominant
TABLE 2.

hemispheres (P5.64, nonparametric Wilcoxon signed rank
test). Therefore, we pooled data from both hemispheres (Table
1) to perform statistical comparisons with the patient group.

Individual Patient Results
Clinical presentations of the eight stroke patients at the onset of
stroke as well as at the time of study are summarized in Table
2. Severity of stroke and motor deficit, measured at the same
two time points by the total and the subgroup motor NIHSS
scores, are given in Table 3. Overall recovery at the time of
study, documented by the modified Rankin score, is also listed
in Table 3. The time required to perform 20 finger oppositions
of each hand and the prolonged time for paretic hand performance are given in Table 3. In the following section, a brief
description of the clinical presentation and fMRI findings of
each stroke patient will be given.
Patient 1, a 25-year-old woman, suffered her first ever
ischemic stroke due to dissection of the left internal carotid
artery (the only subject with extracranial artery involvement)
that caused a solitary infarct (largest dimension 7.9 cm) in the
left anterior cerebral artery territory (Table 2). The infarct
extended from the superior frontal gyrus to the superior
portion of the precentral gyrus, and to the paraCL, and
partially spared the hand motor area (Fig 1). At stroke onset,
she experienced right hemiplegia, right hemisensory loss, and a
transcortical mixed aphasia. At the time of study (5 months

Summary of Clinical Characteristics of Stroke Patients
Ischemic Infarct

Pt No. Sex Age, y Side

Topography

Clinical Features

Artery
Dim,
Territories cm

Onset

Time of Study

Synkinesia

1

F

25

L

Frontal

ACA

7.9

R hemiplegia f-a-l1transcortical mixed R hemiparesis1transcortical motor
aphasia
aphasia

None

2

F

19

L

Frontoparietotemporal

MCAB

9.1

R hemiplegia f-a-l1global
aphasia

R hemiparesis1motor aphasia

None

3

F

41

L

Frontoparietal

MCAB

6.0

R hemiplegia f-a-l1global
aphasia

R hemiparesis f-a1motor aphasia

Yes

4

M

60

R

Frontoparietotemporal

MCA

12.0

L hemiplegia1
neglect1gaze palsy

L S-M f-a

None

5

F

51

R

Frontoparietotemporal

MCAB

5.5

L hemiplegia f-a-l1neglect

L hemiparesis f-a

None

6

F

70

R

Parietal

MCAB

5.0

L S-M f-a-l1neglect

L S-M f-a-l

None

7

M

64

R

Corona radiata

PB

2.4

L pure hemiparesis f-a-l

L pure hemiparesis f-a

Yes

8

F

36

R

Striatocapsular

LSA

3.6

L hemiplegia

L S-M f-a-l

None

Pt indicates patient; S-M, sensorimotor syndrome; f-a-l, face-arm-leg; Dim, largest dimension; MCAB, middle cerebral artery branch; PB, parietal branch; LSA,
lenticulostriates arteries; and ACA, anterior cerebral artery.

Cao et al
TABLE 3.

115

Measurements of Stroke Severity and Motor Deficits of Stroke Patients
At Onset of Stroke

At Time of Study
Time for 20
Finger Oppositions, s

Prolonged
Time, s

Pt.
No.

Total

LL

UL

Hand

Time
Interval,
mo

NH

PH

PH vs NH

1

15

4

3

1

5

4

1

1

0

2

23

34

11

2

24

4

4

2

14

11

3

2

1

3

30

80†

50

3

20

4

4

2

5

8

0

0

1

3

38

51

13

4

13

2

3

2

5

3

0

1

0

1

38

39

1

5

*

*

*

*

43

3

0

1

1

2

27

32

5

6

10

3

2

2

19

7

2

1

1

1

55

75

20

7

5

1

1

1

10

3

0

1

1

1

59

75

16

8

*

*

*

*

24

5

1

1

1

2

40

84

44

NIHSS Scores

NIHSS Scores
Total

LL

UL

Hand

Rankin
Scores

NIHSS, indicates NIH Stroke Scale; LL, lower limb; UL, upper limb; NP, normal hand; and PH, paretic hand.
*Patients 5 and 8 were evaluated at another hospital at stroke onset; therefore, the NIHSS scores were not available.
†Time was obtained from finger’s open-close movements.

Downloaded from http://ahajournals.org by on April 11, 2024

after the stroke), she had recovered sufficiently to perform
finger opposition of her paretic hand at a rate of 2.4 Hz,
compared with 3.5 Hz for her unaffected hand. She continued
to suffer right hemiparesis and transcortical motor aphasia.
Her normal (left) hand activated the contralateral SM cortex
and the lateral premotor cortex. Minor activation was also
observed in the ipsilateral SM cortex and the contralateral SPL
and SMG. The number of activated pixels in each region is
listed in Table 4. Activation produced by her paretic (right)
hand movements was primarily observed in the contralateral
SM cortex, ipsilateral SMG, and lateral premotor cortex (Fig
1). A small volume in ipsilateral SM cortex was also activated.
The location of activation in contralateral SM cortex during
the paretic hand movements was shifted inferiorly, in comparison to that produced by her unaffected hand movements. The
activated volume in the ipsilateral SMG (123 pixels) of this
patient was profoundly enlarged when compared within the
patient group (the mean of the number of activated pixels was
28, see Table 4) and when compared with the control group
(the mean of 9 pixels, see Table 1).
Patient 2, a 19-year-old woman, suffered an ischemic infarct
(largest dimension 9.1 cm) in the left hemisphere involving the
frontal, anterior parietal, and temporal cortices (Table 2). This
was caused by branch occlusion of the MCA. The infarct
further extended into the posterior portion of the putamen and
globus pallidus. At the onset of her stroke, she experienced
right hemiplegia and global aphasia. At the time of the study
(14 months after stroke onset), she suffered the most severe
hand motor impairment among the patient group; she could
only tap three fingers of the paretic hand and took the most
prolonged time (50 seconds) to perform the task.
During her normal (left) hand movements, only the contralateral SM cortex was activated. However, the paretic (right)
hand activated minimal volumes in the contralateral SM, AC,
and paraCL. Instead, the ipsilateral motor cortex, premotor,
SMG, and SPL were activated. A total number of 39 pixels was
activated in SMG and SPL. This was four times the mean value
of the control group.

Patient 3, a 41-year-old woman, suffered a first-ever ischemic stroke due to branch occlusion of the left MCA, resulting
in an infarct of the left frontal and parietal cortices (largest
dimension 6.0 cm) (Fig 2). At stroke onset, she experienced
right hemiplegia and global aphasia (Table 2). Five months
after the stroke, she regained her right hand movement, even
though she continued to suffer right hemiparesis involving face
and arm and motor aphasia. The residual impairment of the
paretic (right) hand was further demonstrated by 13 seconds of
prolonged performance time compared with the normal (left)
hand (Table 3). She also developed minor synkinesia of the
unaffected (left) hand. Her unaffected hand was twitching with
a magnitude of less than 1 mm while the paretic hand
performed a motor task.
The unaffected hand movements of this patient produced
activation in the contralateral SM cortex, premotor cortex, and
SPL. For the paretic hand movements, no activation was
obtained in the contralateral hemisphere (infarct site); instead,
ipsilateral SM and premotor cortex was markedly activated at
a locus nearly identical to that activated by the unaffected hand
movements (Fig 2).
Patient 4, a 60-year-old man, suffered the largest infarct
within the group, due to right MCA occlusion. The infarct
was located in the right frontal region, extending posteriorly to
the parietal cortex and inferiorly to the temporal cortex (largest
dimension 12.0 cm). The infarct spared the striatocapsular
region and the hand area of the primary motor cortex. The
patient was the only subject who had received acute
thrombolytic therapy with tissue plasminogen activator (TPA).
At the onset of the stroke, he experienced left hemiplegia,
neglect, and gaze palsy. At the time of study (5 months after
stroke onset), he had recovered the most completely within the
patient group, even though he continued to suffer left sensorimotor deficit involving face and arm. Both of his hands had
almost identical kinetic ability (38 and 39 seconds required for
20 finger oppositions of the normal and paretic hands, respectively); however, the magnitude of the paretic hand movements was less than the unaffected hand.

116

Motor Function After Poststroke Hemiparesis

Downloaded from http://ahajournals.org by on April 11, 2024

Figure 1. Patient 1 suffered an ischemic infarct, with hyperintensity in T2-weighted MR images (bottom two rows), in the left anterior
cerebral artery territory. The infarct is denoted by arrow heads. Motor activation (red color) produced by movement of the unaffected
(left) and paretic (right) hands is overlaid on T1-weighted MR images, and illustrated in the first and second rows of images, respectively. Note that the location of activation in the contralateral SM cortex during paretic hand movements was shifted inferiorly, compared with locations activated by unaffected hand movements. The right SMG region was profoundly activated during ipsilateral paretic
hand movements. L indicates left.

This patient’s unaffected (right) hand movements activated
the contralateral SM cortex and bilateral SMA and AC gyrus.
Activation from his paretic (left) hand was observed in contralateral SM cortex, bilateral SMG, contralateral AC gyrus,
and ipsilateral lateral premotor cortex. The activated volumes
in the ipsilateral lateral premotor and prefrontal regions were
approximately 12-fold larger than the average value of the
normal control group. No activation was observed in the

ipsilateral SM cortex during the paretic hand movements
(Table 4).
Patient 5, a 51-year-old woman, suffered her first ischemic
stroke due to branch occlusion of the right MCA, resulting in
an infarct (largest dimension 5.5 cm) in the right frontoparieto-temporal cortex (Fig 3). She experienced left hemiplegia and neglect at stroke onset. At the time of the study (43
months after stroke onset), she continued to suffer left hemi-

Cao et al
TABLE 4.

117

Motor Activation of Stroke Patients
Normal Hand Movement

Paretic Hand Movement

Contra
Hemisphere

Ipsi
Hemisphere

Contra
Hemisphere

Ipsi
Hemisphere

Sensorimotor

249

11

204

85

Frontal

56

0

0

28

AC, paraCL

6

0

0

6

SPL, SMG, AG

6

0

0

123

Sensorimotor

111

0

13

42

Frontal

0

0

0

9

AC, paraCL

0

0

11

0

SPL, SMG, AG

0

0

0

39

Sensorimotor

153

0

0

31

Frontal

42

0

0

0

AC, paraCL

0

0

0

0

SPL, SMG, AG

32

5

0

25

Sensorimotor

Patient No.
1

2

3

4
220

7

58

0

Frontal

0

20

33

89

AC, paraCL

28

25

30

0

SPL, SMG, AG

24

16

8

17

Sensorimotor

5
Downloaded from http://ahajournals.org by on April 11, 2024

122

0

16

110

Frontal

0

0

0

12

AC, paraCL

17

0

14

17

SPL, SMG, AG

0

0

6

10

Sensorimotor

11

0

18

0

Frontal

0

6

0

0

AC, paraCL

0

0

8

0

SPL, SMG, AG

0

0

0

0

Sensorimotor

180

7

61

89

Frontal

15

24

35

196

AC, paraCL

0

0

20

11

SPL, SMG, AG

0

0

0

12

Sensorimotor

46

6

17

59

Frontal

0

0

9

6

AC, paraCL

0

0

8

0

6

7

8

0

15

0

0

Sensorimotor (mean6SEM)

SPL, SMG, AG

136628.9

3.961.6

48.4623.5

52.0614.5

Frontal (mean6SEM)

14.167.9

6.363.5

9.665.4

42.5624.3

AC, paraCL (mean6SEM)

6.463.7

3.163.1

11.463.6

4.362.3

SPL, SMG, AG (mean6SEM)

7.864.5

4.562.5

1.861.2

28.3614.3

Total (mean6SEM)

182.5641.8

198.1156.8

Values are number of pixels. Contra indicates contralateral and Ipsi ipsilateral.

118

Motor Function After Poststroke Hemiparesis

Downloaded from http://ahajournals.org by on April 11, 2024

Figure 2. Patient 3 suffered an ischemic infarct, with hyperintensity in T2-weighted MR images (bottom row), in the left MCA territory.
The infarct is denoted by arrow heads. Motor activation (red color) overlaid on T1-weighted MR images is shown in the top row of
images. The left two images were obtained during unaffected (left) hand movements; the right two images represent activation during
paretic (right) hand movements. Note that the ipsilateral activation produced by paretic hand movements was at a locus nearly identical
to that activated by unaffected hand movements.

paresis of face and arm. A minimally prolonged time (5
seconds) was required for her paretic (left) hand to perform the
finger opposition task, in comparison to the unaffected hand.
No synkinesia was observed.
This patient’s unaffected (right) hand movements activated
the contralateral SM cortex, SMA, and lateral premotor cortex.
Her paretic (left) hand movements activated ipsilateral SM and
premotor cortex, and bilateral SMA (Fig 3). The volume
activated in the ipsilateral SM cortex was approximately
10-fold greater than the mean volume of the control group.
Minimal activation was observed in the contralateral SM
cortex (Table 4).
Patient 6, a 70-year-old woman, suffered an infarct (largest
dimension 5.0 cm) in the right parietal cortex due to branch
occlusion of the right MCA. At stroke onset, she experienced
a left SM paresis of face, arm, and leg and neglect. At the time
of study (19 months after stroke onset), she continued to suffer
SM deficits of left face, arm, and leg. Compared with the
unaffected hand, her paretic hand required a prolonged time of
20 seconds to perform the finger opposition task, indicating
moderate recovery.
During her unaffected hand movements, the contralateral
SM cortex was activated. The paretic hand movements activated the contralateral motor cortex and AC gyrus. This
patient is the only one in whom no ipsilateral activation was
observed.

Patient 7, a 64-year-old man, suffered a solitary infarct
(largest dimension 2.4 cm) in a region of the right corona
radiata. At stroke onset, he experienced pure hemiparesis of the
left face, arm, and leg. The severity of the stroke, documented
by the total NIHSS score of 5, was the least of the group. By
the time of the study (10 months after stroke onset), the patient
developed minor synkinesia of the normal (right) hand. The
motor deficits of the paretic hand were indicated by the longer
time (16 seconds) required to perform the finger task, compared with the unaffected hand (Table 3).
The unaffected (right) hand movements activated contralateral SM cortex. The paretic hand movements produced
bilateral activation in SM cortex, SMA, and lateral premotor
cortex. The activated pixels in the ipsilateral SM cortex and the
premotor region were 6 and 28 times, respectively, larger than
those observed in normal control subjects (Tables 1 and 4).
Patient 8, a 36-year-old woman, suffered an ischemic stroke
due to occlusion of lenticulostriate arteries. The infarct (largest
dimension 5.0 cm) was located in the right striatocapsular
region, extending inferiorly to the putamen and globus pallidus. She experienced left hemiplegia at the time of stroke
onset. At the time of the study (24 months after stroke onset),
she continued to experience left SM deficits in the face, arm,
and leg. Based on the finger opposition task, the motor
impairment of her paretic hand was next to the most severe
within the patient group (Table 3).

Cao et al

119

Downloaded from http://ahajournals.org by on April 11, 2024

Figure 3. Patient 5 suffered an ischemic infarct, with hyperintensity in T2-weighted MR images (bottom two rows), in the right MCA
territory. The infarct is denoted by arrow heads. Motor activation (red color) overlaid on T1-weighted MR images is shown in the top
row of images. The left two images were obtained during unaffected (right) hand movements; the right two images represent activation
during paretic (left) hand movements. No activation was observed in the contralateral SM cortex during paretic hand movements.

During unaffected (right) hand movements, the contralateral
SM cortex was activated. During the paretic (left) hand
movements, there was activation of bilateral SM cortex.
Activation of the ipsilateral SM cortex (59 pixels) was at a locus
nearly identical to that activated by the unaffected hand
movement.

Group Results
In this study, the motor deficit of the patient’s paretic hand was
also assessed by the time required to perform 20 finger
oppositions. The time required for the unaffected hand performance (Table 3) was significantly correlated with age
(r5.75, P5.032, n58, linear regression). The prolonged time
required for the paretic hand to carry out the task as compared
with the unaffected hand (Table 3) was significantly correlated
with the summed NIHSS score for the upper limb and hand
(r5.74, P,.036, n58, linear regression). No correlation was
found between the residual motor deficit of the paretic hand
(either measured by the prolonged time for the finger opposition or by the summed NIHSS score of the upper limb and

hand) and the activated volumes in ipsilateral SM cortex
among the six patients who had the increased ipsilateral
activation (rs52.543, P..2, Spearman rank correlation).
In the patient group, paretic hand movements produced the
nearly identical number of total activated pixels (mean6SEM,
198656 pixels) as did their unaffected hand movements
(mean6SEM, 182642 pixels). No significant difference in the
total number of activated pixels was detected between the
patient and control groups (P..45, nonparametric MannWhitney U test). These data establish the bases for comparison
of normal and paretic hands within the patient group as well as
between patient and control groups, even though variability of
effort required to perform the motor task existed between the
two groups and between the two hands within the patient
group.
During paretic hand movements, the volume of activated
ipsilateral SM cortex was extensive in six of eight recovering
stroke patients compared with control subjects. Bilateral activation of the primary SM cortex was recorded in three of these
patients; ipsilateral activation alone was recorded in three

120

Motor Function After Poststroke Hemiparesis

patients who suffered major cerebral infarctions with hemiplegia. Only two patients had mild synkinesia, which alone
cannot be responsible for the increased ipsilateral activation in
primary SM cortex. Furthermore, in two male patients, the
paretic hand movements activated more extensive volumes of
ipsilateral premotor and dorsolateral prefrontal cortex compared with control subjects. In two patients with left frontal
infarction, profound activation was observed in the right SMG
(Brodmann area 40) and in the right premotor cortex during
the ipsilateral paretic hand movements. Even when one
considers that only six patients had ipsilateral activation, in the
total patient group the paretic hand produced a statistically
significant increase in the number of activated pixels in the
ipsilateral SM cortex (mean, 52 pixels) compared with that
observed in the control group (mean, 15 pixels) (P,.03,
Mann-Whitney U test).

Discussion

Downloaded from http://ahajournals.org by on April 11, 2024

The technique of fMRI used in this study is based on BOLD
contrast16 –19 between rest and activated states of human brain.
During activation of normal brain, oxygenation of cerebral
blood increases because the increase of cerebral blood flow
(29% to 50%)20,21 exceeds the increase in oxygen extraction
(5%).22 Although in ischemic tissues the blood flow–metabolism couple is impaired,23–25 it remains normal in noninfarcted
tissue.4,23,25 For example, PET studies of patients with striatocapsular infarction4 confirmed a cerebral blood flow increase
during task performance in ipsilateral primary motor cortex,
lateral prefrontal cortex, and insula. For the purpose of this
fMRI study we focused on functional activation in the
noninfarcted hemisphere and normal tissue regions in the
damaged hemisphere.
Our study has shown different patterns of cortical motor
activation in recovering hemiparetic patients, who suffered
cortical lesions in the majority. The primary SM cortex in
noninfarcted hemispheres of six of the eight recovering
hemiparetic patients was activated during movements of the
ipsilateral (paretic) hand. In three patients who suffered
infarcts that spared the hand area of the primary motor
cortex (two subcortical and one superior frontal), the paretic
hand movements led to bilateral activation of the SM
cortex. In three other patients with precentral gyrus infarction, only ipsilateral activation occurred during paretic hand
movements. These activation patterns were significantly
different from those observed in our normal control subjects. One possible explanation is that in some patients
recovering from hemiparesis after unilateral ischemic stroke,
functional motor pathways may reorganize in an attempt to
recruit any prestroke link between the paretic hand and the
primary motor cortex in the noninfarcted hemisphere via
uncrossed corticospinal tracts or other indirect uncrossed
pathways. In normal humans, uncrossed corticospinal tracts,
comprising about 10% to 15% of all of the corticospinal
fibers,26 rarely participate in distal limb movements. This is
supported by our observation in the control group of 20%
or less ipsilateral activation in each hemisphere, as well as
similar results from other activation studies.12,27–30 In normal
nonhuman primates, the majority of primary motor cortex
neurons subserved contralateral movement; only 7% to 8%

of neurons were associated with ipsilateral distal limb
movement.31 Our findings are further supported by studies
in patients with pyramidal tract infarcts4,32 and in young
patients (mean age, 15 years) with hemiparesis who suffered
perinatal unilateral brain injury.6 A PET study of individual
motor patterns after stratiocapsular stroke found that increased ipsilateral activation in primary SM cortex was
observed only in patients with associated movements,32 so it
could not be determined whether this ipsilateral activation
reflected recruitment of uncrossed corticospinal pathways or
was secondary to the associated movements. In the present
study, however, increased ipsilateral primary sensorimotor
activation could not be explained in this way, because four
of the six patients with this finding had no associated
movements. Furthermore, as observed in one (patient 3) of
our two patients with synkinesia, the fact that there was
exclusive ipsilateral activation during paretic hand movements indicates that the noninfarcted hemisphere must
control both voluntary and associated movements. Therefore, other physiological explanations for ipsilateral activation, such as reorganization of functional motor pathways or
recruitment of the preexisting uncrossed pathways, must be
considered.
In two patients (patients 1 and 2) with the large left frontal
infarcts, we observed extensive activation in the right SMG
(Brodmann area 40) and right premotor cortex during ipsilateral paretic hand movements. Parietal lobe lesions are known
to impair skilled movements and to affect performance of
symbolic gestures, motor production, and ideation.33–35 Neuroimaging patterns of cerebral activation (eg, in premotor,
SMA, and parietal regions) during both mental rehearsal of a
motor task and its execution36,37 add further support that the
posterior parietal cortex is an important component of motor
pathways. Specific contributions of the parietal cortex in motor
behavior were recently studied during the mental representation of hand movements after parietal cortex damage.38
This study demonstrated that parietal cortex lesions selectively
impaired capability of predicting the time necessary to perform
differentiated finger movements and visually guided pointing
gestures. Furthermore, Weiller et al32 have proposed an accessory motor system involving Brodmann area 40 and those
anterior portions of the insula functionally connected to the
motor network via direct projections to the inferior premotor
cortex or the centromedian nucleus of the thalamus. One
possible explanation for our finding of enhanced activation in
the right SMG after left ischemic infarction is that ipsilateral
nonprimary cortical motor areas and their efferent corticocortical pathways may compensate for disruption of the output of
the contralateral primary motor cortex and for inefficiency of
the ipsilateral primary motor cortex and the uncrossed corticospinal tracts. In a separate study, the posterior parietal cortex,
particularly the right hemisphere, was activated in trial-anderror motor sequences.39 Activation of the posterior parietal
cortex was also observed during self-paced complex finger
movements.40 The specific contribution or possibly multiple
roles of Brodmann area 40 in complex motor behavior remain
to be determined, however. Because the right SMG area
(Brodmann area 40) may also be part of a spatial attention
network,41,42 the possibility cannot be ruled out that the right

Cao et al

Downloaded from http://ahajournals.org by on April 11, 2024

SMG activation in our patients is due to enhanced attention
required during paretic hand movements.
Problematic in all activation and task performance studies is
how to control for the effort of subjects and the associated
nonspecific cognitive components. The outcome of increased
effort may be enlarged activation volumes and nonspecific
cognitive components may recruit newly activated regions. In
previous normal motor studies, the pace of movement was
controlled because it is linearly proportional to the magnitude
of activation.43 However residual motor impairment in recovering hemiparetic patients may exaggerate the variability in
effort if the speed of movements is controlled. Also, prolonged
finger movements may cause fatigue and increased effort by
hemiparetic patients, or even normal subjects. Because this
may result in the use of proximal limb muscles in addition to
hand muscles, this might explain expanded regions of motor
cortex activation. We attempted to overcome this in our pilot
study by using a self-paced, more automatic finger opposition
task with comfortable performance time (24 s) for most
patients.
A 3T magnet and GRE pulse sequence were used for this
study. The approach to studying motor function with GRE
pulse sequences has several limitations. First, the examined
volume was limited to a 30-mm thick section of the brain
rather than the entire head. Second, the GRE pulse sequence
is more sensitive to motion than echo-planar imaging (EPI).
Third, T2*-weighted images, acquired by either GRE or
gradient-echo EPI techniques, are sensitive not only to tissue
blood oxygenation changes but also to BOLD signals in
venules and veins.44 However, the percentage of BOLD signals
arising from capillaries and brain tissue increases with magnetic
field strength.45,46 In this respect, then, a 3T magnet is superior
to a 1.5T scanner.
In summary, we have demonstrated that fMRI allows study
of recovery of motor function after ischemic stroke in humans.
We were able to investigate differences in motor activation of
individual patients and to explore the group results. Our study
suggests that BOLD contrast is sensitive to functional recovery
and organization of motor pathways. In the future, fMRI may
be a useful tool to monitor and study poststroke rehabilitation
and may be important in aiding our understanding of the
mechanisms of poststroke motor recovery.

Acknowledgments
This work was supported in part by NIH grant NS 23393 and an AHA
Grant-in-Aid (96009160). The authors would like to thank Patricia
Booker and Susan Ferrari for their assistance in helping to recruit
patients, Dr Rima Dafer for patient referral, Dr Mike Boska for
methodological advice in the NMR Laboratory, and Dr Jay Nelson
for technical support.

References
1. Foulkes MA, Wolf PA, Price TR, Mohr JP, Hier DB. The pilot stroke data
bank: design, methods, and baseline characteristics. Stroke. 1988;19:
547–554.
2. Duncan PW, Goldstein LB, Matchar D, Divine GW, Feussner J. Measurement of motor recovery after stroke: outcome measures and sample size
requirements. Stroke. 1992;23:1094 –1098.
3. Hallet M. The plastic brain. Ann Neurology. 1995;38:4 –5.
4. Weiller C, Chollet F, Friston KJ, Wise RJS, Frackowiak RSJ. Functional
reorganization of the brain in recovery from striatocapsular infarction in
man. Ann Neurol. 1992;31:463– 472.

121

5. Bridgers S. Magnetic cortical stimulation in stroke patients with hemiparesis. In: Chokroverty S, ed. Magnetic Stimulation in Clinical Neurophysiology. Boston, Mass: Butterworths; 1989:233–247.
6. Cao Y, Vikingstad EM, Huttenlocher PR, Towle VL, Levin DN. Functional magnetic resonance studies of the reorganization of the human hand
sensorimotor area after unilateral brain injury in the perinatal period. Proc
Natl Acad Sci U S A. 1994;91:9612–9616.
7. Chollet F, DiPiero V, Wise RJS, Brooks DJ, Dolan RJ, Frackowiak RSJ.
The functional anatomy of motor recovery after stroke in humans: a study
with positron emission tomography. Ann Neurol. 1991;29:63–71.
8. Cao Y, D’Olhaberriague D, Levine SR. Brain reorganization during motor
recovery from stroke in humans. Stroke. 1996;27:194.
9. Oldfield RC. The assessment and analysis of handedness: the Edinburgh
Inventory. Neuropsychologia. 1971;9:97–113.
10. Brott T, Adams HPJ, Olinger CP, Marler JR, Barsan WG, Biller J, Spilker
J, Holleran R, Eberle R, Herztberg V, Rorick M, Moonaw CJ, Walker M.
Measurement of acute cerebral infarction: a clinical examination scale.
Stroke. 1989;20:864 – 870.
11. Van Jign J. Measurement of outcome in stroke prevention trials. Cerebrovasc
Dis. 1992;2(suppl 2):23–34.
12. Cao Y, Towle VL, Levin DN, Balter JM. Functional mapping of human
motor cortical activation by conventional MRI at 1.5 T. J Magn Reson
Imaging. 1993;3:869 – 875.
13. Bandettini PA, Jesmanowicz A, Wong EC, Hyde JS. Processing strategies
for time-course data sets in functional MRI of the human brain. Magn
Reson Med. 1993;30:161–165.
14. Forman SD, Cohen JD, Fitzgerald M, Eddy WF, Mintun MA, Noll DC.
Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold. Mag Reson Med.
1995;33:636 – 647.
15. Damasio H. Human Brain Anatomy in Computerized Images. New York, NY:
Oxford University Press; 1995.
16. Ogawa S, Lee T, Nayak AS, Glynn P. Oxygenation-sensitive contrast in
magnetic resonance image of rodent brain at high magnetic fields. Magn
Reson Med. 1990;14:68 –78.
17. Ogawa S, Tank DW, Menon R, Ellerman JM, Kim SG, Merkle H,
Ugurbil K. Intrinsic signal changes accompanying sensory stimulation:
functional brain mapping using MRI. Proc Natl Acad Sci U S A. 1992;89:
5951–5955.
18. Kwong KK, Belliveau JW, Chesler DA, Goldberg IE, Weisskoff RM,
Poncelet BP, Kennedy DN, Hoppel BE, Cohen MS, Turner R, Cheng
H-M, Brady TJ, Rosen BR. Dynamic magnetic resonance imaging of
human brain activity during primary sensory stimulation. Proc Natl Acad Sci
U S A. 1992;89:5675–5679.
19. Bandettini PA, Wong EC, Hinks PS, Tikofsky RS, Hyde JS. Time course
EPI of human brain function during task activation. Magn Reson Med.
1992;25:390 –397.
20. Fox PT, Meizin FM, Allman JM, Van Essen DC, Raichle ME. Retinotopic organization of human visual cortex mapped with positron
emission tomography. J Neurosci. 1987;7:913–922.
21. Deiber MP, Passingham RE, Colebatch JG, Friston KJ, Nixon PD,
Frackowiak RSJ. Cortical areas and the selection of movement: a study
with positron emission tomography. Exp Brain Res. 1991;84:393– 402.
22. Fox PT, Raichle ME. Focal physiological uncoupling of cerebral blood
flow and oxidative metabolism during somatosensory stimulation in human
subjects. Proc Natl Acad Sci U S A. 1986;83:1140 –1144.
23. Heiss W-D, Huber M, Fink GR, Herholz K, Pietrzyk U, Wagner R,
Wienhard K. Progressive derangement of periinfarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab. 1992;12:193–203.
24. Wise RJS, Bernardi S, Frackowiak RSJ, Legg NJ, Jones T. Serial observations on the pathophysiology of acute stroke. Brain. 1983;106:197–222.
25. Wise RJS, Rhodes CG, Gibbs JM et al. Disturbance of oxidative metabolism of glucose in recent human cerebral infarcts. Ann Neurol. 1983;14:
627– 637.
26. Nyberg-Hansen R, Rinvik E. Some comments on the pyramidal tract,
with serial reference to its individual variations in man. Acta Neurol Scand.
1963;39:1–30.
27. Ramachadran VS. Behavioral and magnetoencephalographic correlates of
plasticity in the adult human brain. Proc Natl Acad Sci U S A. 1993;90:
10413–10420.
28. Colebatch JG, Deiber M-P, Passingham RE, Friston KJ, Frackowiak RSJ.
Regional cerebral blood flow during voluntary arm and hand movements
in human subjects. J Neurophysiol. 1991;65:1392–1401.
29. Fox PT, Burton H, Raichle ME. Mapping human somatosensory cortex
with positron emission tomography. J Neurophysiol. 1987;67:34 – 43.

122

Motor Function After Poststroke Hemiparesis

30. Kim SG, Ashe J, Georgopoulos AP et al. Functional imaging of human
motor cortex at high magnetic field. J Neurophysiol. 1993;69:297–302.
31. Tanji J, Okano K, Sato KC. Neuronal activity in cortical motor areas
related to ipsilateral, contralateral and bilateral digit movements of the
monkey. J Neurophysiol. 1988;60:325–343.
32. Weiller C, Ramsay SC, Wise RJS, Friston KJ, Frackowiak RSJ. Individual
patterns of functional reorganization in the human cerebral cortex after
capsular infarction. Ann Neurol. 1993;33:181–189.
33. Clark MA, Merians AS, Kothari A, Poizner H, Macauley B, Rothi LJG,
Heilman KM. Spatial planning deficits in limb apraxia. Brain. 1994;117:
1093–1106.
34. Jeannerod M. The formation of finger grip during prehesion: a cortically
mediated visuomotor pattern. Behav Brain Res. 1986;19:99 –116.
35. Heilman KM, Rothi LJ, Valenstein E. Two forms of ideomotor apraxia.
Neurology. 1982;32:342–346.
36. Decety J, Perani D, Jeannerod M, Bettinardi V, Tadary B, Woods R,
Mazziotta JC, Fazio F. Mapping motor representations with positron
emission tomography. Nature. 1994;371:600 – 602.
37. Stephan KM, Fink GR, Passingham RE, Silbersweig D, CeballosBaumann AO, Frith CD, Frackowiak RSJ. Functional anatomy of the
mental representation of upper extremity movements in healthy subjects.
J Neurophysiol. 1995;73:373–386.
38. Sirigu A, Duhamel J-R, Cohen L, Pillon B, Dubois B, Agid Y. The mental
representation of hand movements after parietal cortex damage. Science.
1996;273:1564 –1568.
39. Jenkins IH, Brooks DJ, Nixon PD, Frackowiak RSJ, Passingham RE.
Motor sequence learning: a study with positron emission tomography.
J Neurosci. 1994;14:3775–3790.

40. Rao SM, Binder JR, Bandettini PA, Hammeke TA, Yetkin FZ,
Jesmanowicz A, Lisk LM, Morris GL, Mueller WM, Estkowski LD, Wong
EC, Haughton VM, Hyde JS. Functional magnetic resonance imaging of
complex human movements. Neurology. 1993;43:2311–2318.
41. Gitelman DR, Alpert NM, Kosslyn S, Daffner K, Scinto L, Thompson W,
Mesulam M-M. Functional imaging of human right hemispheric activation
for exploratory movements. Ann Neurol. 1996;39:174 –179.
42. Selemon LD, Goldman-Rakic PD. Common cortical and subcortical
targets of the dorsolateral prefrontal and posterior parietal cortices in the
rhesus monkey: evidence for a distributed neural network subserving spatially guided behavior. Neurosci. 1988;8:4049 – 4068.
43. Blinkberg M, Bonde C, Holm S, Svarer C, Anderson J, Paulson OB, Law
I. Rate dependence of regional cerebral activation during performance of
a repetitive motor task: a PET study. J Cereb Blood Flow Metab. 1996;16:
794 – 803.
44. Boxerman JL, Bandettini PA, Kwong KK, Baker JR, Davis TL, Rose BR,
Weisskoff RM. The intravascular contribution to fMRI signal change:
Monte Carlo modeling and diffusion-weighted studies in vivo. Mag Reson
Med. 1995;34:4 –10.
45. Menon RS, Hu X, Adriany G, Andersen P, Ogawa S, Ugurbil K. Comparison of spin-echo EPI, asymmetric spin-echo EPI and conventional EPI
applied to functional neuroimaging: the effect of flow crushing gradients on
the BOLD signal. In: Proceedings of the Society for Magnetic Resonance.
San Francisco, CA: 1994:622. Abstract.
46. Song AW, Wong EC, Jesmanowicz SG, Tan SG, Hyde JS. Diffusion
weighted FMRI at 1.5 T and 3T. In: Proceedings of the Society of
Magnetic Resonance. Nice, France: 1995:457. Abstract.

Downloaded from http://ahajournals.org by on April 11, 2024