вход по аккаунту


Functional reorganization of the brain in recovery from striatocapsular infarction in man.

код для вставкиСкачать
Functional Reorganization
of the Brain in Recovery from
Striatocapsular Idarction in Man
Cornelius Weiller, MD, Francois Chollet, MD, Karl J. Friston, MRCPsych, Richard J. S. Wise, MD,
and Richard S. J. Frackowiak, MD
We used positron emission tomography (PET) to study organizational changes in the functional anatomy of the brain
in 10 patients following recovery from striatocapsular motor strokes. Comparisons of regional cerebral blood flow
maps at rest between the patients and 10 normal subjects revealed significantly lower regional cerebral blood flow in
the basal ganglia, thalamus, sensorimotor, insular, and dorsolateral prefrontal cortices, in the brainstem, and in the
ipsilateral cerebellum in patients, contralaterai to the side of the recovered hand. These deficits reflect the distribution
of dysfunction caused by the ischemic lesion. Regional cerebral blood flow was significantly increased in the contralateral posterior cingulate and premotor cortices, and in the caudate nucleus ipsilateral to the recovered hand. During
the performance of a motor task by the recovered hand, patients activated the contralateral cortical motor areas and
ipsilateral cerebellum to the same extent as did normal subjects. However, activation was greater than in normal
subjects in both insulae; in the inferior parietal (area 40), prefrontal and anterior cingulate cortices; in the ipsilateral
premotor cortex and basal ganglia; and in the contralateral cerebellum. The pattern of cortical activation was also
abnormal when the unaffected hand, contralateral to the hemiplegia, performed the task. We showed that bilateral
activation of motor pathways and the recruitment of additional sensorimotor areas and of other specific cortical areas
are associated with recovery from motor stroke due to striatocapsular infarction. Activation of anterior and posterior
cingulate and prefrontal cortices suggests that selective attentional and intentional mechanisms may be important in
the recovery process. Our findings suggest that there is considerable scope for functional plasticity in the adult human
cerebral cortex.
Weiller C, Chollet F, Friston KJ, Wise RJS, Frackowiak RSJ. Functional reorganization of the brain in
recovery from striatocapsularinfarction in man. Ann Neurol 1992;31:463-472
Several mechanisms have been proposed to account
for the clinical recovery of neurological function after
acute ischemic brain injury 11). The potential role of
ipsilateral cortical efferent pathways in recovery has
been most dramatically suggested by the recovery observed in children following surgical hemispherectomy
{Z}. We previously adduced evidence to support this
notion 131. There are further clinical observations that
the process of recovery from ischemic injury is associated with profound changes in the functional organization of the brain. Unilateral hemispheric lesions can
produce bilateral motor deficits 14, 51. Patients with
motor deficits due to unilateral stroke may show associated movements of the contralateral limb when attempting movement of the paretic limb [b}. Furthermore, ischemic lesions of the brain may cause remote
functional effects 17, 81. W e used positron emission
tomography (PET) to assess the extent of the functional effects of unilateral striatocapsular lesions and to
We studied 10 right-handed patients (7 men, 3 women; age
range, 21-62 years; mean, 41 years) with ischemic, strictly
subcortical, striatocapsular infarctions (maximum diameter
2 2 cm) and no other antecedent neurological or significant
general medical history. None had diabetes; 2 were taking
cardiac glycosides for chronic congestive heart failure. No
other lesions were found on high-resolution brain computed
tomography (CT) scans (n = 6) or magnetic resonance imaging (MRI) (n = 4). We included 2 of the patients from
the study of Chollet and colleagues [3}. These patients met
the inclusion criteria of this study. All lesions were assigned
to the left side of the brain (indeed all but two infarcts were
on the left and these were inverted about the midsagittal
plane). Thus, the formerly plegic, now recovered hand was
From the Medical Research Council Cyclotron Unit, Hammersmith
Hospital, London, United Kingdom.
Address correspondence to Prof Frackowiak, MRC Cyclotron Unit,
Hammersmith Hospital, Ducane Road, London W12 OHS, UK.
investigate the changes in brain activity associated with
the performance of a simple motor task, finger opposition, in patients after stroke.
Materials and Methods
Received Jun 3, 1991, and in revised form Sep 17. Accepted for
publication Oct 3, 1991.
Copyright 0 1992 by the American Neurological Association 463
Table 1. Clinical Findings in 10 Patients with Striatocapsular Infarcts
Sex, Age (yr)
M, 26
M, 60
M, 25
F, 50
F, 45
M, 21
M, 22
F, 35
M, 58
M, 62
48 hr
1 yr
4 mo
4 mo
2 mo
2 mo
4 mo
6 yr
9 mo
2 yr
3 mo
4 mo
7 mo
7 mo
2 mo
dLocation of infarcts is shown in terms of the territories involved in structural imaging (MRI in Patients 1,4,6, and 7) according to the literaturt:
[lb]. LL = lateral group of lenticulostriates; LM = medial group of lenticulostriates; AC = anterior choroidal artery.
'Initial motor deficit is graded in two steps: s H P = severe hemiparesis (arm and hand plegic); m H P = moderate hemiparesis (movement
against gravity possible in rhe arm, some movement possible in the hand).
"'Complete" recovery differs from "Good" in that in the latter the task was performed normally but the patients complained of some mild
'Duration until complete recovery.
'Interval between stroke and PET scanning.
'Associated movements of the conrralateral hand when the recovered hand is moved: + = present; 0 = absent (electromyography assessment
in Patients 5 , 6 , 7, and 9 )
on the right side, for analytical purposes. The patients were
investigated 3 months or longer after the occurrence of
strokes that resulted in unilateral hemipareses of acute onset
and of at least 2 days' duration. All experienced very substantial motor recovery, sufficient to allow the performance of a
simple finger opposition task (Table 1). Associated movements of the unaffected hand were carefully evaluated clinically in all patients and formally by electromyography in 4
(see Table 1). We found associated movements in 4 patients.
For example, in 2, movement of the recovered hand was
accompanied by progressive extension of the little finger of
the unaffected hand, which then flexed as the thumb came
back to the index from the little finger during sequential
finger opposition. At the time of PET, none of the patients
had impairment of superficial or deep sensation. Cervical and
transcranial Doppler sonography or selective angiography
showed normal large extracranial and intracranial vessels in
all patients. Ten healthy right-handed volunteers served as
control subjects (6 men, 4 women; age range, 28-69 years;
mean, 47 years). Written informed consent was given by each
participant. The study was approved by the Hammersmith
Hospital ethics committee and permission to administer radioactivity was given by the Administration of Radioactive
Substances Advisory Committee (United Kingdom).
Each patient or normal subject was scanned six times in
the same scanning session. Two scans were performed at
rest; two, during movement of the right hand (the previously
paralyzed hand in the patients); and two, during movement
of the left hand ("unaffected" hand in the patients). The ordering of tasks was balanced (ABCCBA) to avoid habituation
and order effects. All participants were briefly taught to do
the sequential, finger to thumb opposition task. The rate of
opposition was driven by a metronome at three oppositions
every 2 seconds. The metronome sounded during all six
464 Annals of Neurology Vol 31 No 5 May 1992
Scanning and Image Analysis
The subjects were scanned on an ECAT 931-08/12 (CTI,
Knoxville, TN) PET scanner [QJ. Correction for attenuation
was made by performing a transmission scan with an exposed
germanium 68-gallium 68 ("Ge/"Ga) external ring source
at the beginning of each patient study. Dead time corrections
were applied as described previously [9}. Images were reconstructed by filtered backprojection (Hanning filter, cutoff of
0.5) to give an image resolution of 8.5 X 8.5 X 7.0 mm at
full width, half maximum (FWHM). Regional cerebrz! blood
flow (rCBF) measurements (15 transaxial planes) were made
using inhaled oxygen 15-labeled carbon dioxide as a tracer
[lo]. Estimations were completed in 3.5 minutes at 15minute intervals. Data analysis was performed using SPM
software (MRC Cyclotron Unit, London) in PROMATLAB
(Mathworks, US) on Sun 3/60 computers (Sun Computers,
US) using ANALYZE image display software (BRU, Mayo
Foundation, Rochester, MN). The data from each subject
were first standardized for brain size and shape and reconstructed parallel to the intercommissural line El 1-1 31. Each
image was smoothed to account for the variation in normal
gyral anatomy using a gaussian filter (EWHM of 10 pixels).
In the standard stereotactic space, each voxel was 2 x 2 x
4 mm in size. The confounding effect of global differences
in rCBF between scans was removed using an analysis of
covariance [14]. All image analyses were then performed on
a pixel by pixel basis. Mean rCBF maps were derived for
each of the three conditions: at rest and during movements
of right and left fingers. Comparisons of the appropriate differences in condition means were made by the t statistic
using the adjusted pixel error variances for each condition
estimated from the analysis of covariance. The resulting set
of t values constituted a statistical parametric map (SPM).
The omnibus significance of the SPMs was assessed by comparing the observed and expected number of pixels above
p < 0.001 with reference to the Poisson distribution [15].
Significant pixels from SPMs in this omnibus sense were displayed on coronal, sagittal, and transverse views of the brain.
Planned comparisons within each group (normal subjects and
patients) were made between CBF maps derived during finger opposition of the right hand (i.e., the recovered hand) or
the left hand, and the rest states. The rest states were compared between the group of patients and the group of normal
subjects to show the effects of the lesion on function at distant sites. In patients, we also wanted to identify areas in
which the tasks evoked more or less activation than in normal
subjects. For this we compared the changes in rCBF between
baseline and task between groups to identify pixels at which
these changes were significantly different in normal subjects
and patients. The error variance for these comparisons was
pooled from both groups.
To estimate the size, as opposed to the significance of the
changes in rCBF, the adjusted mean rCBFs for the different
conditions in the two groups were compared at locations
identified by maxima in the SPM (i.e., the pixel with the
most significant rCBF change). As the original scans were
smoothed, rCBF in each pixel was an average of 81 pixels
centered on the pixel chosen (corresponding to a brain region
approximately 18 mm2 in the transaxial plane). To rule out
the possibility that the normalization process by itself accounts for increases in rCBF, a post hoc region-of-interest
(ROI) analysis was performed on stereotactically normalized,
t adjusted for global
but unsmoothed images that had n ~been
flow differences. rCBF was calculated in circular ROIs, each
20 pixels wide, centered on the pixel identified by maxima
in the SPM.
Analysis of Functional Connectivity
To investigate further the relationships between the areas
associated with recovery from brain ischemia, correlation
SPMs were generated. These were designed to show whether
there were any areas in which rCBF covaried systematically
with the changes of rCBF in a reference area. This analysis
assumes that the brain is composed of a number of interconnected regions whose activity is specifically perturbed (up or
down) by finger opposition. By choosing a reference area
within the motor system, we sought to identify the remaining
components of the system associated with finger opposition
in a correlation SPM (rSPM) based on rCBF variation across
the three tasks. Our hypothesis was that these networks are
differently constituted in patients with a lesion in the efferent
motor pathways than in normal subjects. Therefore, reference pixels were chosen in areas activated differently in patients and normal subjects, that is, insular cortex, striatum,
parietal cortex (area 40), premotor cortex, anterior cingulate
cortex, and the cerebellum.
Comparisons between Groups at Rest
In patients at rest (Fig I), we found a significantly decreased rCBF in the left striatum and internal capsule
( - 412 mm relative to the anterior commissuralposterior commissural (ACPC] line) corresponding to
the site of the lesion (Table 2). In addition, there were
significant decreases in rCBF in the left insula (4-16
mm above the ACPC line), left primary sensorimotor
(16-28 mm above the ACPC line) and dorsolateral
prefrontal cortices (area 9/10; 16-44 mm above the
ACPC line), left thalamus (0-12 mm above the ACPC
line), left midbrain/cerebral peduncle, and right cerebellum (8-16 mm below the ACPC line). The trend
to a decreased rCBF in the anterior cingulate cortex
(area 24) was not significant. None of these areas appeared to be infarcted on structural imaging.
There were areas in which the rCBF at rest was
higher in patients than in normal subjects. These were
the right caudate nucleus (12-16 mm above the ACPC
line), right angular gyrus (area 39), right premotor cortex (32-48 mm above the ACPC line), and left posterior cingulate cortex (area 23; 8-28 mm above the
ACPC line). A ROI analysis on the unnormalized data
confirmed the rCBF increases in these regions: caudate
nucleus 24.0 +- 5.2 ml/dl/min versus 30.0 -+ 7.2 ml/dl/
min (normals versus patients); premotor cortex, 40.0
5.0 ml/dl/min versus 46.0 t 6.1 ml/dl/min; posterior
cingulate cortex, 50.0 t 5.5 ml/dl/min versus 56.0
6.3 ml/dl/min; and angular gyrus, 40.0
6.3 ml/dl/
min versus 43.0 k 7.8 ml/dl/min.
Comparison of Conditions between Groups
During finger opposition, the nomzal subjects showed
significant rCBF increases in the contralateral primary
sensorimotor cortex (48-60 mm above the ACPC
line), striatum and insula; the ipsilateral cerebellum
(12-20 mm below the ACPC line); and the bilateral
premotor cortices, inferior parietal cortices (area 40),
and the supplementary motor area.
When patients moved the recovered hand, they activated the ipsilateral cerebellum and contralateral sensorimotor and premotor cortices to the same extent as
normal subjects. However, a bilaterally homologous
area comprising middle to anterior aspects of the insula
and the most ventral part of the premotor cortex (with
the maximum of the SPM in the anterior aspects of
the insula, 0-12 mm above the ACPC line) was more
active in patients than normal subjects (Fig 2). In the
patients there was also significantly greater activation
in bilateral parietal cortices (area 40; 16-32 mm above
the ACPC line), lateral prefrontal cortices (area 46;
- 4- + 20 mm relative to the ACPC line), and anterior
cingulate cortices (area 24/32; 16-24 m above the
ACPC line); and the ipsilateral premotor cortex (lateral
area 6 ; 28-36 mm above the ACPC line), striatum
(0-8 mm above the ACPC line), and contralateral cerebellum (12-16 mm below the ACPC line).
When the unaffected hand was moved by the patients, the pattern of cerebral activation was also found
to differ from normal. Patients activated the right insula, striatum, and the lateral prefrontal, premotor, and
inferior parietal (area 40) cortices more than normal
Weiller et al: Functional Reorganization after Striatocapsular Infarction
466 Annals of Neurology Vol 31 No 5 May 1992
Fig I , Comparison of regional cerebral bloodfEow (rCBF) at rest
between 10 patients with striatocapsular infarcts and 10 normdsubjects. The top row represents mean K B F scans averaged
into Talairach’s three-dimensionalspace from the 10 patients,
and in the second row from 10 normal subjects. The numbers
refer to the distance of the plane from the intercommissural
plane. The lower part of the figure shows sagittal, coronal, and
transverse projections of the statistical parametric maps (SPM)
obtained from a group to group comparifon of patients and normal subjects. The left side shows areas with signifcant decreases
in rCBF and the right side, areas with significant increases of
tCBF in the patients. The grid is the standzrd, proportional,
stereotactic grid o f Takzirach and Tounoux { I I ) , which defines
the three-dimensional space into which all the subjects’ and patients’ brain scans have been normalized. The line drawings are
of the horizontal brain contours at the level of the anterior commissure, with the sagittal outline representing the midsagittal
plane. Only pixels that are significantly different between both
groups at p < 0.001 are displayed. The “hot” end of the rainbow scale (distributed over 255 leuels) shows areas of maximally
significant K B F differences and the “col2 end, the threshold of
p < 0.001. In patients, there were significant decreases in
C B F in the left striatum and internal capsule, the insuh, lateral prejrontal and primary sensorimotor cortices, the thalamus,
cerebral pedunclelmidbrain, and right cerebellum. In patients,
there were also significant increases in rCBF in the right caudzte nucleus and premotor cortex and in the ldt posterior cingulate cortex. With the exception of the midbrainlpeduncular region, all pixels displayed lie in gray matter according to the
4 Fig 2. Comparison of differences in activation between 10 patients with striatocapsular infarcts and 10 normal subjects. The
upper part shows projections during movement of the right (recovered) hand and the lower part, during movement of the left
unaffected'^ hand. On the lejt, sagittal, coronal, and transuerse projections of the statistical parametric maps (SPM) obtained by a planned comparison of the differences of changes in
regional cerebral blood $ow (rCBF) between rest and activation
in patients and normal controls are shown. Only pixels that
showed a significantly greater activation in patients at p <
0.001 are shown. The grid and display are as in Figure 1 .
The right ha&shows projections onto the lateral and medial
surfaces of the brain. During movement of the recovered hand,
patients activated an area bilaterally comprising the anterior
part of the insuh and the most ventral part of the premotor cortex, bilateral area 40, lateral prefrontal and anterior cingulate
cortices, ipsilateral premotor cortex and striatum, and contralatera1 cerebellum more than normal. During movement of the “unaffecte2 hand, patients activated the right striatum, insula,
area 40, and lateral prdroontal, anterior cingulate, and premotor
cortices more than normal.
subjects did (see Fig 2). All these areas were on the
right side of the brain, contralateral to the lesion and
contralateral to the moving fingers. In addition, there
was significantly more activation in the anterior cingulate cortex in the patients than in normal subjects.
No areas were activated more in normal subjects
than in patients. However, a greater decrease of rCBF
was recorded between task and rest conditions in the
patients than in the normal subjects in the angular gyms (area 39; 24-32 mm above the ACPC line) and
posterior cingulate cortex.
Quantitative Changes in
Regional Cerebral Blood Flow
Locations with the most significant change in rCBF between task and rest in patients were identified from
maxima in the SPMs. These locations were used to
tabulate the adjusted mean rCBF for each condition in
both groups. The stereotactic coordinates and rCBF
data for the recovered hand are presented in Table 3,
and for the contralateral, unaffected hand in Table 4.
Patients showed a lower rCBF at rest, compared to
normal subjects, in areas in the left hemisphere and in
the anterior cingulate cortex. During the motor task,
the rCBF increased to normal or near-normal levels in
these areas (insula, inferior parietal, dorsolateral prefrontal, and anterior cingulate cortices). rCBF decreased to near-normal levels in the posterior cingulate
cortex and in area 39, both of which showed higher
than normal rCBF at rest. However, the resting rCBF
was higher than normal in the right premotor cortex
(contralateral to the lesion) and rose further with
movement of the fingers of either the recovered or the
opposite hand. All right hemispheric increases were of
the same order whether the recovered or the opposite
hand was engaged in the task.
Analysis of Functional Connectivity
In normal subjects, the same restricted set of areas
showed correlated rCBF changes when the reference
was sited in the premotor cortex, insula, striatum, or
inferior parietal cortex (area 40). These areas were
characterized by variations in rCBF across the three
conditions that correlated strongly with each other in
the same hemisphere and with the ipsilateral primary
sensorimotor cortex and the contralateral cerebellum.
In patients, the premotor cortex, striatum, insula, area
40, and cerebellum showed correlated rCBF changes
bilaterally, and with the anterior cingulate and lateral
prefrontal cortices. Anterior and posterior cingulate
cortices showed an inverse correlation of rCBF in normal subjects and in patients. rCBF changes in these
two areas were not linked to those in the motor areas
in normal subjects.
Weiller et al: Functional Reorganization after Striatocapsular Infarction 467
Table 2. Areas with Significantly Different Regional Cerebral Blood Flow
IrCBF) at Rest in 10 PatirntJ with Striatocapsular Infarctsa
rCBF at Rest
(Adjusted Group Means)
(ml . dl-’. min-’)
Talairach Coordinates
(mm) x, y, z from the
Anterior Commissure
Z Score
L left striatum
L insula
L thalamus
L peduncleimidbrain
L areas 1-4
L dorsolateral
-32, - 6 , 8
-36, 0, 12
- 10, - 2 6 , 4
- 8 , - 2 , -12
-52, -16, 24
prefrontal cortex
R cerebellum
14, -62, - 16
8, 8, 12
24, 2, 48
38, -64, 12
- 4 , -60, 12
R caudate nucleus
R premotor
R angular gyms
(area 39)
L postcingulate
(area 23)
“Adjured group mean rCBF at rest in the areas indicated on the statistical parametric maps to show significant difference in patients (n = 10)
and normal subiecrs (n = 10) (at p < 0.001). Flow values have been normalized by analysis of covariance to a mean flow of 50 m l ’ d l - “
min - ’ . The Talairach coordinates are in mm and correspond to the stereotactic conventions of the atlas of Talairach and Tournow [I I]. The
coordinates refer to the location of maximum significance, with the K B F in each pixel being the mean of 81 pixels centered o n this pixel. The
upper part of the table shows areas with decreased rCBF in patients: The striatum corresponds to the lesion site; the insula, prefrontal and
primary sensorimotor cortices, thalamus, cerebral peduncle/midbrain, and contralateral cerebellum constitute deactivated areas. The lower parr
of the table shows areas of increased rCBF in patients.
In a first approach to assessing brain areas involved in
the recovery process after injury, we compared patterns of cerebral activity associated with movements of
the recovered side with those recorded on movement
of the opposite, “unaffected” hand [ l b ] . In the latter
case, we found a highly lateralized activation pattern.
However, the activation was bilateral, when the recovered hand was moved. We raised the possibility that
the hemisphere without a lesion may not be entirely
normal and independent in such patients, despite the
presence of a single localized and lateralized lesion.
Therefore, in this study we compared the recovered
and the unaffected sides in patients and normal control
subjects. Comparisons between normal subjects and
stroke patients are liable to artificial group effects due
to the differences in brain shape caused by the infarction or atrophy. Therefore, we chose to study subcortical infarcts in the first instance, and took care to
choose a site that was not to be used as a landmark for
the anatomical reorientation and resizing procedures
necessary for averaging data across subjects. We only
recruited patients with striatocapsular infarctions of a
size that would not have major effects on the morpho-
468 Annals of Neurology
Vol 31
No 5
May 1902
logical structure of the brain surface. This type of subcortical infarction is not due to small-vessel disease, but
to the simultaneous occlusion of perforators at the
level of the middle cerebral artery [I 7 J. The patients
were young; therefore, age-related brain changes were
unlikely. Hemodynamic comproniise of uninfarcted
brain areas was also unlikely as all patients were studied
more than 12 weeks after the stroke occurred and all
large extracranial and cerebral vessels were normal.
There was no significant difference in global flow between patients and normal subjects. T o minimize inherent group differences due to possible distortions of
the anatomical structure, we used a new technique that
compared the differences in activation (from rest to
task) between each group.
In patients at rest, the rCBF was decreased in the
left striatum and internal capsule corresponding to the
site of the lesion. In addition, the rCBF was decreased
in the left insula, primary sensoriniotor and dorsolateral prefrontal cortices, thalamus, midbrain/cerebral
peduncle, and contralateral cerebellum. There was a
trend to relative hypoperfusion in the anterior cingulate cortex. None of these areas appeared infarcted o n
structural imaging. In some areas immediately sur-
Table 3. Areas with Significantly Different Changes in Regional Cerebral Blood Flow
(rCBF ) in Patients with Striatocapsular Infarctions during Movement of the Recovered Handa
rCBF (Adjusted G r o u p Means)
(mm) x, Y, z
from Anterior
Z Score
R basal ganglia
R insula, low
28, 6 , 0
40, 4 , 8
38, 32, 16
54, 26, 28
40, - 6 , 36
0, 22, 20
-44, 6 , 4
-40, 36,20
L area 40
L cerebellum
-52, -40,28
-18, -66, -16
L area 39
R area 39
-32, -58, 36
-32, -64, 28
R lateral prefrontal
R area 40
R high premotor
Anterior cingulare
cortex (area 24)
L insula, low
L lateral prefrontal
"Adjusted group mean rCBF during movement of the right hand (the recovered hand in patients) and at rest (averaged from two estimations
for each condition) in the areas indicated on the statistical parametric maps to have shown significantly different rCBF changes in patients than
in normals (at p < 0,001). Flow values have been normalized by analysis of covariance to a mean flow of 50 m l ~ d l - ' ~ m i n - 'The
coordinate conventions are as in Table 2. The top shows areas with a higher increase in rCBF in patients during the task. The bottom shows
greater decreases in rCBF in patients during the task.
Table 4. Areas with Significantly Different Changes in Regional Cerebral Blood Flow
(tCBF) in Patients with Stviatocapsular Infarcts during Movement of the Primarily UnaffectedHanda
rCBF (Adjusted Group Means)
(ml . dl- .min- I )
( m m ) x, Y, 2
from Anterior
Z Score
R insula, low
38, 8, 0
52, -32, 28
40, - 2 , 36
42, 30, 16
0 , 22, 20
-36, -62, 36
3. I
R area 40
R premotor cortex
R lateral prefrontal
Anterior cingulate
L area 39
L postcingulate
- 12, -50, 8
"Adjusted group mean rCBF during movement of the left hand (the hand contralareral to the formerly plegic hand in patients) and at rest
(averaged from two estimations for each condition) in the areas indicated on the statistical parametric maps to have shown significantly different rCBF changes in patients than in normals (at p < 0.001). Flow values have been normalized by analysis of covariance to a mean flow of
50 rn1.dl-l 'rnin-'. The Talairach coordinate conventions are as in Table 2. The upper part of the table shows areas with a higher increase
in rCBF during the task in parients. The lower part (from L area 39) shows greater decreases in patients during the task.
Weiller e t al: Functional Reorganization after Striatocapsular Infarction
rounding the infarct, such as the insula, the hypoperfusion may have been caused by partial neuronal attrition, short of frank infarction, but this cannot be true
for the more distant areas (e.g., dorsolateral prefrontal
cortex) and those lying outside the territory of the
feeding artery (e.g., thalamus). We suggest that the decreased rCBF in these areas is due to functional deactivation, though structural changes at a microscopic level
due to transsynaptic or retrograde degeneration cannot
be excluded. The concept of functional disconnection,
introduced by von Monakow {7) under the term “diaschisis,” supposes that the lesion disrupts part of a neuronal network, of which both the lesion and the deactivated areas are constituent parts. Significant cerebellar
hypometabolism has been observed repeatedly in the
presence of contralateral basal ganglia and internal capsule lesions 118). There are at least five parallel circuits
involving the basal ganglia and the thalamus that have
cortical projections to the primary motor, premotor,
dorsolateral prefrontal, and anterior cingulate cortices
C19, 20). Disruption of these circuits at the level of the
basal ganglia may result in deactivation of other parts
of these circuits. It is interesting that the premotor
cortex, which was not deactivated in our study, has
bilateral connections and indeed the premotor cortex
is activated bilaterally during unilateral motor tasks
12 1, 22). The hypoperfusion in the midbrain/peduncular region can be interpreted as the result of degeneration of the pyramidal tract.
Remote functional effects in the hemisphere contralateral to a lesion have been described and discussed
for many years 123). rCBF studies have revealed a
“transhemispheric depression of metabolism” lasting
for short periods after ischemic infarction 124). A major shortcoming of these studies, however, has been
the lack of appropriate controls {25). A new finding
we report is an increase in rCBF in structures located
at a distance from the primary infarct. It can be argued
that the increases may be due to the normalization
procedure that accounts for differences in global flow.
However, the changes in rCBF were substantiated by
a ROI analysis of the unnormalized data. We postulate
a functional disinhibition of the contralateral premotor
cortex and caudate nucleus by the striatocapsular lesion. This implies that there is a functional inhibition
in these areas under normal conditions, which is determined by contralateral structures. There is indeed
some evidence for the existence of functional inhibition between homotopic regions of both hemispheres
126,271. Both inhibitory and excitatory neurotransmissions require energy expenditure at the synaptic level.
The results of increased activity at an inhibitory synapse are, however, seen as a decrease in blood flow at
the next synapse downstream of it (the inhibited neuron fails to fire, or fires less) 128). At present, the
470 Annals of Neurology
Vol 3 1 No 5
May 1992
explanation postulating impaired inhibition of transhemispheric pathways for the increased rCBF in the
unaffected hemisphere at rest must remain speculative.
We found two major changes in the activation of the
motor system after striatocapsular infarction, namely
the bilateral activation of motor pathways and the recruitment of the inferior parietal cortex and the anterior aspects of the insula. This part of the parietal cortex was labeled area 40 by Brodmann E29). It projects
to the inferior part of the premotor cortex (area 6 )
{30). In turn, there are projections from inferior area
6 to area 4 131). The inferior parietal Cortex also sends
neural projections to the medial parts of the insula. In
monkeys, the insula itself is related to several somatosensory areas, including SI and SII l19). Another connection of the insula (at the junction of the granular
with the dysgranular cortex) is directed to the ventral
portion of area 6 (the precentral operculum), to which
the inferior parietal cortex also projects. The insula and
the parietal cortex (area 40) both showed a covariation
of rCBF with primary and premotor cortices of the
same side and with the contralateral cerebellum in our
normal subjects. Hence, the most ventral parts of area
6 , insula, and parietal cortex constitute a system that is
functionally connected to the normal motor network.
The recruitment of this system after a stroke, in association with motor recovery, is greater than that normally
seen with the execution of a sequential motor task.
This suggests a potential for plastic change within the
cerebral cortex itself after injury and raises the possibility that higher-order areas of sensorimotor cortex have
a greater capacity for reorganization than do primary
areas 132). Activation of this system may be responsible for the associated movements of the unaffected
hand that were observed in 4 of our patients. The relationship between neuronal activity and the laterality of
hand movements is more complex in the ventral part
of area 6 than in the primary sensorimotor cortex. Firing on one side has been observed with bilateral finger
movements in monkeys 133). Associated movements
of the contralateral hand may also be explained by the
activation of bilateral motor pathways (contralateral
motor cortex and ipsilateral cerebellum and ipsilateral
premotor cortex, striatum, and contralateral cerebellum). The modified activation evoked by movement of
the “unaffected” hand in patients may, in part, be a
cortical concomitant of the mild clinical involvement
of the ipsilateral limb that Jones and coworkers 1341
and Colebatch and Gandevia 14) observed in patients
with cerebral infarcts.
The lateral prefrontal cortex, area 39, and the anterior and posterior cingulate cortices are related less
obviously and directly to the executive motor system.
The anterior and posterior cingulate cortices appear to
form an extensive, reciprocally interconnected system
dealing with selective attention or vigilance [35}. The
lateral prefrontal cortex and the angular gyrus (area
39) constitute a brain system concerned with internally
generated, as opposed to stimulus-driven, responses
{36}. We found exactly this pattern of activation in our
patients. It may be assumed that after striatocapsular
infarction, finger opposition to the rhythm of a metronome is no longer a simple, automated, highly overlearned, stimulus-driven task, but requires additional
selective attention and intention. Indeed, choice selection tasks are especially difficult when performed by
patients with left-sided lesions-8 of our 10 patients
had a left hemispheric stroke 137).
Recovery from motor stroke due to comparatively
small striatocapsular lesions is associated with a complex pattern of functional reorganization of the brain
at rest and during performance of a simple motor task.
At rest, deactivated areas and areas with an increased
K B F were found, illustrating the wide distribution of
the networks affected by the localized lesion. Major
mechanisms for the restoration of function may include
bilateral activation of the motor system, with use of
ipsilateral pathways and recruitment of additional motor areas. Changes in areas not normally activated by
finger opposition and usually associated with attentional and intentional states suggest the importance of
such mechanisms in performing relatively simple, sequential, fractionated movements when the pyramidal
tract output pathway is lesioned on one side. Finally,
it appears that remarkable reorganization and plastic
functional changes occur in the human brain after
Dr Weiller is a Feodor Lynen Research Fellow of the Alexander von
Humboldt Foundation, Bonn-Bad Godesberg, Germany. Dr Chollet
is a member of the Department of Neurology of the Hospital Purpan
in Toulouse (Prof A. Rascol) and was partly funded by INSERM.
Dr Friston was funded by the Wellcome Trust.
We thank colleagues at the Medical Research Council Cyclotron
Unit without whose assistance this study would not have been possible.
1. Merill EG, Wall PD. Plasticity of connections in the adult nervous system. In: Cotman CW, ed. Neuronal plasticity. New
York: Raven Press, 1978:97-111
2. Gardner WJ. Removal of the right hemisphere for infiltrating
glioma JAMA 1933;101:823-826
3. Chollet F, DiPiero V, Wise RJS, et al. The functional anatomy
of motor recovery after stroke in humans: a study with positron
emission tomography. Ann Neurol 1991;29:63-71
4. Colebatch JG, Gandevia SC. The distribution of muscular weakness in upper motor neurone lesions affecting the arm. Brain
5. Thilmann AF, Fellows SJ, Garms E. Pathological stretch reflexes
o n the "good" side of hemiparetic patients. J Neurol Neurosurg
Psychiatry 1990;53:208-214
6. Marie P, Foix C. Les syncinesies des hemiplegique: etude semiologique et classification. Rev Neurol (Paris) 1916;29:3-27
7. von Monakow C. Die Lokalisation im Grobhirn und der Abbau
der Funktion durch Kortikale Herde. Wiesbaden: Bergmann,
8. Feeney DM, Baron JC. Diaschisis. Stroke 1986;17:817-830
9. Spinks TJ, Jones T, Gilardi MC, Heather JD. Physical performance of the larest generation of commercial positron scanner.
IEEE Trans Nucl Sci 1988;35:721-725
10. Lammertsma AA, Cunningham VJ, Deiber MP, et al. Combination of dynamic and integral methods for generating reproducible functional CBF.images. J Cereb Blood Flow Metab 1989;
11. TalairachJ, Tournoux P. Coplanar stereotaxic atlas of the human
brain. New York: Georg Thieme Verlag, 1988
12. Friston KJ, Passingham RE, N u n JG, et al. Localisation in PET
images: direct fitting of the intercommissural (AC-PC) line. J
Cereb Blood Flow Metab 1989;9:690-695
13. Friston KJ, Frith CD, Liddle PF, Frackowiak RSJ. Plastic transformation of PET images. J Comp Assist Tomogr 1991;15:
14. Friston KJ, Frith CD, Liddle PF, et al. The relationship between
global and local changes in PET scans. J Cereb Blood Flow
Metab 1990;10:458-466
15. Friston KJ, Frith CD, Liddle PF, Frackowiak RSJ. Comparing
functional (PET) images: the assessment of significant change. J
Cereb Blood Flow Metab 1991;11:690-699
16. Perlow MJ, Freed WJ, Hoffer BJ, et al. Brain grafts reduce
motor abnormalities produced by destruction of nigro-striatal
dopamine system. Science 1979;204:643-647
17. Weiller C, Ringelstein EB, Reiche W, et al. The large striatocapsular infarction. A clinical and pathophysiological entity. Arch
Neurol 1990;47:1085- 1091
18. Baron JC, Bousser MG, Comar D, Castaigne P. Crossed cerebellar diaschisis in human supratentorial brain infarction. Trans
Am Neurol Assoc 1981;105:459-461
19. Mesulam MM, Mufson EJ. The insula of Reil in man and monkey: architectonics, connectivity, and function. In: Peters A,
Jones EG, eds. Cerebral cortex: association and auditory cortices. New York: Plenum Press, 1985:179-226
20. Alexander GE, Crutcher D. Functional architecture of basalganglia circuits: neural substrates of parallel processing. Trends
Neurosci 1990;13:266-27 1
21. Roland PE, Skinhoj E, Lassen NA, Larsen B. Different cortical
areas in man in organisation of voluntary movements in extrapersonal space. J Neurophysiol 1980;43:137-150
22. ColebatchJG, Deiber MP, Passingham RE, et al. Regional cerebral blood flow during voluntary arm and hand movements in
human subjects. J Neurophysiol 1991;65:1392-1401
23. Kempinsky WH. Spatially remote effects of focal brain injury.
Relation to diaschisis. Trans Am Neurol Assoc 1956;81:79-82
24. Hoedt-Rasmussen K, Skinhoj E. Transneural depression of the
cerebral hemispheric metabolism in man. Acta Neurol Scand
25. Wise RJS, Gibbs JM, Frackowiak RSJ, et al. N o evidence for
transhemispheric diaschisis after human cerebral infarction.
Stroke 1986;17:853-860
26. Calford MB, Tweedale R. Interhemispheric transfer of plasticity
in the cerebral cortex. Science 1990;249:805-807
27. Ferbert A, Priori A, Rothwell JC, et al. Trans-callosaleffects on
motor cortical excitability in man. J Neurophysiol 1990;429:38P
28. Schwartz WJ, Smith CB, Davidsen L, et al. Metabolic mapping
of functional activity in the hypothalamo-neurohypophysealsystem of the rat. Science 1979;205:723-725
29. Brodmann K. Vergleichende Lokalisationslehre der Grobhirnrinde. 2nd ed. Leipzig: Barth, 1925
Weiller et al: Functional Reorganization after Striatocapsular Infarction
47 1
30. Cavada C, Goldman-Ebkic PS. Posterior parietal cortex in rhesus monkeys: evidence for segregated corticocortical networks
linking sensory and limbic areas with the frontal lobe. J Cornp
Neurol 1989;287:422-445
'7 1. Strick P. How do the basal ganglia and cerebellum gain access
to the cortical motor areas? Behav Brain Res 1985;18:107-124
32. Pons TP, Garraghty PE, Mishkin M. Lesion-induced plasriciry
in the second somatosensory cortex of adult macaques. Proc
Natl Acad Sci USA 1988;85:5279-5281
33. Tanji J, Okano K, Sat0 KC. Neuronal activity in cortical motor
are& related to ipsilateral, contralateral, and bilateral digit movements of the monkey. J Neurophysiol 1988;80:325-343
Annals of Neurology
Vol 31
No 5
May 1992
34. Jones RD, Donaldson IM, Parkin PJ. Impairment and recovery
of ipsilateral sensory-motor function following unilateral cerebral infarction. Brain 1989;112:113-132
35. Posnet MI, Peterson SE. The attention system of the human
brain. Annu Rev Neurosci 1990;11:25-42
36. Frith CD, Ftiston K, Liddle PF, Frackowiak RSJ. Milled action
and the prefrontal cortex in man: a srudy with PET. Proc IR SOC
Lond [Biol] 1991;244:241-246
37. Sturm W, Bussing A. EinAuh der Aufgabenkomplexitit aui
hirnorganische R e a k t i o n s b e e i n t r ~ h t i ~ ~ e n - H i r n s c h a d s oder Patienteneffekc? Eur Arch Psychiatry Neurol Sci 1986;
235:2 14-220
Без категории
Размер файла
1 664 Кб
reorganization, infarction, recovery, brain, man, function, striatocapsular
Пожаловаться на содержимое документа