Visual network activation in recovery from sensorimotor stroke.

Abstract

Recovery of finger movements after hemiparetic stroke has been shown to involve sensorimotor brain areas in perilesional and remote locations. Hand use, however, critically depends on visual guidance in such patients with stroke lesions in the middle cerebral artery territory. Using regional cerebral blood flow measurements, we wished to identify interrelated brain areas that are engaged in relation to manual activity in seven patients after their first hemiparetic brain infarction. During the blind-folded performance of sequential finger movements, the patients differed significantly from healthy controls (n = 7) by the recruitment of a predominantly contralesional network involving visual cortical areas, prefrontal cortex, thalamus, hippocampus, and cerebellum. Greater expression of this cortical-subcortical network correlated with a more severe sensorimotor deficit in the acute stage after stroke reflecting its role for post-stroke recovery. Patients also differed from controls on a lesion-related pattern expressed during rest. A third differentiating pattern involved the ipsilesional supplementary motor area and the contralesional premotor cortex. Our results suggest that post-stroke recovery form impaired sensorimotor integration utilizes crossmodal plasticity of a visual network.

Full text

Page 1

Restorative Neurology and Neuroscience 14 (1999) 25–33
0922-6028/99/$8.00 © 1999, IOS Press
RNN71.fm letzte Änderung: 1999-03-08
1. Introduction
Restitution of function after brain damage has been re-
ported to involve adaptive reorganization of both, perilesion-
al and remote neural circuitries [38,47,55]. While early brain
damage is compensated by recruiting atypical cortical areas,
inclusive of the contralateral hemisphere [3,52], motor re-
covery-related plasticity in the adult brain seems to be limit-
ed to the perilesional cortex [9,41,77] and the affected motor
cortical output system [5,22]. Thus, in the adult human brain
restitution of sensorimotor functions appeared to rely on
within system recovery or on the substitutional recruitment
of functionally related systems [59]. Clinical studies, howev-
er, suggest that other functional systems may also come into
play for successful sensorimotor recovery. Specifically, vi-
sual guidance of movement has been shown to contribute
markedly to the post-ischemic restitution of hand function
[2,34].
Motor acts involve a number of specialized and temporal-
ly well structured computations mediating sensorimotor in-
tegration [31,48]. In healthy subjects the internal generation
of voluntary movements necessitates the coherent activation
of motor, somatosensory, cingulate, and subcortical brain ar-
eas [15,46,58,65]. As yet, it is unclear how this essential and
widespread recruitment related to planning, initiation, and
execution of movement is accomplished in patients with im-
paired sensorimotor integration after brain damage. Using
regional cerebral blood flow (rCBF) measurements, we ex-
Correspondence to: Dr. Rüdiger J. Seitz, M.D., Department of Neurology,
Heinrich-Heine-University Düsseldorf, Moorenstraße 5, D–40225 Düssel-
dorf, Telephone: +49-211-81-18974, Fax: +49-211-81-18485, e-mail:
seitz@neurologie.uni-duesseldorf.de
Visual network activation in recovery from sensorimotor stroke
Rüdiger J. Seitz*, Uwe Knorr**, Nina P. Azari, Hans Herzog***,
Hans-Joachim Freund*
*Department of Neurology, Heinrich-Heine-University Düsseldorf, Moorenstraße 5, D–40225 Düsseldorf
**Present address: Department of Neurology, Johann-Wolfgang-Goethe-University Frankfurt, D–60528 Frankfurt
***Institute of Medicine, Research Center Jülich, D–54525 Jülich
Received 19 June 1998; revised 29 September 1998; accepted 29 September 1998
Abstract
Recovery of finger movements after hemiparetic stroke has been shown to involve sensorimotor brain areas in perilesional and remote loca-
tions. Hand use, however, critically depends on visual guidance in such patients with stroke lesions in the middle cerebral artery territory.
Using regional cerebral blood flow measurements, we wished to identify interrelated brain areas that are engaged in relation to manual activity
in seven patients after their first hemiparetic brain infarction. During the blind-folded performance of sequential finger movements, the
patients differed significantly from healthy controls (n = 7) by the recruitment of a predominantly contralesional network involving visual cor-
tical areas, prefrontal cortex, thalamus, hippocampus, and cerebellum. Greater expression of this cortical-subcortical network correlated with a
more severe sensorimotor deficit in the acute stage after stroke reflecting its role for post-stroke recovery. Patients also differed from controls
on a lesion-related pattern expressed during rest. A third differentiating pattern involved the ipsilesional supplementary motor area and the
contralesional premotor cortex. Our results suggest that post-stroke recovery from impaired sensorimotor integration utilizes crossmodal plas-
ticity of a visual network.
Keywords: Stroke, hemiparesis, sensorimotor integration, motor recovery, plasticity, functional imaging

Page 2

26
R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
RNN71.fm letzte Änderung: 1999-03-08
amined functional interactions in the human brain after re-
covery from hemiparetic brain infarction. To this end, we
applied to the rCBF images a principal component analysis
(PCA) that, in contrast to the widely used categorical com-
parisons, does not demonstrate topographic rCBF differenc-
es between groups of subjects or test conditions but rather
distributed systems of functionally connected brain areas
[20]. Specifically, PCA describes without a priori assump-
tion functional interrelations in large data sets, such as rCBF
images, asigning a numerical value to the expression of a
principal component in each subject. Thereby, PCA allows
to formally compare subgroups of patients and test condi-
tions with respect to functional brain systems using inferen-
tial statistics. Here we report that motor activity after hemip-
aretic stroke is accompanied by the abnormal recruitment of
a cortical-subcortical network involving visual cortical ar-
eas. Preliminary data were presented in abstract form [63].
2. Materials and Methods
2.1. Subjects
Seven patients (54 ± 8 (SD) years, 1 female, 6 males)
with their first brain infarction were investigated. Patients
were referred to our clinic because of first completed is-
chemic stroke. Inclusion criteria for this study were acute
hemiplegia or severe hemiparesis with complete loss of frac-
tionated movements of the affected hand, and presence of
only one brain lesion as evident from magnetic resonance
images (MRI) and right handed as assessed with the Eding-
burgh questionnaire [43]. Patients with only moderate or
slight hemiparesis and those who did not recover were ex-
cluded from the study. The study was approved by the Ethics
Committee of the Heinrich-Heine-University Düsseldorf.
Motor impairment was assessed by the Barthel Index [37]
and a multifactorial score that was specifically designed for
examining arm and hand function by separate assessment of
various components contributing to the motor disturbance
[36]. Testing was performed in the acute stage (1–3 days)
and within two days before or after PET scanning. PET
scanning was performed after significant recovery which
was on average six months after brain infarction.
Seven healthy, right-handed volunteers (age range 26 to
29 years, 1 female, 6 males) without a neurological abnor-
mality on history and examination served as controls. Al-
though younger than the patients, they were chosen because
at this age they had no neurological abnormality and present-
ed with a normal MRI of their brains and normal rCBF
scans.
2.2. Lesion assessment
The stroke lesions were outlined in spatially standardized
[60], proton weighted MRI taken at the chronic stage after in-
farction (3 days before or after PET scanning) and plotted in
stereotaxic space [66]. Five patients had right and two had left
hemispheric infarctions. Functions of the cortico-spinal tract
and of the afferent somatosensory tract were determined elec-
trophysiologically and compared to the non-affected ipsile-
sional side as reported in detail elsewhere [5,60]. Specifically,
motor evoked potentials (MEP) were recorded after transcra-
nial magnetic stimulation in the first dorsal interosseus mus-
cles and somatosensory evoked potentials (SEP) after electri-
cal stimulation of the median nerves on the scalp. For both
methods, amplitude ratios were assessed as the percentage of
the amplitude (%) on the affected side in comparison with the
amplitude on the nonaffected side (100 %) as described pre-
viously [5].
3. RCBF-measurements
An eight-ring GE/SCANDITRONIX PC4096 plus PET
camera was used to measure the rCBF following intravenous
bolus injection of [15O]-butanol. This PET camera had an
optimal spatial resolution of 4.9 mm in plane, and a slice dis-
tance of 6.4 mm [50]. A transmission scan using a rotating
68Ge pin source was obtained prior to the emission scans to
correct for attenuation. The 15 PET image slices were recon-
structed with a Hanning filter to an effective image resolu-
tion (FWHM) of 8 mm.
PET scanning started at the time of the intravenous bolus
injection of 40 mCi [15O]-butanol into the right brachial
vein. Quantitation of the rCBF was performed with a com-
bined dynamic-autoradiographic approach [27] using a com-
mon look-up table as detailed elsewhere [67].
3.1. Sensorimotor activation
The patients and controls were blind-folded during the
rCBF measurements. In one rCBF scan, the patients had to
perform finger movement sequences of the recovered hand as
accurately and fast as they could. The control subjects had
to perform the finger movement sequences with their right
hand. Prior to scanning, the patients and controls were in-
structed to sequentially touch the index, long, ring, and little
finger with the thumb of the recovered hand and were trained
until they exactly knew what to do. In a second scan, the pa-
tients performed the same sequence, however, with the unaf-
fected ipsilesional hand. A further scan was a rest condition,
which was taken either as the first or the last scan both in pa-
tients and controls. The sequence of tasks was randomized
across the patients and controls. One of the investigators ob-
served the subjects and registered the number of finger move-
ments.
3.2. Data analysis
The rCBF images were spatially standardized as detailed
elsewhere [60,61]. Standardization yielded 21 axial image
slices that were 6.43 mm apart with a matrix of 128 × 128
pixels, each of 2.55 × 2.55 mm. After image standardization,
the rCBF images obtained during rest and activation in the
patients were compared with those in the healthy control
subjects pixel-by-pixel.
Since the correlated changes in the complex functional
networks of the human brain cannot be assessed by categori-

Page 3

R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
27
RNN71.fmletzte Änderung: 1999-03-08
cal comparisons, we employed a PCA for data evaluation
[20]. The PCA was applied to the rCBF data of the patients
and controls after normalization across groups and condi-
tions, which standardized the variances across the PET im-
ages. The PCA extracted the important features of the cova-
riance matrix in terms of principal components (PCs) or
eigenvectors without a priori assumptions. These vectors
were the linear combinations that accounted for independent
or orthogonal amounts of variance in the observed data.
Only a few PCs are usually required to explain the majority
of observed variance. In terms of functional connectivity, a
principal component (PC) represented a truly distributed
brain system within which there were high intercorrelations.
Because any of these PCs is orthogonal to the remaining,
these systems were functionally unconnected from each oth-
er [20]. The expression of such a pattern in each subject was
given by a single number, the so-called factor score or PC-
value. The coupling of a single brain voxel with such a PC
was given by a number, the PC-load. Thus, a pattern of func-
tional connectivity networks could be visualized as an image
in which each voxel showed the PC-load.
Since each individual subject was given a numerical val-
ue for each PC, groups and conditions were formally com-
pared on these PC-values (two-tailed t-test of independent
means) to determine which pattern of functional connectivi-
ty networks were differentially expressed (significance
p < 0.05). Of these, eight PCs accounted for 80 % of the
variance. PC1, PC3, and PC8 showed significant group dif-
ferences. Constituting areas (loading factors > 0.5) were lo-
calized in stereotaxic coordinates [66]. The loads of the
group-differentiating PCs were displayed in pseudocolor
scale and superimposed onto the spatially standardized MRI
of the patient with the largest stroke lesion. Also, linear re-
gressions of the PC-values and external measures were cal-
culated; R denotes the regression coefficient; p-values corre-
spond to testing if the slope was different from zero.
In addition, an omnibus testing of the mean rCBF differ-
ences between the patients and the controls was performed.
The calculated descriptive t-maps were thresholded at a t-
value of 2.681. To correct for the autocorrelation of adjacent
pixels that is limited by the spatial resolution (FWHM) of
the reconstructed PET images, only clusters of at least 17
spatially contiguous suprathreshold pixels in the PET image
slices were accepted corresponding to p < 0.01 [76]. This de-
scriptive analysis is similar to theoretical cluster analysis ap-
proaches that incorporate the degree of smoothness in the
images [21,75]. In addition, the stroke lesions as defined in
MRI were excluded for the comparison of the PET images
recorded at rest.
4. Results
The patients had all recovered remarkably well from their
first, severely disabling hemiparetic stroke in spite of the in-
Fig. 1: Stroke induced changes. a) Largest extent of the stroke lesions in axial planes dorsal to the intercommissural line (z, mm) in stereotaxic space [66].
b) Impairment after stroke and recovery (p < 0.0001) as assessed by the Barthel Index [37]. c) Persistent reduction (p < 0.02) of somatosensory evoked potentials
(SEP) of the median nerve compared to almost normalized motor evoked potentials (MEP) of small hand muscle after recovery. d) Lesion-related (PC1) pattern
affecting cortical areas and subcortical structures ipsilateral to the infarction (white) and with opposite relation in the contralesional cerebral hemisphere (grey).
e) Positive relation of lesion volume and lesion-related (PC1) pattern.

Page 4

28
R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
RNN71.fm letzte Änderung: 1999-03-08
TABLE I: Group differentiating principal components
PC ConstitutingAreas Loads
Coordinates
Group comparisons Functional Implication
xyz
PC1– Lesion extent –NORM-REST vs. STROKE-REST Lesion-related pattern
Mid. frontal gyrus IL+
−39
− 3
−47
Precentral gyrus CL
−−46
− 5
−32
Precuneus CL
−−12
−53
−38
Post. cingulate gyrus CL
−− 7
−54-16
Mid. occipital gyrus IL
−−25
−85 -06
Mid. occipital gyrus CL
−−32
−82
−12
Basal ganglia CL
−−25
− 6
− 0
Thalamus IL+
−15
−17
− 8
Thalamus CL
−−18
−20
− 4
PC3Cuneus IL+
− 4
−77
−31 NORM-ACTIVATION vs. STROKE-ACTIVATION Recovery-related pattern
Cuneus CL+
− 9
−73
−21
Sup. frontal gyrus IL
−
− 7
−47
− 3
Sup. frontal gyrus CL+
− 7
−50
− 3
Postcentral gyrus CL+
−57
−15
−43
Parietal operculum CL+
−52
−15
−16
Ant. cingulate gyrus CL+
− 5
−25
−25
Hippocalamus CL+
−27
−28
− 4
Lateral thalamus CL+
−18
−20
− 8
Pulvinar thalami CL
−−22
−20
−14
Ant. cerebellum IL
−
−20
−76
−10
Ant. cerebellum CL
−−20
−76
−10
PC8Sup. frontal gyrus IL+
− 6- 2 -43NORM-REST vs. NORM-ACTIVATION
NORM-REST vs. STROKE-ACTIVATIONi
NORM-REST vs. STROKE-ACTIVATION
STROKE-REST vs. STROKE-ACTIVATIONi
Motor learning pattern
Mid.Frontal gryrus CL+
−27 - 6-43
Principal component, PC; coordinates of stereotaxis space (mm)66; Stroke patients,STROKE; Controls, NORM; Resting condition, REST; recovered hand move-
ments, ACTIVATION; non-affected ipsilesional hand movements, ACTIVATIONi; hand movements in controls are ACTIVATION. IL: ipsilesional,
CL: contralesional. Loads: +, loading factors > 0.5; −, loading factors < 0.5. Group separation was significant at p < 0.05, two-tailed t-test of independent means.
Fig. 2: Motor performance. a) Simplified finger movement sequence to be
performed while blindfolded during the rCBF-measurements. b) Reduced
finger movement rate of the recovered (Recover, p < 0.001) and non-affect-
ed ipsilesional (Ipsi, p < 0.01) hand as compared to healthy controls.
volvement of large portions of the motor and somatosensory
cortex around the central sulcus (Fig. 1a). The stroke lesions
also affected the parietal cortex and the striatum but spared
the motor hand area, inclusive of the dorsolateral precentral
gyrus, the internal capsule, and the thalamus. The largest ex-
tent of the stroke lesions is depicted in Fig. 1a. In the acute
stage, the patients were severely disabled by dense hemi-
paresis including absolute inability to move their hand or
fingers. However after recovery, the patients were unsup-
ported ambulatory again and had regained their ability to
perform individual finger movements with the contralesional
hand (Fig. 1b). At the time of PET scanning, the MEPs were
slightly reduced in the recovered hand, while the SEPs were
still significantly affected (Fig. 1c). When performing the
simple finger movement sequences while being blind-fold-
ed, the patients were significantly impaired as compared to
healthy controls in both the recovered contralesional arm, as
well as in the non-affected ipsilesional arm (Fig 2a,b).
To study the brain areas participating in movement activity,
the rCBF was measured. The rCBF images comprised three
functional connectivity patterns that separated the patients
from controls as well as the activation and the rest conditions.

Page 5

R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
29
RNN71.fm letzte Änderung: 1999-03-08
a
b
Fig. 3: Plasticity-related changes. a) Recovery-related pattern (PC3) with
functional interaction of bilateral visual association areas, cingulate cortex,
hippocampal formation, and bilateral cerebellum; red = positive PC-load,
blue = negative PC-load. b) Relation of poor motor score initially after stroke
as assessed by a dedicated motor score [36] and greater expression of the re-
covery-related pattern (PC3) during finger movements of the recovered hand
suggesting compensatory reorganization. c) Motor learning pattern (PC8) in-
volving contralesional dorsolateral premotor cortex and ipsilesional supple-
mentary motor area.
c
Two functional connectivity patterns were differently ex-
pressed in the patients and controls (Table 1). The first pattern
(PC1) was characterized by interactions during the resting
state involving the core and perilesional stroke area, the con-
tralesional motor, medial parietal and lateral occipital cortex,
as well as the bilateral basal ganglia and thalamus. This pattern
was related to lesion volume showing positive effects for ip-
silesional and negative effects for contralesional brain areas
(Fig. 1d,e). That is, the larger the effective brain damage, the
more PC1 was expressed. The categorical comparison of the
rCBF images obtained during rest demonstrated significant
mean rCBF depressions in the patients only in the ipsilesional
cerebral hemisphere. Apart from the infarction they occurred
in the premotor and parietal cortex but spared the thalamus.

Page 6

30
R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
RNN71.fmletzte Änderung: 1999-03-08
Patients and controls also differed on a second pattern
(PC3) expressed during sequential finger movements. This
pattern was characterized by interactions among bilateral
occipital and prefrontal cortical areas, contralesional cingu-
late, hippocampal formation and pulvinar, as well as the bi-
lateral cerebellum (Table 1, Fig. 3a). Greater expression of
this pattern in the patients correlated with lower motor
scores initially after stroke (Fig. 3b). This meant that less
motor impairment was positively related to the cerebellum,
while a severe hemiparesis was related to occipital, parietal,
prefrontal, and hippocampal interactions. Because the pa-
tients were blindfolded during the rCBF-measurements, this
second pattern probably reflected a recovery-related reorga-
nization involving a compensatory recruitment of a visual
network. Comparing the rCBF images obtained during fin-
ger movements in a categorical manner, the healthy subjects
showed a greater mean rCBF only in the contralateral motor
cortex and the ipsilateral anterior cerebellum than the pa-
tients. In contrast, there was no greater mean rCBF in the pa-
tients compared to control.
The third differentiating pattern (PC8) separated move-
ment activity from resting conditions in both, the patients
and controls. This pattern involved the contralesional dorso-
lateral premotor cortex and the ipsilesional frontomesial cor-
tex probably inclusive of the supplementary motor area
(Table 1, Fig. 3c).
5. Discussion
The clinically recovered patients had regained the ability
to generate individual finger movements of the contralesion-
al recovered hand, while their somatosensory cortex was
damaged by the stroke lesion (Fig. 1). Still, the finger move-
ment rate during the blind-folded execution of the simplified
finger movement sequences was reduced in the recovered
arm compared to healthy controls (Fig 2a,b). As became evi-
dent from our study, the stroke lesion had a widespread ef-
fect on the rCBF in the patients. The first pattern successful-
ly identified the stroke lesions explaining the variance in the
image data related to this pathology. Moreover, PC1 differ-
entiated patients and controls reflecting also the spatial ex-
tent and distribution of the remote stroke effects in the af-
fected and contralesional hemisphere [1,60]. While the cate-
gorical comparisons showed that the mean rCBF depressions
in the patients were restricted to the affected cerebral hemi-
sphere, the pattern of functional interrelations revealed that
also the contralesional hemisphere was markedly involved
(Fig. 1d). Interestingly, however, there was no significant
mean rCBF depression in the ipsilesional thalamus which
was in line with the critical role of the thalamus for motor re-
covery after hemiparetic stroke [1,5].
While being blind-folded, the patients not only had a re-
duced finger movement rate but also failed to perform a
more complicated finger movement sequence that they could
perform under visual guidance outside the PET scanner and
that can be performed by naive healthy subjects [58]. Corre-
spondingly, the mean rCBF in the affected motor cortex and
the ipsilateral anterior cerebellum were lower in the patients
compared to controls probably due to the known movement
rate effect on rCBF [6,30,53,56,58,61]. In addition, move-
ment performance was similarly reduced in the non-affected
ipsilesional arm. Both, the residual deficits in the recovered
arm and the non-affected ipsilateral arm were not unexpect-
ed [12,44,74]. Although a slight reduction of finger move-
ment frequency occurs in proportion with age [25], the sig-
nificant impairment in our patients indicated that they suf-
fered a more generalized impairment of sensorimotor
integration. This impairment affected the dominant as well
as the non-dominant hand in our patients (Fig. 1). Therefore,
sensorimotor conception and sensory guidance of movement
normally subserved by the parietal cortex [15,35,58,61,65]
was impaired in our patients. It is tempting to speculate that
the lesion-related rCBF pattern involving both cerebral
hemispheres was the cerebral counterpart to this bilateral be-
havioral deficit. Since sensorimotor guidance was impaired
in our patients, they had to use a new strategy to create a
mental representation of movement to successfully perform
the finger movement sequence task with the recovered hand
during the 60 seconds of the rCBF measurements [4,31]. It is
a well established clinical observation that patients with sen-
sorimotor deficits after cortical lesions and peripheral deaf-
ferentation employ visual guidance for the control of even
simple movements [2,19,29,51]. Since our patients were
blind-folded, they could not directly observe their perfor-
mance but used the mental movement representation for
guidance of the required motor task.
Indeed, our imaging data showed that the patients utilized
a predominantly contralesional network, which may have
promoted the processing, encoding, and imagery of visual
and visuomotor information [14,35,49,61,65,71]. The occip-
ital cortical areas mediate visual information processing in-
duced by both visual stimulation and visual imagery [49,71].
The hippocampus and the prefrontal cortex participate in
memory processes; the former in storage [17,64], the latter
in encoding and retrieval [70]. Further, prefrontal cortical
connections to the cerebellum serve to facilitate practice-in-
duced performance improvement [33,45]. The parietal areas
and the parietal operculum which were involved in the con-
tralesional hemisphere assisted probably in guidance of ac-
curate movements as also evident in healthy subjects
[33,58,61]. Since movements were not executed in relation
to external cues, there was predictably no engagement of the
superior parietal cortex such as Brodmann areas 5 or 7
[15,33,35]. Further, in healthy subjects visual cortical areas
are deactivated during blindfolded movement activity [58].
We therefore propose that the damaged human brain may
engage a dedicated cortical-subcortical network related to
mental visualization of movement to compensate for im-
paired somatosensory-motor integration. Interestingly, our
data also indicated a non-significant trend suggesting that
the older patients demonstrated the greatest motor improve-
ment. Because such a finding is contrary to the evidence for

Page 7

R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
31
RNN71.fm letzte Änderung: 1999-03-08
age-related “impairments” in recovery from stroke [26,69],
one may speculate that our older patients had a “greater” po-
tential than younger patients for plasticity and compensatory
reorganization as reflected in the PC3 pattern.
The third differentiating pattern involved areas in lateral
and medial premotor cortex that participate in normal motor
learning [33,35,58,61,65] and become activated after recov-
ery from striatocapsular infarction [72,73]. In accordance
with recent neuroimaging studies there was a lack of sen-
sorimotor cortex activation in the affected middle cerebral
artery territory [8,13,62]. Due to their corticospinal projec-
tions [16,23], the engaged premotor cortical areas can serve,
however, also as substitutional systems for the damaged mo-
tor cortex. Moreover, the dorsolateral premotor cortex plays
an important role in visuomotor integration [14,35,48,61].
Normally, there is a tight anatomical and functional link be-
tween sensorimotor representations in parietal and premotor
areas [24,32]. In our patients, however, the motor learning
pattern in premotor cortex appeared to be independent from
the recovery-related pattern that subserved the creation of a
visual movement representation.
Recently, network analysis approaches have been shown
to be important complements to categorical analyses of neu-
roimaging data in identifying functionally interrelated brain
systems in both health as well as disease [7,28,42]. Here we
show that PCA revealed functional connectivity patterns
evident in our data which differentiated patients from con-
trols. Specifically, we were able to show that sensorimotor
recovery after hemiparetic stroke seems to utilize cross-
modal plasticity of visual system pathways. In our patients, a
visuomotor brain system appeared to compensate the post-
stroke somatosensory-motor deficit during movement activi-
ty. Such crossmodal plasticity has been observed in animal
models of focal brain lesions and in the developing human
visual and auditory systems [10,11,40, 54]. Our findings
suggest that crossmodal plasticity also may occur in the
sensorimotor domain of the adult human brain. It should be
pointed out that the recruited network utilized anatomical
structures not affected by the stroke lesion. Thus, the interac-
tions did not take place via newly formed structural
connections as shown to occur in experimental models of
postlesional repair in the central nervous system [39,57,68].
6. Acknowledgements
The authors thank W. Hamkens, Institute of Radiochem-
istry, Research Center Jülich, for skillful tracer production
and L. Theelen and L. Tellmann for expert assistance in PET
scanning and PET data pre-processing. N.P. Azari was re-
cipient of an Alexander-von-Humboldt fellowship.
References
[1] ???????????? ?????????????????????????????????????????????????????
??????????????????????????????????????????????????????????????
??????????????????????????????????????????????????????????????
?????????????????????????????????????
[2] Basso, A., Capitani, E., Della Sala, S., Liacona, M., Spinnler, H. Re-
covery from ideomotor apraxia. A study on acute stroke patients.
Brain 110 (1987) 747–760.
[3] Benecke, R., Meyer, B.-U., Freund, H.-J. Reorganization of descend-
ing motor pathways in patients after hemispherectomy and severe
hemispheric lesions demonstrated by magnetic brain stimulation. Exp.
Brain Res. 83 (1991) 419–426.
[4] Berthoz, A. The role of inhibition in the hierarchical gating of execut-
ed and imagined movements. Cog. Brain Res. 3 (1996) 101–113.
[5] Binkofski, F., Seitz, R.J., Arnold, S., Claßen, J., Benecke, R., Freund,
H.-J. Thalamic hypometabolism and corticospinal tract integrity deter-
mine motor recovery in stroke. Ann. Neurol. 39 (1996) 460–470
[6] Blinkenberg, M., Bonde, C., Holm, S., Svarer, C., Andersen, J., Paul-
son, O.B., Law, L. Rate dependence of regional cerebral activation
during performance of a repetitive motor task: a PET study. J. Cereb
Blood Flow Metab. 16 (1996) 794–803.
[7] Büchel, C., Friston, K.J. Modulation of connectivity in visual path-
ways by attention: cortical interactions evaluated with structural equa-
tion modelling and fMRI. Cereb. Cortex 7 (1997) 768–778.
[8] Cao, Y., D'Olhaberriague, L., Vikingstad, E.M., Levine, S.R., Welch,
K.M.A. Pilot study of functional MRI to assess cerebral activation
of motor function after poststroke hemiparesis. Stroke 29 (1998) 112–
122.
[9] Classen, J., Schnitzler, A., Binkofski, F., Werhahn, K.J., Kim, Y.-S.,
Kessler, K.R., Benecke, R. The motor syndrome associated with exag-
gerated inhibition within the primary motor cortex. Brain 120 (1997)
605–619.
[10] Clark, R.E., Delay, E.R. Reduction of lesion-induced deficits in visual
reversal learning following cross-modal training. Restorative Neurol.
Neurosci. 3 (1991) 247–255.
[11] Cohen, L.G., Celnik, P., Pascual-Leone, A., Corwell, B., Falz, L.,
Dambrosia, J., Honda, M., Sadato, N., Gerloff, C., Catala, M.D., Hal-
lett, M. Functional relevance of cross-modal plasticity in blind hu-
mans. Nature 389 (1997) 180–183.
[12] Colebatch, J.G., Gandevia, S.C. The distribution of muscular weak-
ness in upper motor neurons lesions affecting the arm. Brain 112
(1989) 749–763.
[13] Cramer, S.C., Nelles, G., Benson, R.R., Kaplan, J.D., Parker, R.A.,
Kwong, K.K., Kennedy, D.N., Finklestein, S.P., Rosen, B.R. A func-
tional MRI study of subjects recovered from hemiparetic stroke. Stroke
28 (1997) 2518–2527.
[14] Decety, J., Perani, D., Jeannerod, M., Bettinardi, V., Tadardy, B.,
Woods, R., Mazziotta, J.C., Fazio, F. Mapping motor representations
with positron emission tomography. Nature 371 (1994) 600–602.
[15] Deiber, M.-P., Passingham, R.E., Colebatch, J.G., Friston, K.J., Nixon,
P.D., Frackowiak, R.S.J. Cortical areas and the selection of movement:
a study with positron emission tomography. Exp. Brain. Res. 84 (1991)
393–402 .
[16] Dum, R.P., Strick, P.L. The origin of corticospinal projection from the
premotor areas in the frontal lobe. J. Neurosci. 11 (1991) 667–689.
[17] Eichenbaum, H., Otto, T., Cohen, N.J. The hippocampus – what does
it do? Behav. Neural. Biol. 57 (1992) 2–36.
[18] Feeney, D.M., Baron, J.C. Diaschisis. Stroke 17 (1986) 817–830.
[19] Foerster, O. Motorische Felder und Bahnen. In: Bumke, O., Foerster,
O. (eds) Handb. Neurol., Vol. 6, Allgemeine Neurologie, Julius
Springer Verlag, Berlin. (1936) pp 1–357.
[20] Friston, K.J., Frith, C.D., Liddle, P.F., Frackowiak, R.S.J. Functional
connectivity: the principal-component analysis of large (PET) data
sets. J. Cereb. Blood Flow Metab. 13 (1993) 5–14.
[21] Friston, K.J., Worsley, K.J., Frackowiak, R.S.J., Mazziotta, J.C.,
Evans, A.C. Assessing the significance of focal activations using their
spatial extent. Human Brain Mapping 1 (1994) 210–220.
[22] Furlan, M., Marchal, G., Viader, F., Derlon, J.-M., Baron, J.-C. Spon-
taneous neurological recovery after stroke and the fate of the ischemic
penumbra. Ann. Neurol. 40 (1996) 216–226.
[23] Galea, M.P., Darian-Smith, I. Multiple corticospinal neuron popula-
tions in the macaque monkey are specified by their unique cortical ori-
gins, spinal terminations, and connections. Cereb. Cortex 4 (1994)
166–194.

Page 8

32
R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
RNN71.fm letzte Änderung: 1999-03-08
[24] Godschalk, M., Lemon, R.N., Kuypers, H.G.J.M., Ronday, H.K. Corti-
cal afferents and efferents of monkey postarcuate area: an anatomical
and electrophysiological study. Exp. Brain Res. 56 (1984) 410–424.
[25] Hefter, H., Hömberg, V., Lange, H.W., Freund, H.-J. Impairment of
rapid movement in Huntington's disease. Brain 110 (1987) 585–612.
[26] Heiss, W.D., Emunds, H.G., Herholz, K. Cerebral glucose metabolism
as a predictor of rehabilitation after ischemic stroke. Stroke 24 (1993)
1784–1788.
[27] Herzog, H., Seitz, R.J., Tellmann, L., Müller-Gärtner, H.-W. Quantita-
tion of regional cerebral blood flow using an autoradiographic-dynam-
ic approach in positron emission tomography. J. Cereb. Blood Flow
Metab. 16 (1996) 645–649
[28] Horwitz, B., McIntosh, A.R., Haxby, J.V., Furey, M., Salerno, J.A.,
Schapiro, M.B., Rapoport, S.I., Grady, C.L. Network analysis of PET-
mapped visual pathways in Alzheimer type dementia. Neuroreport 6
(1995) 2287–2292.
[29] Hufschmidt, A., Deuschl, G., Lücking, C.H. Motor habits in visuo-
motor tracking: manifestation of an unconscious short-term memory?
Behav. Neurol. 3 (1990) 217–232.
[30] Jäncke, L., Peters, M., Schlaug, G., Posse, S., Steinmetz, H., Iller, G.
Differential magnetic resonance signal changes in human sensori-
motor cortex to finger movements of different rate of the dominant
and subdominant hand. Brain Res. Cogn. Brain Res. 6 (1998) 279–
284.
[31] Jeannerod, M. The representing brain: neural correlates of motor im-
agery and intention. Behav. Brain Sci. 17 (1994) 187–245.
[32] Jeannerod, M., Arbib, M.A., Rizzolatti, G., Sakata, H. Grasping ob-
jects: the cortical mechanisms of visuomotor transformation. Trends
Neurosci. 18 (1995) 314–320.
[33] Jenkins, I.H., Brooks, D.J., Nixon, P.D., Frackowiak, R.S.J., Passing-
ham, R.E. Motor sequence learning: a study with positron emission to-
mography. J. Neurosci. 14 (1994) 3775–3790.
[34] Kalra, L., Perez, I., Gupta, S., Wittink, M. The influence of visual ne-
glect on stroke rehabilitation. Stroke 28 (1997) 1386–1391.
[35] Kawashima, R., Roland, P.E., O’Sullivan, B.T.O. Fields in human mo-
tor areas involved in preparation for reaching, actual reaching, and
visuomotor learning: a positron emission tomography study. J. Neuro-
sci. 14 (1994) 3462–3474.
[36] Kunesch, E., Binkofski, F., Steinmetz, H., Freund, H.-J. The pattern of
motor deficits in relation to the site of stroke lesions. Eur. Neurol. 35
(1995) 20–26.
[37] Mahoney, F.I., Barthel, D.W. Functional evaluation: the Barthel index.
Md. State Med. J. 14 (1965) 61–65.
[38] Merzenich MM, Sameshima K. Cortical plasticity and memory. Curr.
Opinion Neurobiol. 3 (1993) 187–196.
[39] Müller, H.W., Matthiessen, H.P., Schmalenbach, C., Schroeder, W.O.
Glial support of CNS neuronal survival, neurite growth and regenera-
tion. Restorative Neurol. Neurosci. 2 (1991) 229–232.
[40] Newman, A., Corina, D., Tomann, A., Bavelier, D., Braun, A., Clark,
V., Jezzard, P., Neville, H. Effects of age of acquisition on cortical or-
ganization for American sign language (ASL): a functional magnetic
resonance imaging (fMRI) study. Neurosci. Abstracts 23 (1997) 1059.
[41] Nudo, R.J., Wise, B.M., SiFurentes, F., Milliken, G.W. Neural Sub-
strates for the effects of rehabilitative training on motor recovery after
ischemic infarct. Science 272 (1996) 1791–1794.
[42] Nyberg, L., McIntosh, A.R., Cabeza, R., Nilsson, L.-G., Houle, S.,
Habib, R., Tulving, E. Network analysis of positron emission tomogra-
phy regional cerebral blood flow data: ensemble inhibition during epi-
sodic memory retrieval. J. Neurosci. 16 (1996) 3753–3759.
[43] Oldfield, R.C. The assessment and analysis of handednes: the Edin-
burgh inventory. Neuropsychologia 9 (1971) 97–113.
[44] Platz, T., Denzler, P., Kaden, B., Mauritz, K.H. Motor learning after
recovery from hemiparesis. Neuropsychologia 32 (1994) 1209–1223.
[45] Raichle, M.E., Fiez, J.A., Videen, T.O., MacLeod, A.-M.K., Pardo,
J.V., Fox, P.T., Petersen, S.E. Practice-related changes in human brain
functional anatomy nonmotor learning. Cereb. Cortex 4 (1994) 8–26.
[46] Rao, S.M., Binder, J.R., Bandettini, P.A., Hammeke, T.A., Yetkin,
F.Z., Jesmanowicz, A., Lisk, L.M., Morris, G.L., Mueller, W.M., Est-
kowski, L.D., Wong, E.C., Haughton, V.M., Hyde, J.S. Functional
magnetic resonance imaging of complex human movements. Neurolo-
gy 43 (1993) 2311–2318.
[47] Rauschecker, J.P. Compensatory plasticity and sensory substitution in
the cerebral cortex. Trends Neurosci. 18 (1995) 36–43.
[48] Rizzolatti, G., Fadiga, L., Fogassi, L., Gallese, V. The space around us.
Science 277 (1997) 190–191.
[49] Roland, P.E., Gulyas, B. Visual imagery and representation. Trends
Neurosci. 17 (1994) 281–287.
[50] Rota Kops, H., Herzog, H., Schmid, A., Holte, S., Feinendegen, L.E.
Performance characteristics of an eight ring whole body PET scanner.
J. Comput. Assist. Tomogr. 14 (1990) 437–445.
[51] Rothwell, J.C., Traub, M.M., Day, B.L., Obeso, J.A., Thomas, P.K.,
Marsden, C.D. Manual motor performance in a deafferentated man.
Brain 105 (1982) 515–542.
[52] Sabatini, U., Toni, D., Pantano, P., Brughitta, G., Padovani, A.,
Bossao, L., Lenzi, G.L. Motor recovery after early brain damage. A
case of brain plasticity. Stroke 25 (1994) 514–517.
[53] Sadato, N., Ibanez, V., Deiber, M.P., Campbell, G., Cohen, M., Hallett,
M. Frequency-dependent changes of regional cerebral blood flow dur-
ing finger movements. J. Cereb. Blood Flow Metab. 16 (1996) 23–33.
[54] Sadato, N., Pascual-Leone, A., Grafman, J., Ibanez, V., Deiber, M.P.,
Dold, G., Hallett, M. Activation of primary visual cortex by Braille
reading in blind subjects. Nature 380 (1996) 526–528.
[55] Schallert, T., Koslowski, D.A., Humm, J.L., Cocke, R.R. Use-dependent
structural events in recovery of function. Adv. Neurol. 73 (1997) 229–
238.
[56] Schlaug, G., Sanes, J.N., Thangaraj, V., Darby, D.G., Jäncke, L., Edel-
man, R.R., Warach, S. Cerebral activation covaries with movement
rate. Neuroreport 7 (1996) 879–883.
[57] Schnell, L., Schwab, M.E. Rat corticospinal tract fibers regenerate in
the presence of antibodies to myelin-associated neurite growth inhibi-
tors. Restorative Neurol. Neurosci. 2 (1991) 249–250.
[58] Seitz, R.J., Roland, P.E. Learning of sequential finger movements in
man: a combined kinematic and positron emission tomography study.
EJN 4 (1992) 154–165
[59] Seitz, R.J., Freund, H.-J. Plasticity of the human motor cortex. Adv.
Neurol. 73 (1997) 321–333.
[60] Seitz, R.J., Schlaug, G., Kleinschmidt, A., Knorr, U., Nebeling, B.,
Wirrwar, A., Steinmetz, H., Benecke, R., Freund, H.-J. Remote depres-
sions of cerebral metabolism in hemiparetic stroke: topography and re-
lation to motor and somatosensory functions. Hum. Brain Mapping 1
(1994) 81–100.
[61] Seitz, R.J., Canavan, A.G.M., Yagüez, L., Herzog, H., Tellmann, L.,
Knorr, U., Huang, Y., Hömberg, V. Representation of graphomotor tra-
jectories in human parietal cortex: evidence for controlled processing
and skilled performance. EJN 9 (1997) 378–389.
[62] Seitz, R.J., Höflich, P., Binkofski, F., Knorr, U., Tellmann, L., Herzog,
H., Freund, H.-J. The role of the premotor cortex for recovery from mid-
dle cerebral artery infarction. Arch. Neurol. 55 (1998) 1081–1088.
[63] Seitz, R.J., Knorr, U., Azari, N.P., Herzog, H. Recruitment of a visuo-
motor network in recovery from sensorimotor stroke. Neuroimage 7
(1998) S484.
[64] Squire, L.R. Mechanisms of memory. Science 232 (1986) 1612–1619.
[65] Stephan, K.M., Fink, G.R., Passingham, R.E., Silberzweig, D., Cebal-
los-Baumann, A.O., Frith, C.D., Frackowiak, R.S.J. Functional anato-
my of the mental representation of upper extremity movements in
healthy subjects. J. Neurophysiol. 73 (1995) 373–386.
[66] Talairach, J., Tournoux, P. Co-planar stereotaxic atlas of the human
brain, Thieme, Stuttgart, New York, 1988.
[67] Tellmann, L., Herzog, H., Seitz, R.J., Schmitt, G., Müller-Gärtner,
H.W. Simplified quantitation of 15O-butanol rCBF-images based on
single-scan images and a common look-up-table. Neuroimage 5
(1997) S373.
[68] Thanos, S., Vanselow, J. Fetal tectal transplants in the cortex of adult
rats become innervated both by retinal ganglion cell axons regenerat-
ing through peripheral nerve grafts and by cortical neurons. Restor-
ative Neurol. Neurosci. 2 (1990) 63–75.
[69] Toni, D., Fiorelli, M., Bastianello, S., Falcou, A., Sette, G., Ceschin, V.,
Sacchetti, M.L., Argentino, C. Acute ischemic strokes improving during

Page 9

R.J. Seitz et al. / Restorative Neurology and Neuroscience 14 (1999)
33
RNN71.fmletzte Änderung: 1999-03-08
the first 48 hours of onset: predictability, outcome, and possible mech-
anisms. A comparison with early deteriorating strokes. Stroke, 22 (1997)
10–14.
[70] Tulving, E., Kapur, S., Craik, F.I.M., Moscovitch, M., Houle, S. Hemi-
spheric encoding/retrieval asymmetry in episodic memory: positron emis-
sion tomography. Proc. Natl. Acad. Sci. USA, 91 (1994) 2016–2020.
[71] vanEssen, D.C., Anderson, C.H., Felleman, D.J. Information process-
ing in the primate visual system: an integrated systems perspective.
Science, 255 (1992) 419–423.
[72] Weder, B., Knorr, U., Herzog, H., Nebeling, B., Kleinschmidt, A.,
Huang, Y., Steinmetz, H., Freund, H.-J., Seitz, R.J. Tactile exploration
of shape after subcortical ischemic infarction studied with PET. Brain,
117 (1994) 593–605.
[73] Weiller, C., Chollet, F., Friston, K.J., Wise, R.J.S., Frackowiak, R.S.J.
Functional reorganization of the brain in recovery from striatocapsular
infarction in man. Ann. Neurol., 31 (1992) 463–472.
[74] Winstein, C.J., Pohl, P.S. Effects of unilateral brain damage on the
control of goal-directed hand movements. Exp. Brain Res., 105 (1995)
163–174.
[75] Worsley, K.J., Evans, A.C., Marrett, S., Neelin, P. A three-dimensional
statistical analysis for CBF activation studies in human brain. J. Cereb.
Blood Flow Metab., 12 (1992) 900–918
[76] Wunderlich, G., Knorr, U., Stephan, K.M., Tellmann, L., Azari, N.P.,
Herzog, H., Seitz, R.J. Dynamic scanning of 15O-butanol with
positron emission tomography can identify regional cerebral activa-
tions. Human Brain Mapping, 5 (1997) 364–378.
[77] Wunderlich, G., Knorr, U., Huang, Y., Tellmann, L., Herzog, H., Ki-
wit, J.C.W., Freund, H.-J., Seitz, R.J. Precentral glioma location deter-
mines the displacement of cortical hand representation. Neurosurgery,
42 (1998) 18–27.