64, 215230 (1998)

BL981961

BRAIN AND LANGUAGE

ARTICLE NO.

Brain Plasticity in Poststroke Aphasia: What Is the

Contribution of the Right Hemisphere?
Hans Karbe,* Alexander Thiel, Gerald Weber-Luxenburger,* Karl
Herholz,* Josef Kessler, and Wolf-Dieter Heiss*,
*Department of Neurology, University Hospital, and the Max Planck Institute for
Neurological Research, Cologne, Germany.
The brain may use two strategies to recover from poststroke aphasia: the structural
repair of primarily speech-relevant regions or the activation of compensatory areas.
We studied the cortical metabolic recovery in aphasic stroke patients with positron
emission tomography (PET) at rest and during word repetition. The left supplementary motor area (SMA) showed the most prominent compensatory activation in the
subacute state of stroke. The restitution of the left superior temporal cortex determined the long-term prognosis of aphasia. The brain recruited right-hemispheric
regions for speech processing, when the left-hemispheric centers were permanently
impaired. This strategy, however, was significantly less effective than the repair of
the original speech-relevant network. 1998 Academic Press

INTRODUCTION

PET studies of aphasic stroke patients showed a significant correlation

between the neuropsychological deficit and the metabolic impairment of
speech-relevant areas in the left cerebral hemisphere (Metter, Kempler, Jackson, Hanson, Mazziota, & Phelps, 1989). An especially close relationship
was found between the left superior temporal metabolic impairment and the
receptive language disorder as measured with the Token test (De Renzi &
Vignolo, 1962; Karbe, Herholz, Szelies, Pawlik, Weinhard, & Heiss, 1989).
Studies of aphasia outcome additionally pointed out that the degree of regional left hemispheric hypometabolism as measured early after stroke predicts the long-term prognosis of aphasia (Karbe, Kessler, Herholz, Fink, &
Heiss, 1995a). However, the restitution of the left hemispheric speech regions may not be the only means by which the brain can cope with strokecaused impairments of language. PET activation studies of aphasic stroke
Reprint requests should be addressed to W. D. Heiss, MD, at the Max-Planck-Institut fur
neurologische Forschung, Gleueler Strasse 50, 50931 Koln, Germany. Fax: xx49 221 4726
298.
215
0093-934X/98 $25.00
Copyright 1998 by Academic Press
All rights of reproduction in any form reserved.

216

KARBE ET AL.

patients gave evidence for a compensatory potential of the right hemisphere

in some patients with severe left hemispheric damage (Weiller, Isensee, Rjintjes, Huber, Muller, Bier, Dutschka, Woods, North & Diener, 1995). This
compensatory potential presumably relies on a preexisting network of speech
areas which involves both cerebral hemispheres. A preceding investigation
of healthy individuals which used the identical methodology of cortical activation as the present study confirmed this hypothesis by finding a network
of significant correlations which connects cortical areas of both cerebral
hemispheres during speech processing (Karbe, Herholz, Weber, Luxenburger, Ghaemi, & Heiss, in press). The present study focuses on the question
how the aphasic brain can execute basic language tasks such as single word
repetition despite the stroke-caused functional impairment of this speechrelevant network. The study therefore analyzes the task-induced metabolic
activation in key regions of language processing of both hemispheres in the
subacute state of stroke in comparison with the activation pattern in normal
individuals. Follow-up studies which were carried out in the second year
after stroke give additional information about the later reorganization of the
speech-relevant network, especially with respect to the contribution of the
right cerebral hemisphere to the long-term outcome of aphasia.
PATIENTS AND METHODS
We studied 12 consecutively admitted patients (7 men, 5 women, mean age 57, range 34
78 years) who had acute aphasia resulting from a single ischemic stroke in the territory of
the left middle cerebral artery. A history of stroke, preexisting language disorders, the complete
inability to repeat single words, or severe medical diseases were exclusion criteria. We compared the findings of the stroke patients with a control group of 10 normal subjects (9 men,
1 woman, mean age 39 years, range 2568 years). The normal subjects were recruited mainly
from the staff of the PET laboratory so that the control group was not balanced for age and
gender. This selection of control subjects, however, had the advantage of only including presumably healthy individuals without preexisting ischemic lesions. All patients and control
subjects gave their informed consent to participate in the study. A laterality questionnaire
(Oldfield, 1971) and, whenever possible, standardized tests of skilled motor performance classified all patients and control subjects as right-handed.
MRI was performed in all patients and control subjects with a superconducting 1-T instrument (Magnetom, Siemens, Erlangen, Germany). The MRI scanner produced 64 adjacent
transaxial T1-weighted tomograms with a slice thickness of 2.5 mm and a pixel size of .98
.98 mm. PET was performed in all individuals with a high-resolution scanner (ECAT EXACT HR, Siemens-CTI, Knoxville, TN), which provided 47 contiguous transaxial slices
through the brain with a center-to-center distance of 3.125 mm and with a spatial resolution
(full width-half maximum of point-spread function) of 3.6 mm transaxially and 4 mm axially
(Wienhard, Dahlbom, Eriksson, Michel, Pietrzyk, & Heiss, 1994). After bolus injection of
185-MBq FDG, six frames of 10 min each were recorded. Images were reconstructed on
a 128 128 matrix with 2.17-mm pixel size by applying a Hanning filter with a cutoff frequency of .4-cycle per pixel. This procedure included corrections for random coincidences, scatter and attenuation; with regard to attenuation a threshold algorithm was used to
define the contour of the head. Images of the regional cerebral metabolic rate of glucose were
calculated based on frames 3 to 6 and multiple arterialized blood samples with adjustment of
rate constants K 1 and k 3 to measured activity (Wienhard, Pawlik, Herholz, Wagner, & Heiss,
1985).

BRAIN PLASTICITY IN APHASIA

217

Two PET studies were performed on each individual within 1 week. In the first PET study,
the individuals were examined in a resting condition, with eyes closed and ears unplugged in
a room with low ambient noise (Heiss, Pawlik, Herholz, Goldner, & Weinhard, 1984). The
second PET study was performed under the same condition but during verbal stimulation.
The subject had to repeat German nouns which one of the investigators randomly presented
aloud. The total number of presented nouns was adapted to the individuals repetition pace.
The total number of repeated nouns was 751 (SD 295) words in stroke patients and 1224 (SD
242) words in control subjects. There was no significant correlation between the number of
repeated words and the regional metabolic activation within the groups of patients and normal
controls. The 30-min period of stimulation started immediately before[18 F] 2-fluoro-2-deoxyD-glucose injection. The MRI and PET images were coregistered in parallel to the intercommissural line as defined on the three-dimensional MRI display by using the multipurpose
matching program. The uncertainties from the coregistration were found to be less than 1.5
mm in any direction (Pietrzyk, Herholz, Fink, Jacobs, Mielke, Slansky, Wurker, & Heiss,
1994).
The original three-dimensional registered MRI and PET images were interpolated to 1 mm 3
voxels. Sagittal, coronal, and transaxial 1-mm slices were created based on the 1 mm 3 voxels.
A software program which was written in house allowed a simultaneous, three-dimensional
display of MRI and PET images. Anatomical regions of interest (ROIs) were placed subsequently on several adjacent MRI slices in one plane. The software program automatically
superimposed the ROIs on the corresponding PET images and simultaneously displayed them
on the two other planes so that the three-dimensional extent of the ROIs could immediately
be seen (Herholz, Dickhoven, Karbe, Halber, Pietrzyk, & Heiss, 1996). The ROIs included
the classical regions of language processing of both cerebral hemispheres, i.e., the inferior
frontal speech area (Brodmanns areas (BA) 44 and 45), the superior temporal speech area
(Heschls gyrus, planum temporale, BA 22), the SMA and the vocalization areas of the precentral gyrus. The ROI boundaries were defined according to the individuals brain anatomy:
1. The inferior frontal speech region included the convolution anterior and posterior to the
anterior ascending ramus of the Sylvian fissure (Foundas, Leonard, & Heilmann, 1995). In
comparison with the standardized maps of Brodmann (1909), these ROIs were identified as
BA 45 and BA 44, respectively.
2. The superior temporal speech region included Heschls gyrus, the planum temporale
which was defined according to the criteria of Steinmetz, Rodemacher, Huang, Hefter, Zilles,
Thron, and Freund (1989), and BA 22. The anterior and posterior borders of BA 22 coincided
with the anterior border of Heschls gyrus and with the posterior border of the planum temporale, respectively. The medial border was defined by the medial fold of the superior temporal
sulcus; the lateral border stretched to the lateral surface of the superior temporal gyrus.
3. The SMA was defined as the part of BA 6 that was located within the longitudinal fissure,
anterior to the motor cortex for foot movements (Rademacher, Galaburda, Kennedy, Filipek, &
Cariness, 1992).
4. The vocalization area of the motor cortex included the inferior half of the precentral
gyrus. The precentral gyrus was outlined following the central sulcus from its origin above
the Sylvian fissure to its entrance into the longitudinal fissure. It was then divided into its
superior and inferior halves. The inferior half consequently included the region where face,
mouth, tongue, and throat have their motor representation.
All ROIs were drawn on parasagittal slices. The regional metabolic increase was calculated
as previously published (Karbe, Wurker, Herholz, Ghaemi, Pietrzyk, Kesster, & Heiss, 1995b).
The infarct volume and location were determined in all patients based on the MRI images.
The initial PET measurements in patients were done 34 weeks after stroke (mean 24 days,
SD 11 days). All patients were tested within the same week with an extensive neuropsychological test battery which included the Token test (De Renzi & Vignolo, 1962). Seven patients
with severe aphasia agreed to have follow-up PET studies and neuropsychological tests more
than 1 year after stroke (mean 19 months, SD 2 months). The commercially available software
package SPSS for Windows, Version 6.0, was used for all statistical calculations.

218

KARBE ET AL.

TABLE 1
Mean Metabolic Changes during Single Word Repetition in 12 Aphasic Patients and in 10
Normal Individuals (in Percentage with Standard Deviation and p According to t Test for
Paired Samples
Aphasic patients
%
Inferior frontal
BA 45
BA 44
Superior temporal
Heschls gyrus
Planum temporale
BA 22
Supplementary motor
Inferior precentral

2.9
3.3
2.4
.9
.8
1.8
.8
4.8
.8

Inferior frontal
BA 45
BA 44
Superior temporal
Heschl gyrus
Planum temporale
BA 22
Supplementary motor
Inferior precentral

3.0
2.5
3.4
8.3
8.4
5.1
9.0
4.3
5.6

SD

Normal individuals
p

Left hemisphere
4.4
.04*
4.9
.04*
5.3
.15
12.0
.80
13.5
.84
15.5
.70
11.0
.82
3.4
.001***
5.4
.62
Right hemisphere
5.1
.07
5.3
.13
5.5
.06
4.4
.001***
4.2
.001***
5.3
.007**
5.4
.001***
4.8
.01**
5.4
.004**

SD

2.1
1.9
2.3
8.2
8.5
8.8
8.1
.6
6.6

3.0
3.4
3.0
3.1
4.2
3.7
3.1
3.5
2.0

.05*
.11
.04*
.001***
.001***
.001***
.001***
.58
.001***

1.1
.9
1.2
9.4
8.3
7.4
9.8
1.3
4.5

3.0
3.2
3.3
2.7
2.6
4.9
3.3
3.0
4.4

.96
.40
.26
.001***
.001***
.001***
.001***
.19
.01**

* p .05.
** p .01.
*** p .001.

RESULTS

Six patients had infarcts that included the left superior temporal cortex
(mean infarct size 68 cm 3, range 27133 cm 3). Five of these six patients had
global aphasia (between 47 and 41 errors in Token test), and one patient had
severe Wernickes aphasia (32 errors). Three patients had cortical infarcts
outside the left superior temporal cortex (one inferior frontal infarct of 61
cm 3 causing global aphasia with 47 errors in Token test; one parietal infarct
of 63 cm 3 causing global aphasia with 44 errors; one precentral infarct of
10 cm 3 causing anomic aphasia with 5 errors). The remaining three patients
had subcortical infarcts (size 52, 13, and 2 cm 3) which caused global aphasia
in the patient with the large infarct (42 errors), and mild, anomic aphasia in
the two patients with the smaller infarcts (9 and 3 errors).
Table 1 shows the mean regional metabolic changes during word repetition in the aphasic patients and in the normal controls. The regional metabolic
activity significantly increased in several speech-relevant regions. Some nor-

BRAIN PLASTICITY IN APHASIA

219

mal individuals and aphasic patients showed decreases of regional metabolic

activity during word repetition. These decreases were generally mild and did
not reach the level of significance. A reduced or even missing metabolic
activation of the left superior temporal cortex corresponded to the functional
impairment of the posterior speech region in the aphasic group. In contrast
to the normal individuals, the aphasic patients additionally activated both
SMAs, especially on the left side. A t test for independent samples confirmed
that the mean left SMA metabolic increase was significantly higher in the
aphasic patients in comparison with the control subjects ( p .01). Some
aphasic patients, especially those with severe aphasia, also activated the right
inferior frontal cortex in the subacute state of poststroke aphasia (Fig. 1).
This activation of right inferior frontal areas, however, was not as consistent
as the SMA activation so that the mean metabolic increase of all aphasic
patients did not reach the level of significance.
The seven severely aphasic patients who agreed to be reexamined more
than 1 year poststroke included five patients with large left perisylvian infarcts, the patient with an inferior frontal infarct, and one patient with a
capsulostriatal infarct. The comparison between the initial and the followup PET studies showed quite interesting changes of the activation patterns
over time. The additional activation of right hemisphere regions, such as the
right SMA or the right inferior frontal cortex had disappeared in the followup studies in some patients while the left-hemispheric superior temporal activation had recovered (Fig. 2). We therefore compared the Token test scores
with the regional metabolic activation, separately for the initial and the follow-up examinations (Table 2). The number of errors of the initial Token
test did not significantly correlate with the regional metabolic increase of
the initial PET studies. The comparison between the follow-up Token test
scores and the follow-up PET studies, however, demonstrated significant direct correlations between the number of test errors and the regional metabolic
increase of the left SMA and several right hemispheric regions. The closest
correlation was found between the right SMA and the Token test score (p
.001; Fig. 3). This direct correlation meant that a permanent additional
metabolic activation of the right SMA especially occurred in patients who
had poor recovery and a persisting language deficit.
Interestingly, the left superior temporal region and the Token test were
inversely correlated in the follow-up examinations. This correlation, however, did not reach the level of .05 significance. The negative relationship
became more evident when the functional recovery of the left superior temporal region was regarded, i.e., the difference of the follow-up minus the
initial metabolic increase (Table 3). This comparison demonstrated a highly
significant correlation between the functional recovery of the left superior
temporal cortex and the recovery of language, as measured with the error
rate of the follow-up Token test (Fig. 4). Furthermore, the recovery of the
left superior temporal areas was paralleled by less metabolic increase in all

220

KARBE ET AL.

BRAIN PLASTICITY IN APHASIA

221

222

KARBE ET AL.

TABLE 2
Pearson Correlation Coefficients between the Regional Metabolic Increase
during Word Repetition and Token Test Scores (Number of Errors) in Seven
Patients with Severe Aphasia Who Had Follow-up Studies
Initial PET
and initial
Token test

Left inferior frontal

Left superior temporal
Left supplementary motor
Left inferior precentral
Right inferior frontal
Right superior temporal
Right supplementary motor
Right inferior precentral

Follow-up PET and

follow-up Token test

.31
.20
.16
.12
.19
.25
.19
.28

.50
.66
.74
.79
.68
.59
.69
.54

.61
.72
.80
.32
.71
.78
.97
.87

.14
.07
.03*
.49
.07
.04*
.001***
.01**

right hemisphere regions, especially in the right superior temporal cortex

(Table 3; Fig. 5).
DISCUSSION

The purpose of this study was to investigate how the speech-relevant cerebral network copes with stroke-caused impairments of language, and in particular, to what degree the right hemisphere contributes to language processing after left hemispheric strokes. Because we studied unselected aphasic
patients, except for the fact that they had to be able to repeat single words
to some degree, we had to use a neuropsychological test which was suitable
for measuring language impairment in aphasias of different types and severities. The Token test that we used in this study meets this requirement

FIG. 1. MRI (first row), FDG-PET at rest (second row), and FDG-PET during word repetition (third row) of a 41-year-old man with global aphasia (Token test 47 errors) 2 weeks after
stroke. Transaxial (first column), coronal (second column), and parasagittal (third column)
through the brain. The cross-hairs mark the left Heschls gyrus. The morphological infarct
zone reaches the posterior bottom of the left Sylvian fissure on the parasagittal MRI scan.
The color scales give the regional cerebral glucose metabolism in mol/100 g per min. Each
scale is adapted to the level of global cerebral metabolism in order to allow comparison between the regional color patterns at rest and during repetition. The arrowhead marks the additionally activated right inferior frontal cortex.
FIG. 2. Thirteen-month follow-up study of the same patient as in Fig. 1; same arrangement
of images. The patient has recovered from aphasia fairly well. The left superior temporal
cortex now shows some metabolic activation during word repetition. The additional activation
of the right inferior frontal cortex has disappeared.

223

BRAIN PLASTICITY IN APHASIA

FIG. 3. Correlation plot between the age-corrected Token test scores and the metabolic
increase of the right SMA in the long-term follow-up studies. The number of errors is directly
correlated with metabolic activation of the right SMA.

TABLE 3
Pearson Correlation Coefficients of the Left Superior Temporal Recovery
(Metabolic Increase of Follow-up PET Minus Initial PET) with the Token Test
Scores and with the Metabolic Increase in Right Hemispheric Regions

Token test
Right inferior frontal
Right superior temporal
Right supplementary motor
Right inferior precentral

Initial study

Follow-up study

.60
.13
.23
.05
.48

.15
.78
.61
.92
.27

.91
.82
.94
.86
.81

.005**
.02*
.002**
.01**
.03*

224

KARBE ET AL.

FIG. 4. Correlation plot between the age-corrected Token test scores in the follow-up
studies and the left superior temporal functional recovery (metabolic increase of follow-up
PET studies minus metabolic increase of initial PET studies). A left superior temporal recovery
corresponds to a good outcome of aphasia.

(Swisher & Sarno, 1969; Gallaher, 1979). The Token test reliably diagnoses
even subtle or latent aphasic disturbances (Orgass & Poeck, 1966) and it is
equally good in evaluating language comprehension deficits in fluent and
nonfluent aphasias (Zaidel, 1979).
Lesion localization studies with computer tomography have pointed out
that the crucial area for auditoryverbal comprehension, as measured with
the Token test, is located between the middle and posterior third of the first
temporal gyrus (Vignolo, 1979). Interestingly, some language comprehension deficit, however, occurs almost independently from a specific lesion
location in all types of aphasia, also in pure motor aphasias (De Renzi &
Vignolo, 1962). This receptive language disturbance which the Token test
can detect is due to remote neural impairment of the left superior temporal
cortex which occurs also in aphasic patients with left frontal or subcortical
lesions (Karbe et al., 1989, 1995a). It was therefore advantageous to use a
language activation task such as single word repetition from which we already knew that it significantly activated the superior temporal areas in normal individuals (Howard, Patterson, Wise, Brown, Friston, Weiller, Fracko-

BRAIN PLASTICITY IN APHASIA

225

FIG. 5. Correlation plot between the metabolic increase of the right superior temporal
cortex in the follow-up studies and the left superior temporal functional recovery (metabolic
increase of follow-up PET minus metabolic increase of initial PET). A persisting left superior
temporal impairment corresponds to a compensatory right superior temporal activation.

wiak, 1992; Karbe et al., 1995b) and which also patients with major language
impairments could execute.
In the subacute state of aphasia, PET revealed stroke-caused impairments
of metabolic activation within the left cerebral cortex, as expected. The left
superior temporal and the left precentral metabolic activations were significantly reduced or even lost because of the ischemic brain damage, whereas
the corresponding right superior temporal and precentral areas showed the
typical activation pattern due to the ongoing auditory and motor processing
in the right hemisphere. In comparison with the healthy control subjects, the
aphasic stroke patients, however, activated additional, presumably compensatory cortical areas in both cerebral hemispheres. The most consistent of
these initial additional activations were found in the SMAs, more prominently on the left than on the right side. Furthermore, some patients with
severe aphasia clearly activated the right inferior frontal cortex, which is the
right hemispheric counterpart of Brocas area.
The language competence of the left SMA is well established, especially

226

KARBE ET AL.

since the basic intraoperative studies of Penfield and coworkers (Penfield &
Roberts, 1959). Finding speech-induced activation of the left SMA therefore
was not surprising or completely new. Several language activation studies
with PET also confirmed that the SMAs significantly contribute to language
processing in the brain. Subjects who had to speak aloud auditorily presented
stimuli showed a bilateral activation of the SMAs (Petersen, Fox, Posner,
Mintun, & Raichle, 1988) and the left SMA was even activated during silent
verb generation (Wise, Chollet, Habar, Friston, Hoffner, & Frackowiak,
1991). Furthermore, studies that investigated the verbal short-term memory
suggest that the SMA is part of the articulatory loop which includes a
subvocal rehearsal system and a phonological store (Baddely, 1986; Paulesu,
Frith, & Frackowiak, 1993). Besides the left SMA, this functional system
supposedly encompasses the left BAs 44 (Brocas area) and 22 (auditory
association cortex) (Zatorre, Evans, Meyer, & Gjedde, 1992; Habib & Demonet, 1996). Awh and coworkers especially focused on the question whether
storage and rehearsal processes are implemented by different brain regions.
In verbal working memory, frontal regions seem to mediate rehearsal,
whereas posterior regions mediate storage (Awh, Smith, & Jonides, 1995).
Auditory word repetition which we have used as activation paradigm in this
study requires an intact system of short-term verbal memory, i.e., of shortterm storage and rehearsal. A significant activation of the left SMA in aphasic
stroke patients therefore potentially means that the brain tries to cope with
the aphasic impairment by activating those parts of the articulatory loop that
are still intact because of their location outside the territory of the left middle
cerebral artery. Moreover, if the supplementary motor cortex is part of the
internal rehearsal system and the superior temporal cortex of the short-term
storage, the activation of the SMA may indicate that aphasic patients change
their strategy of short-term memory by using their mental rehearsing system
instead of the impaired areas of short-term verbal storage in the left superior
temporal cortex. Based on this hypothesis, the additional activation of right
hemisphere regions such as the right inferior frontal cortex, which we found
in some severely impaired patients, may be interpreted as a further involvement of functionally connected and parallel processing areas which have
some speech competence but which are usually not needed for language
processing in the intact brain.
The additional activation of right hemispheric regions that we observed
in the subacute state of stroke, however, did not simply continue during the
period of long-term reorganization of the speech-relevant network, as one
might expect. A good long-term outcome apparently depended mainly on
the repair of left superior temporal cortex function, whereas the recruitment
of right hemisphere regions was significantly less effective. Previous PET
studies already demonstrated that recovery of the left superior temporal activation sometimes occurred even several months after stroke and that this
repair or reorganization of the primarily speech-relevant cortex areas was

BRAIN PLASTICITY IN APHASIA

227

more important for a good outcome of aphasia than largely activated regions in the right hemisphere (Heiss, Karbe, Weber-Luxenburger, Herholz,
Kessler, Pietrzyk, & Pawlik, 1997; Karbe, Herholz, Kessler, Wienhard,
Pietrzyk, & Heiss, 1997). Moreover, these results confirmed computer tomographic studies (Naeser, Gaddie, Palumbo, & Stiassny-Eder, 1990) which
demonstrated late recovery of auditory comprehension in global aphasia after
left subcortical temporal lesions, in contrast to lesions devastating Wernickes cortical area itself. That means that the subsistence of neurons at or
closely around a specific speech area allows ipsilateral structural compensation of lost function even months or years after the injury.
A completely new finding of this study is the significant inverse correlation
between the functional recovery of the left superior temporal cortex and the
task-induced right hemispheric activations in the long-term studies. The negative correlation means that the recuperation of speech competence within
the left superior temporal region reduces the permanent compensatory activity of right hemispheric regions or, argued from another point of view, that
the permanent loss of left superior temporal activity reinforces the compensatory activity of right hemispheric speech regions. Many researchers have
postulated that the speech-relevant network primarily is bilaterally organized
and parallel processing, although the left hemispheric speech regions usually
are much more effective in the adult brain (Benson & Zaidel, 1985). The
compensatory potential of right hemisphere regions after damage of their
left hemispheric counterparts, however, remained a matter of controversy
(Kertesz, 1988). Jackson, for example, considered the right hemisphere capable of automatic utterances and responsible for the residual speech output
in global aphasia ( Jackson, 1873), whereas Wernicke speculated that the
contralateral homologous cerebral cortex is responsible for the return of language function after left hemispheric injuries (Wernicke, 1886). A previous
PET study of six male patients who had completely recovered from Wernickes aphasia confirmed Wernickes assumption by finding significant increase of cerebral blood flow in right prefrontal and inferior premotor cortex
during speech processing (Weiller et al., 1995). In contrast to our study,
these patients, however, were selected retrospectively from a large data bank
of aphasic patients with respect to an excellent outcome. The observed type
of full speech recovery based on right hemisphere activation therefore presumably represents a type of compensation which is found only in some
specially organized brains.
In conclusion, our data indicate that the reorganization of the speech-relevant network is a process which is ongoing for at least several months or
even more than 1 year after the ischemic brain injury. The dynamics of recovery within the left hemispheric key regions of language processing determines the lasting coordination between speech-competent regions of both
cerebral hemispheres. The left hemispheric structural reorganization is significantly more effective than the right hemispheric compensation. The long

228

KARBE ET AL.

period of neural reorganization opens a wide time window for language

training.
REFERENCES
Awh, E., Smith, E. E., & Jonides, J. 1995. Human rehearsal processes and the frontal lobes:
PET evidence. Annals of the New York Academy of Sciences, 769, 97117.
Baddely, A. 1986. Working memory. Oxford: Clarendon Press.
Benson, D. F., & Zaidel, E. 1985. The dual brain. Hemispheric specialization in humans.
New York: Guilford Press.
Brodmann, K. 1909. Vergleichende Lokalisationslehre der Grohirnrinde in ihren Prinzipien
dargestellt auf Grund des Zellenbaues. Leipzig: Johann Ambrosius Barth.
Foundas, A. L., Leonard, C. M., & Heilmann, K. M. 1995. Morphologic cerebral asymmetries
and handedness. The pars triangularis and the planum temporale. Archives of Neurology,
52, 501508.
Gallaher, A. J. 1979. Temporal reliability of aphasic performance on the Token test. Brain
and Language, 7, 3441.
Habib, M., & Demonet, J. F. 1996. Cognitive neuroanatomy of language: the contribution of
functional neuroimaging. Aphasiology, 10, 217234.
Heiss, W. D., Pawlik, G., Herholz, K., Goldner, H., & Wienhard, K. 1984. Regional kinetic
constants and cerebral metabolic rate for glucose in normal human volunteers determined
by dynamic positron emission tomography of 18 F-2-fluoro-2-deoxy-D-glucose. Journal
of Cerebral Blood Flow and Metabolism, 4, 212223.
Heiss, W. D., Karbe, H., Weber-Luxenburger, G., Herholz, K., Kessler, J., Pietrzyk, U., &
Pawlik, G. 1997. Speech-induced cerebral metabolic activation reflects recovery from
aphasia. Journal of the Neurological Sciences, 145, 213217.
Herholz, K., Dickhoven, S., Karbe, H., Halber, M., Pietrzyk, U., & Heiss, W. D. 1996. Integrated quantitative analysis of functional and morphological 3D data by volume of interest. In T. Jones, V. Cunningham, R. Myers, & Bailey, D. (Eds) Quantification of brain
function using PET. San Diego, CA: Academic Press Pp. 175179.
Howard, D., Patterson, K., Wise, R., Brown, W. D., Friston, K., Weiller, C., & Frackowiak,
R. 1992. The cortical localization of the lexicons. Positron emission tomography evidence. Brain, 115, 17691782.
Jackson, J. H. 1873. On the anatomical and physiological localization of movements in the
brain. Lancet, 1, 232234.
Karbe, H., Herholz, K., Szelies, B., Pawlik, G., Wienhard, K., & Heiss, W. D. 1989. Regional
metabolic correlates of Token test results in cortical and subcortical left hemispheric
infarction. Neurology, 39, 10831088.
Karbe, H., Kessler, J., Herholz, K., Fink, G., & Heiss, W. D. 1995a. Long-term prognosis of
post-stroke aphasia studied with positron emission tomography. Archives of Neurology,
52, 186190.
Karbe, H., Wurker, M., Herholz, K., Ghaemi, M., Pietrzyk, U., Kessler, J., & Heiss, W. D.
1995b. Planum temporale and Brodmanns area 22. Magnetic resonance imaging and
high-resolution positron emission tomography demonstrate functional left-right asymmetry. Archives of Neurology, 52, 869874.
Karbe, H., Herholz, K., Kessler, J., Wienhard, K., Pietrzyk, U.K.E., & Heiss, W. D. 1997.
Recovery of language after brain damage. In H. J. Freund, B. A. Sabel, and O. W. Witte,
(Eds.). Brain plasticity. Advances in neurology. Philadelphia: Lippincrott-Raven. Vol.
73, Pp. 347358.

BRAIN PLASTICITY IN APHASIA

229

Karbe H., Herholz K., Weber-Luxenburger G., Ghaemi M., & Heiss W. D. (in press) Cerebral
networks and functional brain asymmetry. Evidence from regional metabolic changes
during word repetition. Brain and Language, 63, 108120.
Kertesz, A. 1988. Recovery of language disorders. Homologous contralateral or connected
ipsilateral compensation? In T. E. LeVere, R. Almli, & D. G. Stein, (Eds.), Brain injury
and recovery. Theoretical and controversial issues. New York: Plenum: Pp. 307321.
Metter, E. J., Kempler, D., Jackson, C., Hanson, W. R., Mazziotta, J. C., & Phelps, M. E. 1989.
Cerebral glucose metabolism in Wernickes, Brocas, and conduction aphasia. Archives of
Neurology, 46, 2734.
Naeser, M. A., Gaddie, A., Palumbo, C. L., & Stiassny-Eder, D. 1990. Late recovery of auditory comprehension in global aphasia. Improved recovery observed with subcortical temporal isthmus lesion vs. Wernickes cortical area zone. Archives of Neurology, 47, 425
432.
Oldfield, R. 1971. The assessment and analysis of handedness: The Edinburgh inventory.
Neuropsychologia, 9, 97113.
Orgass, B., & Poeck, K. 1966. Clinical validation of a new test for aphasia: An experimental
study of the Token test. Cortex, 2, 222243.
Paulesu, E., Frith, C. D., & Frackowiak, R. S. J. 1993. The neural correlates of the verbal
component of working memory. Nature, 362, 342345.
Penfield, W., & Roberts, L. 1959. Speech and brain mechanisms. Princeton, NJ: Princeton
Univ. Press.
Petersen, S. E., Fox, P. T., Posner, M. I., Mintun, M., & Raichle, M. E. 1988. Positron emission
tomographic studies of the cortical anatomy of single-word processing. Nature, 331, 585
589.
Pietrzyk, U., Herholz, K., Fink, G., Jacobs, A., Mielke, R., Slansky, I., Wurker, M., & Heiss,
W. D. 1994. An interactive technique for three-dimensional image registration: Validation
for PET, SPECT, MRI, and CT brain studies. Journal of Nuclear Medicine, 35, 2011
2018.
Rademacher, J., Galaburda, A. M., Kennedy, D. N., Filipek P. A., & Caviness, V. S. 1992.
Human cerebral cortex: Localization, parcellation, and morphometry with magnetic resonance imaging. Journal of Cognitive Neuroscience, 4, 352374.
De Renzi, E., & Vignolo, L. A. 1962. The Token test: A sensitive test to detect receptive
disturbances in aphasics. Brain, 85, 665678.
Steinmetz, H., Rademacher, J., Huang, Y., Hefter, H., Zilles, K., Thron, A., & Freund, H. J.
1989. Cerebral Asymmetry: MR planimetry of the human planum temporale. Journal of
Computer Assisted Tomography, 13, 9961005.
Swisher, L. P., & Sarno, M. T. 1969. Token test of three matched patient groups: Left braindamaged with aphasia; right brain-damaged with aphasia; non-brain-damaged. Cortex,
5, 264273.
Vignolo, L. A. 1979. Lesions underlying defective performances on the Token test: A CT
scan study. In F. Boller, & M. Dennis, (Eds.), Auditory comprehension. Clinical and
experimental studies with the Token test. New York: Academic Press. Pp. 135159.
Weiller, C., Isensee, C., Rijntjes, M., Huber, W., Muller, S., Bier, D., Dutschka, K., Woods,
R. P., Noth, J., & Diener, H. C. 1995. Recovery from Wernickes aphasia: A positron
emission tomography study. Annals of Neurology, 37, 723732.
Wernicke, C. 1886. Neuere Arbeiten uber Aphasie. Fortschritte der Medizin, 4, 371377.
Wienhard, K., Pawlik, G., Herholz, K., Wagner, R., & Heiss, W. D. 1985. Estimation of local
cerebral glucose utilization by positron emission tomography of 18 F-2-fluoro-2-deoxy-D-

230

KARBE ET AL.

glucose: A critical appraisal of optimization procedures. Journal of Cerebral Blood Flow

and Metabolism, 5, 115125.
Wienhard, K., Dahlbom, M., Eriksson, L., Michel, C., Pietrzyk, U., & Heiss, W. D. 1994.
The ECAT EXACT HR: Performance of a new high resolution PET scanner. Journal of
Computer Assisted Tomography, 18, 110118.
Wise, R., Chollet, F., Hadar, U., Friston, K., Hoffner, E., & Frackowiak, R. 1991. Distribution
of cortical neural networks involved in word comprehension and word retrieval. Brain,
114, 18031817.
Zaidel, E. 1979. Long-term stability of hemispheric scores on the Token test following brain
bisection and hemidecortation. In F. Boller & M. Dennis (Eds.), Auditory comprehension.
Clinical and experimental studies with the Token test. New York: Academic Press. Pp.
135159.
Zatorre, R. J., Evans, A. C., Meyer, E., & Gjedde, A. 1992. Lateralization of phonetic discrimination in speech processing. Science, 256, 846849.