DOI: 10.1093/brain/awh159
Brain (2004), 127, 1229±1236
Language reorganization in children with earlyonset lesions of the left hemisphere: an fMRI study
F. LieÂgeois,1 A. Connelly,2 J. Helen Cross,3 S. G. Boyd,4 D. G. Gadian,2 F. Vargha-Khadem1 and
T. Baldeweg1
1Developmental
Cognitive Neuroscience Unit and
and Physics Unit, Institute of Child Health,
University College London, 3Neurosciences Unit and
4Department of Clinical Neurophysiology,
Great Ormond Street Hospital for Children NHS Trust,
London WC1N 3JH, UK
2Radiology
Summary
It is widely assumed that following extensive damage to
the left hemisphere sustained in early childhood, language functions are likely to reorganize and develop in
the right hemisphere, especially if the lesion affects the
classical Broca's or Wernicke's language areas. In the
present study, functional MRI (fMRI) was used to
examine language lateralization in 10 children and
adolescents with intractable epilepsy who sustained an
early lesion in the left hemisphere. Lesions were adjacent to or within anterior language cortex in ®ve
patients, while they were remote from both Broca's and
Wernicke's areas in the remainder. A lateralization
index was calculated on the basis of the number of
voxels activated in the left and right inferior frontal
gyri when performing a covert verb generation task.
Only two patients were right-handed, suggesting a high
incidence of functional reorganization for motor control
Correspondence to: Dr F. LieÂgeois, Developmental
Cognitive Neuroscience Unit, Institute of Child Health,
UCL, 4th ¯oor, 30 Guilford Street, London WC1N 1EH,
UK
E-mail: [email protected]
in the remaining patients. Five out of 10 showed bilateral or right language lateralization, but lateralization could not be inferred from the proximity of
lesions to classical language areas on an individual
basis. Lesions in or near Broca's area were not associated with inter-hemispheric language reorganization in
four out of ®ve cases, but with perilesional activation
within the damaged left hemisphere. Paradoxically,
lesions remote from the classical language areas were
associated with non-left language lateralization in four
out of ®ve cases. Finally, handedness, age at onset of
chronic seizures, and site of EEG abnormality also
showed no obvious association with language lateralization. In conclusion, it is dif®cult to infer intra- versus
inter-hemispheric language reorganization on the basis
of clinical observations in the presence of early
pathology to the left hemisphere.
Keywords: cerebral dominance; language; early pathology
Abbreviations: fMRI = functional MRI; ICA = intracarotid amobarbital; LI = lateralization index
Received October 6, 2003. Revised January 20, 2004. Accepted February 9, 2004. Advance Access publication April 6, 2004
Introduction
It is now well established that extensive damage to the left
cerebral hemisphere in adults can result in profound and
chronic aphasia. In contrast, comparable lesions of the left
hemisphere suffered during childhood rarely result in
pronounced speech and language disorders (Woods and
Teuber, 1978; Woods and Carey, 1979; Vargha-Khadem
et al., 1985, 2000; Thal et al., 1991). The gross sparing of
speech and language functions after early lesions of the left
hemisphere has been attributed to the impressive plasticity
and reorganizational capacity of the immature brain, which
may enable speech and language functions to develop in the
right hemisphere.
In cases of early left hemisphere injury, it is widely
assumed (Rasmussen and Milner, 1977; Devinsky et al.,
1993; Isaacs et al., 1996; Lazar et al., 2000) that such interhemispheric language reorganization occurs when lesions
either encroach on language areas (i.e. Broca's and
Wernicke's) or are very extensive, while lesions remote
from language cortex do not result in inter-hemispheric
language reorganization (Devinsky et al., 1993; Isaacs et al.,
1996; Lazar et al., 2000). However, recent ®ndings have
called this assumption into question. Using cortical stimulation to map eloquent cortex, Duchowny et al. (1996)
have shown that developmental lesions accompanied by
Brain Vol. 127 No. 6 ã Guarantors of Brain 2004; all rights reserved
1230
F. LieÂgeois et al.
early-onset seizures originating in the left hemisphere can be
associated with language functions within that hemisphere.
Similarly, developmental tumours near language areas do not
necessarily lead to speech and language reorganization to the
right hemisphere, as assessed by the intracarotid amobarbital
(ICA) procedure (DeVos et al., 1995). The effect of lesion
site on speech and language lateralization following early
pathology of the left hemisphere therefore remains unclear.
Unfortunately, the ICA procedure does not address the issue
of intra-hemispheric reorganization, and electrocortical
stimulation does not provide information about the functions
subserved by the non-stimulated hemisphere. Functional MRI
(fMRI) is a useful non-invasive tool to address these issues.
Neuroimaging studies of both adults (Muller et al., 1999b)
and children (Duncan et al., 1997; Muller et al., 1998a, b,
1999a) have indicated that right hemisphere language
lateralization is more likely to be observed after early (i.e.
sustained before age 5 years) rather than late (i.e. after age 5
years) left hemisphere injury. This ®nding is consistent with
results of studies using ICA (Rasmussen and Milner, 1977;
Satz et al., 1988; Helmstaedter et al., 1997) and the dichotic
listening task (Isaacs et al., 1996) to assess language
lateralization. In the latter study, right hemisphere language
representation deduced from left-ear advantage was associated with pronounced motor de®cits in the right hand,
suggesting that right hemisphere language representation was
more likely to result after extensive lesions of the left
hemisphere. Unfortunately, in patients with early injury, the
effect of the proximity of the lesion to language areas on
reorganization of this function has never been examined
using fMRI. Clarifying this issue could improve outcome
prognosis in children who are candidates for the surgical
resection of the lesion.
In the present study, an fMRI task was used to determine
lateralization of activation within the anterior language area
(Broca's area) in 10 children and adolescents who suffered an
early lesion of the left hemisphere. In half the patients, lesions
were adjacent to or within anterior language cortex, while in
the other half lesions were remote from both Broca's and
Wernicke's areas. fMRI results obtained from the patient
group were compared with those obtained from a group of
neurologically intact children and adolescents. In addition to
lesion location, the relationship of handedness, seizure focus
and age at onset of epilepsy to language lateralization was
investigated in the patient group.
Material and methods
Control group
This group consisted of 10 participants, ®ve children aged 8±12:4
years (mean age 10:1 years), and ®ve adolescents aged between 14:4
and 16:10 years (mean age 15:8 years). Informed consent was
obtained from the parents of all participants as outlined by the Great
Ormond Street Hospital for Children/Institute of Child Health
Research Ethics Committee.
Patients
Ten patients (Table 1), aged 7±18 years, with early pathology of the
left hemisphere on MRI (Fig. 1) participated in this study. Their right
hemispheres showed no structural abnormality on MRI. Verbal IQs
ranged from 65 (exceptionally low range) to 126 (high range), but no
patient presented with clinical dysphasic symptoms at the time of
scanning. In ®ve patients, the lesions were remote from classical
language areas, located in the hippocampal, parahippocampal, or
temporal pole regions. In the remaining ®ve patients, the lesions
were adjacent to or within anterior language cortex, located within or
close to Broca's area in the frontal or centro-frontal regions in four
cases, and in the anterior superior temporal gyrus (Devinsky et al.,
1993) in one case. Two patients presented with a right-sided motor
weakness.
All patients suffered from drug-resistant seizures arising from the
left hemisphere and were investigated as potential candidates for
epilepsy surgery. During clinical investigation, the patients underwent standard video-EEG telemetry and structural MRI. An
extensive neuropsychological assessment was also performed to
assess language and memory functions, as well as handedness
(Crovitz and Zener, 1962). As part of the handedness inventory,
participants were asked which hand they favoured to perform 18
everyday activities, some of which used both hands (e.g. hammering
a nail, opening a tin). Rankings between 18 and 34 represent righthand preference, those between 35 and 54 indicate mixed
handedness, whilst those between 55 and 90 suggest left-hand
preference.
Evaluation of EEG abnormalities
Children underwent EEG video-telemetry as part of the pre-surgical
investigation, using standard techniques of electrode placement and
recording. Interictal and ictal EEG changes including abnormal
rhythmic activity, spikes and sharp waves were evaluated by a
paediatric clinical neurophysiologist, who was blind to the results of
the fMRI and MRI ®ndings. EEG abnormalities were categorized as
unilateral or bilateral, and, if the latter, any lateralized predominance
was noted. Ictal and interictal changes were assessed separately.
Localization was assessed using normal criteria for clinical
interpretation. Changes seen at F7 or F9 were interpreted as
temporal unless maximal involvement was seen at F3 or Fp1, when
they were interpreted as being of frontal origin.
fMRI data acquisition
MRI was performed on a 1.5 T Siemens Vision System (Erlangen,
Germany). Anatomical images were obtained from multi-slice T1weighted FLASH (fast low angle shot) images [TR (repetition
time) = 31 ms, TE (echo time) = 11 ms, ¯ip angle 40°, matrix size
256 3 256 3 64, voxel size 0.75 3 0.75 3 3 mm]. Functional MRI
data were acquired using a whole brain 3D EPI (echo-planar
imaging) sequence (TR = 87 ms, TE = 40 ms, ¯ip angle = 30°, matrix
64 3 64 3 64, 3 mm isotropic voxels). The total acquisition time for
each 3D data set was 5.6 s. For each participant, two consecutive
runs of 120 3D data sets were collected. Each run consisted of 10
task/rest cycles with six data sets in each state. The presentation of
the ®rst stimulus of each cycle was triggered automatically by the
®rst image in that cycle.
Since good correlations have been reported (Hertz-Pannier et al.,
1997; Benson et al., 1999; LeheÂricy et al., 2000) between
lateralization of fMRI activation during covert generation tasks
The images were analysed using the Statistical Parametric Mapping
software (SPM99, Wellcome Department of Imaging Neuroscience,
London, UK. http://www.®l.ion.ucl.ac.uk/spm). Images were
realigned, co-registered and then smoothed to three times the
original voxel size. The statistical analysis involved the comparison
of task versus rest states in a block design.
Left (±0.90)
Right (0.71)
Left (±0.79)
Right (0.84)
Right (0.97)
Left (±0.91)
Left (±0.25)
Bilateral (±0.06)
Left (±0.23)
Right (0.64)
Right (22)
Left (90)
Mixed² (36)
Mixed (41)
Left (76)
Mixed* (46)
Left (NA)
Right (22)
Mixed* (35)
Left³ (NA)
70
126
70
65
90
84
59
72
81
73
right
right
right
right
=
>
>
>
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
1
2
3
4
5
6
7
8
9
10
F
M
F
F
M
F
F
F
M
F
1:0
2:6
4:0
4:6
5:0
5:0
6:0 (®rst seizure at age 3 years)
8:0 (febrile seizure at 0:10)
14:0 (febrile seizure at 0:6)
11:0
Temporo-parietal
Anterior temporal/frontal
Frontal
Frontal
Fronto-temporal
Centro-temporal/frontal
Fronto-central/parietal
Anterior temporal/central
Fronto-temporal/frontal
Frontal
Side
In the control group, the fMRI pattern of activation (Table 2)
was very similar to that found in other studies using covert
generation paradigms (e.g. Holland et al., 2001) Activation
Patient Sex Age at onset of habitual seizures EEG abnormality
(years:months)
Location
Results
fMRI lateralization index
Table 1 Patient details
For each participant, a lateralization index (LI) was computed. First,
the highest activation voxel in the inferior frontal gyrus was
identi®ed [note: LI was calculated at P < 0.05 for patient 10 as no
voxel was signi®cantly activated at P < 0.01]. The number of
signi®cantly activated voxels within a 20 mm radius spherical
volume of interest centred on this voxel was then computed. A
homologous voxel relative to the midline was chosen in the
contralateral hemisphere, and an equivalent spherical volume
centred on this voxel was again used to calculate the number of
signi®cantly activated voxels. The LI was obtained by computing
(nR ± nL)/(nR + nL), where nL and nR are the number of activated
voxels in the left and in the right hemisphere, respectively.
Participants with a negative LI were considered left lateralized for
language, while those with a positive LI were considered right
lateralized. Participants whose LI ranged from ±0.1 to 0.1 were
considered to have bilateral language representation. This choice of
criteria for lateralization (which particularly in¯uences the classi®cation of patients 7, 8 and 9) was made on the basis of lateralization
results obtained from both invasive investigations (ICA and electrocortical stimulation) and a non-invasive method which involves the
direct statistical comparison of left and right hemisphere fMRI
activation (Statistical Lateralization Maps). Lateralization maps
could not be used in the present study because normalization of
MRIs to a template, which is required for analysis, would have been
considerably affected by the presence of large lesions and vascular
abnormalities in some patients described here (for details on the
method see LieÂgeois et al., 2002).
Aetiology
Lateralization index
M = male; F = female; VIQ = verbal IQ assessed using the UK versions of the WISC-III-R (Wechsler, 1992) and WAIS-R (Wechsler, 1998); DNET = dysembryoplastic
neuroepithelial tumour; CD = cortical dysplasia; AVM = arteriovenous malformation; *parental left-handedness; NA = not available; ³mild right hemiplegia; ²right-hand motor
weakness. Patients 4, 5, 7, 8 and 9 have been reported previously in a methodological article (LieÂgeois et al., 2002).
fMRI data analysis
17:5
7:10
12:4
14: 4
15:0
18:0
13: 8
16:6
16:4
15:0
and language lateralization determined with invasive techniques
such as ICA, a verb generation task was used here. During the task
period, participants were asked to generate covertly single verbs to
concrete nouns (1±3 syllables in length) presented via earphones
every 2.8 s (~12 nouns per task period). The word lists were
generated from a PC using in-house software. Before entering the
scanner, all participants were asked to generate verbs to nouns
overtly to ensure that they could perform the task successfully. The
inter-stimulus interval was increased to 4 s for those patients who did
not manage to produce a response within 3 s. During the rest period,
bursts of amplitude-modulated noise were presented at the same rate
as the nouns (every 2.8 or 4 s). Scanning time was 12 min per run,
which made the total experiment (anatomical and functional images)
~40 min long.
Hippocampal sclerosis
Anterior temporal DNET
Fronto-parietal AVM
Parahippocampal DNET
Anterior temporal DNET
> right Fronto-central DNET
> right Inferior frontal CD
Hippocampal sclerosis
Superior temporal CD or DNET
> Right Perisylvian polymicrogyria
Age at testing VIQ Hand preference fMRI language
(years:months)
(score)
lateralization (LI)
Language lateralization and early left lesions
1231
1232
F. LieÂgeois et al.
Fig. 1 Lesion location (circled) in each patient, as shown on T2- (patients 1, 2, 3, 5, 8 and 10) and T1-weighted (patients 4, 6, 7 and 9)
MRIs. The patient number is indicated in the left-hand corner in each case. (A) Lesions remote from classical language areas; (B) lesions
near or within anterior language cortex. Left hemisphere is on the left. Sagittal images show the left hemisphere.
involved the left inferior frontal gyrus (i.e. Broca's area), as
well as the posterior middle temporal region (Wernicke's
area). LIs were calculated on the basis of fMRI activation in
the inferior frontal gyrus, as this locus has been shown to be
most reliable for assessment of language lateralization
(LeheÂricy et al., 2000).
The LIs of all participants are presented in Fig. 2. All
control children and adolescents showed left hemisphere
lateralization. Five patients also showed left hemisphere
lateralization, while ®ve others showed non-left lateralization, either bilateral (one case) or right dominant (four
cases). Overall, the LI of patients (mean = 0.002, SD = 0.74)
and that of control participants (mean = ±0.71, SD = 0.23)
differed signi®cantly (Mann±Whitney test, P = 0.026),
control participants being more strongly left lateralized.
fMRI language lateralization and lesion
location
Out of the ®ve cases whose lesions were located near or in
Broca's area, four showed left hemisphere lateralization for
language (Fig. 3A). In these four cases, fMRI activation was
detected at the edge of the lesion, either more anteriorly
(patients 6 and 3) or more dorsally (patients 7 and 9).
Patient 7 is remarkable, as she is the only patient with a
congenital lesion restricted to Broca's area. fMRI activation
was detected above the lesion in the middle frontal gyrus.
This atypical language locus was con®rmed with electrocortical stimulation of this region, which induced speech
disturbance. Moreover, this patient did not present any speech
or language disorder following focal resection of the lesion.
Patient 10, who had polymicrogyria surrounding the left
sylvian ®ssure, showed activation not only in the right inferior
frontal gyrus, but also within the polymicrogyric cortex in the
left hemisphere (not shown), slightly more posteriorly to
Broca's area. The LI was calculated on the basis of activation
in the right inferior frontal region and, since no activation was
detected in the left hemisphere homotopic region, it was
concluded that there was right lateralization. It may be more
accurate to classify this patient as having bilateral language
representation, but this is impossible to ascertain as invasive
monitoring and surgery were not performed. What is clear in
this case, however, is that language activation was not left
lateralized. This patient also illustrates the limits of the LI
method when bilateral activation is not symmetrical
(LieÂgeois et al., 2002).
It may have been predicted that an early extensive lesion
affecting the perisylvian region would be associated with
right hemisphere language lateralization. The results from
patients 3 and 6 revealed that, in those two cases, frontal
activation was detected anterior to the lesion, with no
detectable activation in the homologous region of the right
hemisphere.
Out of the ®ve patients with lesions remote from the
classical language areas, four did not show left hemisphere
language lateralization (three right, one bilateral, see Fig. 3B).
In three patients, anterior fMRI activation was detected in the
inferior frontal gyrus, i.e. the Broca's area homologue in the
right hemisphere. The only exception was patient 8 who
showed bilateral activation in the frontal operculum.
fMRI language lateralization and clinical
observations (Table 1)
Only two patients were right-handed, while four were lefthanded and four showed mixed handedness. Non-right
handedness was equally often associated with right/bilateral
Language lateralization and early left lesions
1233
Table 2 Brain regions activated in the control group (activation signi®cant at P < 0.05,
corrected for multiple comparisons), obtained from averaged normalized data
Anatomical region
BA
Coordinates in mm
(x, y, z)
T value
Precentral gyrus
Anterior cingulate
Inferior frontal gyrus, pars triangularis
Inferior frontal gyrus, pars opercularis (dorsal)
Premotor cortex (ventral)
Posterior middle temporal gyrus
Precentral gyrus (ventral)
Inferior frontal gyrus, ventral
Supplementary motor area
L4
L24
L45
L44
L6
L21/22
L4
L45/47
L6
±48,
±3,
±54,
±57,
±45,
±57,
±48,
±48,
±3,
7.02
6.39
6.19
6.04
5.64
5.23
5.16
5.09
4.98
0, 48
24, 39
30, 12
18, 21
18, 21
±54, 12
9, 30
21, ±9
3, 54
BA = Brodmann's area; L = left hemisphere. Coordinates indicate local maxima in the x medial±lateral
axis (negative, left), y anterior±posterior axis (negative, posterior) and z dorsal±ventral axis (negative,
ventral). Origin: anterior commissure.
(four patients) as with left (four patients) fMRI language
lateralization (Fisher's exact test, P = 0.78).
EEG abnormality in the left frontal lobe was equally often
associated with left (four patients) as with non-left (four
patients) fMRI language lateralization (Fisher's exact test,
P = 0.10). Similarly, the presence of EEG abnormality in the
right hemisphere was associated with left hemisphere
language lateralization in four cases and with non-left
lateralization in three cases (Fisher's exact test, P = 0.50).
Finally, early onset of chronic epilepsy (i.e. before age 5
years) emanating from the left hemisphere was equally often
associated with left (three patients) as with non-left (three
patients) language lateralization (Fisher's exact test, P = 0.74).
The same ®nding applied if the ages at ®rst seizure were taken
into account.
Post-surgical outcome
Seven patients underwent surgery, including one who had
resective surgery following invasive subdural electrode
monitoring (patient 7) and one with functional stimulation
during awake craniotomy (patient 6). Patients 2, 5 and 8
underwent a temporal lobectomy. Lesionectomies were performed on patients 9, 4 and 7, and an anterior lesion resection
was performed on patient 6. None of these patients presented
with any reported postoperative speech or language disorders.
Formal postoperative neuropsychology assessment is
available in two of the seven patients. One year after surgery,
verbal intelligence had remained stable in patient 5 and had
increased by 16 points in patient 7. Formal postoperative
neuropsychological assessment is yet to be performed in the
®ve remaining patients, but no deterioration in language
abilities has been observed clinically.
Discussion
The present study provides further evidence that fMRI is
feasible in young epileptic patients, even in the presence of
Fig. 2 fMRI lateralization index (LI) in control participants and
patients. Bars indicate group means.
restricted intellectual abilities (Hertz-Pannier et al., 1997;
Booth et al., 1999). In this group of children and teenagers
with epilepsy associated with cortical lesions, while the
majority (eight out of 10) were of left or mixed handedness,
atypical language lateralization was found in only half.
Moreover, contrary to the ®ndings reported by Rasmussen
and Milner (1977), lesions encroaching on one or both
classical language areas rarely resulted in the reorganization
of speech and language functions to the right hemisphere. In
contrast, remote lesions were often associated with atypical
language lateralization. Most patients showed lateralization
patterns that could not have been inferred from clinical
observations.
Language lateralization in children with early
lesions in the left hemisphere
Whilst all control participants showed a left hemisphere
lateralization for language, half of the patient group showed
1234
F. LieÂgeois et al.
Fig. 3 Horizontal and sagittal views of fMRI maximal activation in the frontal cortex (crosshair) during covert verb generation, displayed
at a threshold of P < 0.01, uncorrected for multiple comparisons, except for patient 7 where P < 0.05. (A) Left lateralized language
representation; (B) right lateralized or bilateral language representation. Activation maps are superimposed on each patient's T1-weighted
image. The patient number is indicated in the left-hand corner in each case. L = left hemisphere; R = right hemisphere. Slice position is
indicated in the z dorsal±ventral axis (negative, ventral). Sagittal images show the left hemisphere in A and the right hemisphere in B.
right or bilateral language representation. Although our
sample is small, this high proportion of non-left lateralization
(50%) is consistent with ®ndings from previous studies using
ICA (55% in Rasmussen and Milner, 1977), dichotic listening
(60% in Isaacs et al., 1996) and fMRI (Springer et al., 1999)
to assess language lateralization. Furthermore, the cases of
right hemisphere representation showed fMRI activation in
the region homologous to Broca's area in the right hemisphere, consistent with other reports (Booth et al., 1999;
Staudt et al., 2002). Despite our small sample, a 50%
proportion of non-left language lateralization is particularly
high compared with the 4% found in right-handed adults, and
the 27% (10% in Pujol et al., 1999) found in the strongly lefthanded adult population (Knecht et al., 2000). These results
corroborate suggestions that the developing brain has far
greater reorganizational capacity than the adult brain with
respect to language (Woods and Teuber, 1978; Woods and
Carey, 1979; Vargha-Khadem et al., 1985, 2000; Thal et al.,
1991), since the majority of adults remain left lateralized for
language following left hemisphere injury (Benson et al.,
1999).
Language lateralization and lesion location
Based on earlier studies of large groups of patients
(Rasmussen and Milner, 1977; Isaacs et al., 1996), it was
hypothesized that early lesions near or within Broca's area
would be associated with language reorganization to the right
hemisphere. In our group, the data from only one patient (case
10) out of ®ve with lesions encroaching on, or very close to,
language cortex agreed with this prediction. In the remaining
four patients, fMRI activation was located very close to
abnormal cortex, and no signi®cant activation in the right
inferior frontal region was detected in patients 3, 6 and 7.
These results are consistent with ®ndings from invasive
techniques indicating that developmental lesions in language
areas do not always lead to inter-hemispheric language
reorganization (Duchowny et al., 1996), and that language
functions can develop within the left hemisphere around
childhood onset tumours (DeVos et al., 1995). Similarly, in
our sample, when lesions were located near or encroaching on
Broca's area, there seemed to be a strong predisposition for
intra-hemispheric language reorganization. Perilesional reorganization has also been observed in adults who suffered a
Language lateralization and early left lesions
left inferior frontal stroke and showed good language
outcome (Karbe et al., 1995; Rosen et al., 2000).
Another unexpected ®nding was that the majority of our
patients with lesions in the hippocampal, parahippocampal or
temporal pole regions showed right lateralized or bilateral
language representation. These lesions were much less
extensive than those previously described (Rasmussen and
Milner, 1977), and were remote from Wernicke's area. Right
hemisphere language lateralization has been observed previously in adults who suffered from early temporal epilepsy,
using PET (Muller et al., 1999b) and ICA (Rausch and
Walsh, 1984), and in a small proportion of children and adults
with temporal lobe epilepsy using fMRI (Springer et al.,
1999; Gaillard et al., 2002). Atypical speech representation
assessed by ICA has also been reported in 20 out of 64 adults
with medial temporal lobe epilepsy due to left hippocampal
sclerosis (Janszky et al., 2003). The reason why early
abnormality in temporal regions would cause language to
be predominantly subserved by the right hemisphere remains
unclear. It is possible that the spreading of epileptic
discharges to language regions located in the left temporal
cortex, such as the inferior temporal cortex (Luders et al.,
1991) and the anterior temporal pole (Devinsky et al., 1993),
may cause the frontal language network to represent this
function within the right hemisphere. The overall results from
the present study indicate that lateralization of language
functions cannot be inferred from the proximity of lesions
to classical language areas in children who suffer from
developmental or early brain pathology.
Language lateralization and clinical
observations
It was expected that patients with right-hand preference were
likely to show left hemisphere language representation,
whereas those with mixed or left-hand preference were likely
to show right or bilateral language representation. Results
from the literature (Rasmussen and Milner, 1977; Hinz et al.,
1994; DeVos et al., 1995; Janszky et al., 2003) and those
from the present study suggest that although right-handedness
is often associated with left hemisphere lateralization, non
right-handedness is not a good predictor of inter-hemispheric
reorganization of language functions. Indeed, in our group,
only two patients were right-handed, indicating a high
proportion of functional reorganization for hand motor
control. One example of the lack of a systematic relationship
between handedness and language lateralization is patient 3,
who suffered from right congenital motor weakness and
whose lesion involved the sensorimotor cortex, but showed
fMRI activation in the left inferior frontal gyrus. This ®nding
is in disagreement with the claim of a causal link between
pathological left-handedness and right lateralization for
language (Devinsky et al., 1993). It similarly has been
shown that a small proportion of patients who suffer from
early-onset left-sided temporal epilepsy can develop right
1235
language lateralization and remain right-handed (Rausch and
Walsh, 1984). It is therefore important to distinguish between
results obtained at the group level (i.e. Woods et al., 1988;
Springer et al., 1999) and their predictive value at an
individual level. Because of evidence of a certain degree of
dissociation between hand motor control and language
lateralization, it is not possible to predict language reorganization based on a shift in handedness, especially in patients
with either neurodevelopmental or early pathology of the left
hemisphere.
Our results also suggest that the age at onset of chronic
epilepsy does not predict language lateralization, consistent
with conclusions drawn from a recent study on patients with
focal medial temporal lobe epilepsy of early onset (Janszky
et al., 2003). Additionally, although it was hypothesized that
EEG abnormality in the left frontal lobe or in the right
hemisphere would prevent language organization in those
regions, no such evidence was observed. This may not be
surprising if one considers that one of our patients, case 10,
even showed fMRI activation within polymicrogyric cortex,
consistent with the idea that abnormal cortex can be
functionally active (Marusic et al., 2002). It seems possible,
therefore, for cognitive functions to develop in structurally
abnormal regions, particularly in the presence of developmental pathology.
In conclusion, results from the present study con®rmed that
non-left language lateralization is more common in patients
who suffer from early or developmental left hemisphere
lesions than in control children. Nevertheless, in these same
patients, the factors that have been thought to impact on
language reorganization, such as handedness and lesion
proximity to language regions, do not seem to have a
predictive value at an individual level.
Acknowledgements
We wish to thank Dr W. K. Chong for reviewing the MRI
scans. F.L. was supported by fellowships from INSERM
(France) and the Fyssen Foundation. We also wish to thank
the Wellcome Trust for support. Research at the Institute of
Child Health and Great Ormond Street Hospital for Children
NHS Trust bene®ts from Research and Development Funding
from the NHS Executive.
References
Benson RR, FitzGerald DB, LeSueur LL, Kennedy DN, Kwong KK,
Buchbinder BR, et al. Language dominance determined by whole brain
functional MRI in patients with brain lesions. Neurology 1999; 52:
798±809.
Booth JR, MacWhinney B, Thulborn KR, Sacco K, Voyvodic J, Feldman
HM. Functional organization of activation patterns in children: whole
brain fMRI imaging during three different cognitive tasks. Prog
Neuropsychopharmacol Biol Psychiatry 1999; 23: 669±82.
Crovitz HF, Zener K. A group-test for assessing hand- and eye-dominance.
Am J Psychol 1962; 75: 271±6.
Devinsky O, Perrine K, Llinas R, Luciano DJ, Dogali M. Anterior temporal
1236
F. LieÂgeois et al.
language areas in patients with early onset of temporal lobe epilepsy. Ann
Neurol 1993; 34: 727±32.
DeVos KJ, Wyllie E, Geckler C, Kotagal P, Comair Y. Language dominance
in patients with early childhood tumors near left hemisphere language
areas. Neurology 1995; 45: 349±56.
Duchowny M, Jayakar P, Harvey AS, Resnick T, Alvarez L, Dean P, et al.
Language cortex representation: effects of developmental versus acquired
pathology. Ann Neurol 1996; 40: 31±8.
Duncan JD, Moss SD, Bandy DJ, Manwaring K, Kaplan AM, Reinian EM,
et al. Use of positron emission tomography for presurgical localization of
eloquent brain areas in children with seizures. Pediatr Neurosurg 1997;
26: 144±56.
Gaillard WD, Balsamo L, Xu B, Grandin CB, Braniecki SH, Papero PH, et al.
Language dominance in partial epilepsy patients identi®ed with an fMRI
reading task. Neurology 2002; 59: 256±65.
Helmstaedter C, Kurthen M, Linke DB, Elger CE. Patterns of language
dominance in focal left and right hemisphere epilepsies: relation to MRI
®ndings, EEG, sex, and age at onset of epilepsy. Brain Cogn 1997; 33:
135±50.
Hertz-Pannier L, Gaillard WD, Mott SH, Cuenod CA, Bookheimer SY,
Weinstein S, et al. Noninvasive assessment of language dominance in
children and adolescents with functional MRI: a preliminary study.
Neurology 1997; 48: 1003±12.
Hinz AC, Berger MS, Ojemann GA, Dodrill C. The utility of the intracarotid
amytal procedure in determining hemispheric speech lateralization in
pediatric epilepsy patients undergoing surgery. Childs Nerv Syst 1994; 10:
239±43.
Holland SK, Plante E, Weber BA, Strawsburg RH, Schmithorst VJ, Ball WS
Jr. Normal fMRI brain activation patterns in children performing a verb
generation task. Neuroimage 2001; 14: 837±43.
Isaacs EB, Christie D, Vargha-Khadem F, Mishkin M. Effects of
hemispheric side of injury, age at injury, and presence of seizure
disorder on functional ear and hand asymmetries in hemiplegic children.
Neuropsychologia 1996; 34: 127±37.
Janszky J, Jokeit H, Heinemann D, Schulz R, Woermann FG, Ebner A.
Epileptic activity in¯uences the speech organization in medial temporal
lobe epilepsy. Brain 2003; 126: 2043±51.
Karbe H, Kessler J, Herholz K, Fink GR, Heiss W-D. Long-term prognosis
of poststroke aphasia studied with positron emission tomography. Arch
Neurol 1995; 52: 186±90.
Knecht S, Drager B, Deppe M, Bobe L, Lohmann H, Floel A, et al.
Handedness and hemispheric language dominance in healthy humans.
Brain 2000; 123: 2512±8.
Lazar RM, Marshall RS, Pile-Spellman J, Duong HC, Mohr JP, Young WL,
et al. Interhemispheric transfer of language in patients with left frontal
cerebral arteriovenous malformation. Neuropsychologia 2000; 38:
1325±32.
LeheÂricy S, Cohen L, Bazin B, Samson S, Giacomini E, Rougetet R, et al.
Functional MR evaluation of temporal and frontal language dominance
compared with the Wada test. Neurology 2000; 54: 1625±33.
LieÂgeois F, Connelly A, Salmond CH, Gadian DG, Vargha-Khadem F,
Baldeweg T. A direct test for lateralization of language activation using
fMRI: comparison with invasive assessments in children with epilepsy.
Neuroimage 2002; 17: 1861±7.
Luders H, Lesser RP, Hahn J, Dinner DS, Morris HH, Wyllie E, et al. Basal
temporal language area. Brain 1991; 114: 743±54.
Marusic P, Najm IM, Ying Z, Prayson R, Rona S, Nair D, et al. Focal
cortical dysplasias in eloquent cortex: functional characteristics and
correlation with MRI and histopathologic changes. Epilepsia 2002; 43:
27±32.
Muller RA, Rothermel RD, Behen ME, Muzik O, Mangner TJ, Chakraborty
PK, et al. Brain organization of language after early unilateral lesion: a
PET study. Brain Lang 1998a; 62: 422±51.
Muller R-A, Rothermel RD, Muzik O, Becker C, Fuerst DR, Behen ME,
et al. Determination of language dominance by [15O]-water PET in
children and adolescents: a comparison with the Wada Test. J Epilepsy
1998b; 11: 152±61.
Muller R-A, Behen ME, Rothermel RD, Muzik O, Chakraborty PK, Chugani
HT. Brain organization for language in children, adolescents, and adults
with left hemisphere lesion: a PET study. Prog Neuropsychopharmacol
Biol Psychiatry 1999a; 23: 657±68.
Muller R-A, Rothermel RD, Behen ME, Muzik O, Chakraborty PK, Chugani
HT. Language organization in patients with early and late left-hemisphere
lesion: a PET study. Neuropsychologia 1999b; 37: 545±57.
Pujol J, Deus J, Losilla JM, Capdevila A. Cerebral lateralization of language
in normal left-handed people studied by functional MRI. Neurology 1999;
52: 1038±43.
Rasmussen T, Milner B. The role of early left-brain injury in determining
lateralization of cerebral speech functions. Ann NY Acad Sci 1977; 299:
355±69.
Rausch R, Walsh GO. Right-hemisphere language dominance in righthanded epileptic patients. Arch Neurol 1984; 41: 1077±80.
Rosen HJ, Petersen SE, Linenweber MR, Snyder AZ, White DA, Chapman
L, et al. Neural correlates of recovery from aphasia after damage to left
inferior frontal cortex. Neurology 2000; 55: 1883±94.
Satz P, Strauss E, Wada J, Orsini DL. Some correlates of intra- and
interhemispheric speech organization after left focal brain injury.
Neuropsychologia 1988; 26: 345±50.
Springer JA, Binder JR, Hammeke TA, Swanson SJ, Frost JA, Bellgowan
PS, et al. Language dominance in neurologically normal and epilepsy
subjects: a functional MRI study. Brain 1999; 122: 2033±46.
Staudt M, Lidzba K, Grodd W, Wildgruber D, Erb M, Krageloh-Mann I.
Right-hemispheric organization of language following early left-sided
brain lesions: functional MRI topography. Neuroimage 2002; 16:
954±967.
Thal D, Marchman V, Stiles J, Aram D, Trauner D, Nass R, et al. Early
lexical development in children with focal brain injury. Brain Lang 1991;
40: 491±527.
Vargha-Khadem F, O'Gorman AM, Watters GV. Aphasia and handedness in
relation to hemispheric side, age at injury and severity of cerebral lesion
during childhood. Brain 1985; 108: 677±96.
Vargha-Khadem F, Isaacs E, Watkins KE, Mishkin M. Ontogenetic
specialization of hemispheric function. In: Oxbury JM, Polkey CE,
Duchowny M, editors. Intractable focal epilepsy. London: W.B. Saunders;
2000. p. 405±18.
Wechsler D. Wechsler Intelligence Scale for Children. WISC III. 3rd edn.
Sidcup (UK): Psychological Corporation; 1992.
Wechsler D. Wechsler Adult Intelligence Scale. WAIS III. 3rd edn. Sidcup
(UK): Psychological Corporation; 1998.
Woods BT, Carey S. Language de®cits after apparent clinical recovery from
childhood aphasia. Ann Neurol 1979; 6: 405±9.
Woods BT, Teuber HL. Changing patterns of childhood aphasia. Ann Neurol
1978; 3: 273±80.
Woods RP, Dodrill CB, Ojemann GA. Brain injury, handedness, and speech
lateralization in a series of amobarbital studies. Ann Neurol 1988; 23:
510±8.