DOI: 10.1093/brain/awh204
Brain (2004), 127, 1796–1810
Intra-operative mapping of cortical areas involved
in reading in mono- and bilingual patients
Franck-Emmanuel Roux,1,2 Vincent Lubrano,1,2 Valérie Lauwers-Cances,3 Michel Trémoulet,2
Christopher R. Mascott2 and Jean-François Démonet1,4
1
Institut National de la Santé et de la Recherche Médicale,
Unité 455, 2Federation of Neurosurgery, 3Service
d’Epidémiologie and 4Federation of Neurology, University
Hospitals, 31059 Toulouse, France
Correspondence to: Franck-Emmanuel Roux, Service de
Neurochirurgie et INSERM 455, Hôpital Purpan, F-31059
Toulouse, France
E-mail: [email protected]
Summary
In order to identify the cortical areas involved in the reading process and to spare them during surgery, we systematically studied cortical areas by direct cortical stimulation
in patients operated on for brain tumours. Seventy-six
cortical stimulation mapping studies for language were
performed in 35 monolingual and 19 bi- or multilingual
patients over a 5-year period. We systematically searched
for reading interference areas in addition to standard naming areas using an ‘awake surgery’ technique for brain
mapping. A ‘reading aloud’ task (translated into different
languages in multilingual patients) was used. Brain mapping was performed in left (44 patients) and right (10
patients) hemispheres. Cortical areas involved in reading
were identified according to the type of interference, location and distinctness from naming areas. Stimulation of
several major hemispheric regions resulted in significant
interference with reading aloud: (i) the lower part of the
pre- and postcentral gyri (P < 0.00001); (ii) the dominant
supramarginal, angular and the posterior part of the superior temporal gyri (P < 0.00001); (iii) in the dominant inferior and middle frontal gyri (P < 0.001); and (iv) in the
posterior part of the dominant middle temporal gyrus
(P < 0.05). Interferences in reading were generally found in
small cortical areas, with intervening areas evoking no
reading interferences. Only partial overlap between reading and naming sites was found. Reading-specific sites were
preferentially found when stimulating dominant inferior
parietal or posterior temporal areas. Different types of
reading interferences were noted. While ‘articulatory’
interferences were found in pre- and postcentral gyri bilaterally, and ocular-induced movements in bilateral middle
frontal gyri, paraphasias were found mainly in the dominant supramarginal and posterior superior temporal gyri.
Reading arrest sites were found in many regions. Reading
interference sites were also occasionally found in the nondominant hemisphere. In bilingual patients, if common
cortical areas could be found, language- and reading-specific areas were sometimes detected, lending support to the
concept that bilinguals can have relatively distinct cortical
representation of their language skills. Finally, in this series,
the location of reading interference sites and their relative
specialization showed considerable individual variability.
Keywords: bilingualism; brain tumour; cortical mapping; language; reading
Abbreviations: fMRI = functional MRI; F2 = middle frontal gyrus; F3 = inferior frontal gyrus; T1 = superior temporal
gyrus; T2 = middle temporal gyrus
Received August 7, 2003. Revised February 1, 2004. Second revision March 23, 2004. Accepted March 29, 2004.
Advanced Access publication July 7, 2004
Introduction
In the history of mankind, the emergence of writing abilities
(and of its corollary, reading) constituted a major step in civilization by allowing the preservation and transmission of language and the facilitation of exchange of knowledge beyond
the scope of purely spoken languages. The ability to read and
write is an acquired process, sometimes affected by developmental (Taylor Sarno, 1998; Paulesu et al., 2001) or acquired
Brain Vol. 127 No. 8
#
(Déjerine, 1891, 1892; Hinshelwood, 1902; Beauvois and
Derouesne, 1979; Damasio and Damasio, 1983; Kawamura
et al., 1987; Coslett and Saffran, 1989; Anderson et al.,
1990; Taylor Sarno, 1998) deficits. The earliest stages of the
reading processes stem from visual input (or alternative inputs
in blind Braille readers) and classically involve the dominant
fusiform, lingual and angular gyri (Damasio and Damasio,
Guarantors of Brain 2004; all rights reserved
Cortical sites of reading
1797
1983; Cohen et al., 2000, 2002). The organization of the whole
reading process is, however, far more complex. Several recent
reports based on clinical or brain activation studies have shown
that multiple areas could be involved in the reading process
throughout the left and right hemispheres (Petersen et al., 1990;
Price et al., 1994; Schäffler et al., 1996; Ojemann, 1998; Cohen
et al., 2000; Simos et al., 2000). Moreover, the cortical organization of reading function may depend on the type of writing/
reading paradigms used, the manner of reading acquisition and
training, and on the number of languages mastered by the
subjects (‘monolingual’ versus ‘bilingual’ people). These factors have raised many questions about brain organization for
reading (Hinshelwood, 1902; Dedic, 1926; Lyman et al., 1938;
Albert and Obler, 1978; Price et al., 1994; Sakurai et al., 1994;
Paulesu et al., 2000; Nakada et al., 2001; Roux and Trémoulet,
2002; Tan et al., 2003).
In order to identify the cortical areas involved in the reading
process and to spare them during surgery, we studied monoand multilingual patients operated on for brain tumours. In
these patients, operated on using an ‘awake surgery’ technique
(Ojemann and Whitaker, 1978; Ojemann et al., 1989, 2002),
the different language areas involved in reading were localized
by direct cortical stimulation mapping. The cortical organization of the reading process was analysed based on these
surgical data.
Hand dominance, language and reading testing
Material and methods
Patients
(iv) early acquisition (before the age of 7 years), low proficiency in
All patients in this series had language examinations pre- and postoperatively in order to rule out language-specific deficits. This testing
(of all languages studied) included the evaluation of written and oral
understanding, naming, language fluency, reading, computation,
dictation, repetition, written transcription and object handling,
using appropriate tests in languages spoken by the patients. Dysphasic
patients (with >10% of errors in naming tests) or patients with
pre-operative reading deficits were excluded from this study. The
degree of handedness of the subjects was assessed with the Edinburgh
Handedness Inventory test (Oldfield, 1971). In our study, the handedness indices ranged between 100 and 90. Fifty-one patients in this
series were right-handed (42 with a lesion in the left and nine in the
right hemisphere) while three were left-handed. All these data are
detailed in Table 1.
All subjects were French natives, but in bi- or multilingual patients
(n = 19), all languages in which the subject was fluent were tested.
Among the languages tested in bilingual patients, English was tested
in eight cases, Spanish in three, Occitan in five, German in three,
Russian in one, Chinese Mandarin in one and Arabic in one. For
this study of a small group of patients, languages and patients were
classified as follows (see Table 2):
(i) early acquisition (before the age of 7 years, considered as native
language), high proficiency in language: L. I;
(ii) late acquisition (after the age of 7 years), high proficiency in
language: L. II;
(iii) late acquisition (after the age of 7 years), low proficiency in
language: L. III; or
Over a 5-year period (from December 1997 to December 2002), 76
languagemappingswereperformedinourdepartmentusingatleasttwo
standard tasks (naming and reading) in 54 patients (aged 13–76 years
old, average 54 years;23 male, 31 female patients)operated on forbrain
tumour or occasionally for other lesions (such as cortical dysplasia).
Naming and reading sites were identified in order to spare them during
surgery.Datafromthesesuccessivebrainmappingswereprospectively
collected by the same team using the same protocol over the years.
Lesions were in the left hemisphere in 44 cases, while 10 patients had
brain tumours located in the right hemisphere (18%). As suggested by
other authors (Taylor and Bernstein, 1999), it has been the policy of our
department to propose, when feasible, the ‘awake surgery’ technique
with direct brain mapping, in patients with brain tumours where functional areas were potentially ‘at risk’ during surgery. Accordingly,
patientswithbraintumourslocatedintherighthemispherewereoffered
‘awake surgery’ if mapping of the Rolandic area was required.
language: L. IV (i.e. with a serious loss of a previously acquired
native language resulting from the acquisition of another
language).
Occitan is a regional French dialect, spoken in the south of France,
that contains many words with the same lexical roots as modern
French (French and Occitan belong to the same ‘romance languages’
family), but also with many distinct words. Both languages share a
number of cognate words and orthographic features (similar to the
relationship between French and Spanish). Occitan is often spoken,
but more rarely read. In our five bilingual Occitan-French patients, two
of them were used to reading in Occitan. The other three patients,
although able to read and translate Occitan reading materials into
French, did not usually read in Occitan. Our institution serves an
area with a population of 3.5 million. The bilingualism in this area
is composed of Occitan-speaking people >60 years (Occitan
often being their first language learned at home with their parents),
Table 1 Patients’ mono- or bilingual status and number of language studies performed
Patients (n = 54)
Tumour side
Monolingual patients (n = 35),
multilingual patients (n = 19)
Language studies,
total (n = 76)
Right-handed patients (51)
Left hemispheric (42)
Monolinguals (28)
Bilinguals (12)
Patient speaking three languages (1)
Patient speaking four languages (1)
Monolinguals (5)
Bilingual patients (4)
Monolinguals (2)
Bilingual patient (1)
28
24
3
4
5
8
2
2
Right hemispheric (9)
Left-handed patients (3)
Left hemispheric (2)
Right hemispheric (1)
1798
F.-E. Roux et al.
Table 2 Language classification used in this study
for bilingual patients
L. I
L. II
L. III
L. IV
Early-acquired language (before age 7 years), spoken
daily or frequently. The patient was able to name
various objects in both languages and could understand
all the sentences he/she read without any problem.
He/she could also have a conversation on concrete
or abstract topics.
Late-acquired language (after age 7 years), spoken daily
or frequently. The patient was able to name various
objects in this language and could understand all the
basic sentences he/she read without any problem. He/she
could also have a conversation on concrete topics.
Late-acquired language (after age 7 years), not spoken
every day. The patient was able to name various basic
objects in this language and could understand all the
sentences he/she read without any problem. He/she
could also have a conversation on concrete or
abstract topics.
Early-acquired language (before age 7 years), not spoken
every day (serious loss of a previously acquired
native language resulting from the acquisition of another
language). The patient was able to name various basic
objects in the language without any problem. He/she
could understand all basic sentences he/she read
without any problem. He/she could also have a
conversation on concrete topics.
composed of Spanish- or Arabic-speaking persons (because of the
vicinity of Spain or immigration from Spain or North Africa), people
employed by the aeronautic and high-technology industries, and
university students, usually speaking English.
Fig. 1 Examples of basic sentences from our set of 30 different
sentences used for brain mapping in our patients.
Cortical mapping procedures
All patients were operated on using the ‘awake surgery’ technique
(Penfield and Roberts, 1959; Ojemann et al., 1989). We used a
neuro-navigational system in 46 patients (85%). Intra-operative
cortical stimulation was used to localize areas of functional cortex
after determination of the after-discharge threshold by electrocorticography. The cortex was directly stimulated using the bipolar
electrode of the Ojemann cortical stimulator (1 mm electrodes
separated by 5 mm; Radionics1, Burlington, MA, USA). Our strategy was to spare the language areas identified by stimulation during
tumour removal by resecting tumour to within no more than 1 cm
from the eloquent cortex (distance of the resection margin from the
nearest functional site). Direct brain mapping was usually
performed in <20 min in monolingual patients. In each patient,
each site studied was tested for both reading and naming. Each
site was studied twice in monolinguals: language A (Lang. A)
naming and reading. In bilinguals, each site was studied four
times (Lang. A naming; Lang. B naming; Lang. A reading; Lang.
B reading), six and eight times, respectively, in our two patients
speaking three or four languages. All the patients and their families
gave their informed consent to study language areas (naming and
reading) by direct brain mapping.
For brain mapping, patients were asked to perform two different
tasks in the appropriate languages: a naming task to detect standard
anomia (i.e. ‘This is a . . . ’), and a reading task involving various basic
unrelated sentences (and not previously rehearsed by the patients).
During naming procedures, the patients were shown a set of 30 pictures
of various objects (a car, a wheelbarrow, a plane, a chair, a bird, a ball, a
lion, a tomato, grapes, a bomb, a bottle, a boat, a key, a strawberry, a
screwdriver, a rose, a saw, a chicken, a computer, slippers, bread,
hands, a book, a rubber, a clock, a fridge, a pencil, a flute, a camera, a
door). The patients were trained to name each object as it appeared.
These pictures were shown randomly and the stimulation was applied
as soon as they appeared. For the reading tasks, we used a set of 30
different sentences (see examples in Fig. 1). The patients were asked to
read rather slowly and stimulation was applied randomly on the cortex
while they were reading. Naming and reading tasks are elaborate tasks
known to activate similar or different areas in the brain (Ojemann
et al., 1989; Ojemann, 1991). One of the differences between the
reading and picture naming paradigms is that reading involves syntactic processing and multiple word types (verbs, grammatical words)
in addition to concrete nouns. To test reading in bilingual persons, we
have a data set of sentences translated from French into the major
world languages or specific languages spoken in our region.
Conditions of validation
To be accepted as language location, the language sites found were
tested meticulously at least three separate times. During direct brain
mapping, single anomia sites can occasionally be found
(Ojemann et al., 1989) and the interpretation of this phenomenon
remains complex. Although this is a matter of debate, sites showing
no reproducible language interference were not included in this
Cortical sites of reading
study. In multilingual patients, languages were tested sequentially
(one language mapping session after another). Nevertheless, at the
end of language mapping, the language-specific sites found were
confirmed by alternate mapping in both languages in order to carefully validate such sites. A picture of the brain was systematically
taken after each brain mapping along with a written record of the
findings during stimulation. Intra-operative photographs of the brain
indicated the sites of positive or negative response to cortical
stimulation. In the last 2 years of this study, cortical mapping
procedures were all video recorded, the patients’ answers being
recorded with a microphone placed near their mouths (as evidence
of the patient’s language cortical organization and to be further
analysed in team meetings).
Postoperative tests
Postoperatively, patients were asked to perform the same tests as preoperatively in order to detect any language difficulties. The results
were compared with the intra-operative findings. These postoperative
tests are important in determining recovery or surgery-related deficits.
In practice, they can be very difficult to analyse because of several
factors: first of all, as noted by the earliest authors (Penfield and
Roberts, 1959), postoperative deficits evolve with time, and may
sometimes not be present immediately after surgery but only a few
days later. This is usually not a reflection of direct surgical trauma, but
of subsequent oedema. The meaning of such deficits with regard
to localization is therefore questionable. Several factors can also
interfere with post surgical language evaluations, such as patient
post-surgical stress or fatigue (especially if radiotherapy is initiated).
Because partial or total of language recovery can occur between the
date of operation and the date of postoperative language evaluation,
specific language deficits can be underestimated or missed.
Furthermore, this study included some patients with high-grade
brain tumours or metastases, limiting long-term follow-up. Finally,
the patterns of language performance can be related not only to surgery, but also to environmental (speech therapy) or other individual
factors (Minkowky, 1965; Paradis, 1995; Fabro, 1999).
Analysis of language sites
All authors presenting direct brain mapping data face the challenge of
the spatial and anatomical localization of the functional responses. To
address these issues, brain regions can be defined according to different approaches: the brain surface can be divided into small squares
drawn schematically in relation to surface landmarks (Ojemann et al.,
1989) or by using a grid on a template (Schäffler et al., 1996). We think
that the terms of ‘Broca’s’ or ‘Wernicke’s’ regions remain imprecise
and not very informative. For present our data, we chose to define
regions using the gyral/sulcal anatomy. For instance, the supramarginal gyrus was considered as a region, as was the angular gyrus. Large
gyri, such as the temporal gyri, for example, were arbitrarily divided
into three segments by drawing an imaginary line extending the preand postcentral sulci.
Several issues have to be discussed in the analysis of data found
during direct brain mappings. Direct brain mapping in monolinguals
can yield a large amount of data that includes: (i) the location of the
language interference sites; (ii) the variation of the response when two
different tasks are used (Ojemann et al., 2002); and (iii) the type of
response obtained by direct cortical stimulation; the term ‘reading
interference’ can include typical reading arrest, hesitations in reading
or semantic errors, as shown in Table 3. Bilinguals add further complexity to the analysis and presentation of the results. It must be also
1799
Table 3 Classification used in this study of the different
types of language interferences found during the ‘naming’
and ‘reading’ tasks
Different types of language
interference found
during the ‘naming’ task
Different types of language
interference found
during the ‘reading’ task
Articulatory/interference
sites:
Stimulation evokes a visible
face or tongue contraction.
Almost always found in the
pre-and/or postcentral gyri.
Articulatory/interference
sites:
Stimulation evokes a
visible face or tongue
contraction. Almost
always found in the
pre- and/or postcentral gyri.
Typical anomia sites:
The patient is unable to find
the word appropriate to the
image shown. Example: ‘This
is a . . . I know what it is but
I can’t find the word . . . ’. Once
stimulation is no longer applied,
the patient is successful in
finding the word corresponding
to the image.
Pure reading arrest sites:
The patient stops reading.
Example: ‘The car is blue
and . . . ’. Once stimulation
is no longer applied, the
patient is again successful
in reading the sentence.
No face or tongue
contraction is visible. No
underlying mechanism
explaining reading arrest
is found.
Speech arrest sites:
The patient is unable to say
anything after the image
has been shown. No
visible contraction/face
and mouth.
Paraphasia sites:
The speech is fluent but
the text read by the
patient is incomprehensible
or the words are
inappropriate. Example: ‘
The tar is cue and straw . . . ’.
Miscellaneous findings:
hesitations, repeated words,
use of uncommon synonyms.
Example: ‘This is a . . .
yes . . . a . . . this a car . . .
I am right? This a car?’, or
‘This is a Chrysler’.
Miscellaneous findings:
repeated words,
hesitations during reading.
Other findings: single error
In a given site, interference for
the task used is found
only once.
Ocular movements:
The patient stops reading
because of ocular
movements; he/she
sometimes acknowledges
that reading is difficult.
Other findings: single error
In a given site, interference
for the task used is found
only once.
understood that the absolute number of reading sites that we found in a
region does not necessarily mean that this region is more involved in
the reading process than another. Indeed the left hemispheric regions
known as language regions were preferentially tested; conversely,
areas not classically implicated in language processes such as the
upper parietal or anterior frontal areas in the left hemisphere or the
entire right hemisphere were less frequently tested by direct cortical
mapping. Thus, the relative number of reading sites found when
related to the total number of stimulations performed in a given region
is more informative. The absolute number of sites found needed to
1800
F.-E. Roux et al.
be normalized to the number of the stimulations performed in a
given region.
We analysed brain mapping data by separating reading sites from
naming sites (thus defining specific reading sites within both hemispheres). Secondly, we analysed the inter-dependence of reading sites
to naming sites as a function of the location and the type of response
obtained. To avoid confusion in the presentation of results, we have
presented all the localizations of reading sites on a figure using a
standard brain according to operative anatomical data, whereas the
type of response and the variations of the response obtained are presented elsewhere in the text. Bilinguals were included in the global
results (two brain mappings for each subject).
Finally, and most importantly, it must be underscored that when we
qualified a site as ‘reading-specific’, no naming deficit was found in
that site. We cannot exclude, however, that our sites defined as
‘reading-specific’ may lead to stimulation interference of other
functions not tested in this study.
Statistical analysis
We analysed the percentage of responses for each zone stimulated
according to the number of stimulations performed. The analysis was
carried out with unbalanced repeated-measures analysis of variance
(ANOVA) after transformation of percentage values to obtain a
normal distribution (Zar, 1999).
The respective P-values were corrected for lack of sphericity using
Box’s conservative epsilon. Differences were estimated significant at
P < 0.05. The analysis was performed using Stata 7.0 statistical software (Stata Corporation, College Station, TX, USA).
Results
Overall results
A total of 76 language studies (naming and reading) using
naming and reading tasks were performed in 54 patients (35
monolingual, 17 bilingual and two speaking three and four
languages, respectively) in frontal, temporal and parietal dominant and non-dominant regions (152 different brain mappings
in all). In this study, no stimulation was performed in occipital,
inter-hemispheric and basal temporal regions. We studied
1577 different cortical sites, testing each of them with a naming
and reading task in one language (for 35 monolingual patients)
or in two, three and four languages for the remaining multilingual patients. At least one reading site and one language site
(anomia or speech arrest) was found in all patients but one. In
this patient, stimulation over the non-dominant (right hemisphere) supramarginal, superior temporal gyrus (T1) and angular gyri evoked no response. Including all types of naming and
reading interferences, 344 naming interferences and 406 reading interferences (205 in our 35 monolingual patients, 201 in
our 19 bilingual patients) were found. These 406 reading interferences were found in 309 different cortical sites (some sites
were involved in two or more reading interferences in bilingual
patients). Naming interferences were found in 251 sites (some
sites were involved in two or more naming interferences in
bilingual patients). Table 4 presents the overall naming and
reading results in all patients; Table 5 presents the results of
reading stimulations in bilinguals.
On average, we found more reading interference than
naming interference; mean for 44 patients that had their left,
dominant hemisphere studied: 8.13 [95% confidence interval
(CI) 5.99–9.95] reading interferences per mapping compared
with 6.88 (95% CI 5.54–8.57) interferences found for naming
per brain mapping (P < 0.01). Hence, reading interference sites
extended over a larger zone than those found during naming.
Although naming sites and reading sites were usually located
within the same brain region, there was only partial overlap
between sites producing naming interferences when stimulated
and those in which stimulation produced reading interferences.
As for naming sites, reading interference sites were small areas
with intervening areas where no reading interferences were
evoked. The margins of these sites were usually distinct, the
displacement of the electrode into an adjacent cortical area
located in the same gyrus producing no interference. The greatest current that did not evoke after-discharges varied from one
patient to another, ranging from 3 to 9 mA.
Specific results in all patients
Location of reading interference sites
Reading interference sites were found in several cortical
regions of the left hemisphere. Figures 2 and 3 summarize
the reading sites found in relation to intra-operative anatomy,
showing the number of mappings performed in each region
(circles) and the percentage of stimulations that caused reading
interference over the total of stimulations performed in each
region (squares).
Although reading interference sites were found in several
regions of the left temporo-parietal and frontal lobes, several
major left hemisphere regions were significantly associated
with ‘reading aloud’ interference: (i) the lower part of the
Rolandic (pre- and postcentral gyri) region (P < 0.00001);
(ii) the angular and supramarginal gyri and the posterior
part of the T1 gyrus (P < 0.00001); (iii) the posterior part of
the inferior frontal gyrus (F3) and middle frontal gyrus (F2)
(P < 0.001); and (iv) the posterior part of the middle temporal
gyrus (T2) gyrus (P < 0.05) (see Fig. 4).
Reading aloud interferences were found sometimes far
beyond the traditional Broca’s and Wernicke’s areas. Reading
interferences sites were sometimes found in anterior temporal
areas or in middle frontal gyri. Although rare, reading interferences were occasionally found in dominant upper parietal and
superior frontal gyri, or in the right hemisphere of right-handed
patients. The main area producing reading interferences in the
non-dominant hemisphere was the lower part of the precentral
gyrus (‘articulatory’ process interferences). Nevertheless,
some reading interference sites were also occasionally
found in non-dominant supramarginal or frontal gyri.
Types of reading interference
Different mechanisms producing reading interferences could
be identified. Using the classification of the types of reading
interferences (presented in Table 4), articulatory interference
Cortical sites of reading
1801
Table 4 Number of stimulations performed in each region and number of naming and reading interferences found
(number of mappings = 76)
Region studied
Left hemisphere
F1
Upper precentral gyrus
Upper postcentral gyrus
Upper parietal
F2
Middle precentral gyrus
Middle postcentral gyrus
Supramarginal gyrus
Angular gyrus
F3
Lower precentral gyrus
Lower postcentral gyrus
Anterior T1
Middle T1
Posterior T1
Anterior T2
Middle T2
Posterior T2
Anterior T3
Middle T3
Posterior T3
Right hemisphere
F1
Upper parietal
Angular gyrus
Supramarginal gyrus
Middle precentral gyrus
F2
Lower postcentral gyrus
Lower precentral gyrus
F3
Posterior T1
Middle T1
Number of sites
stimulated in the
region studied (n = 1577)
Number of naming
interferences (n = 344)
Number of reading
interferences (n = 406)
32
24
15
13
131
40
25
160
25
189
97
59
58
92
138
32
50
138
5
28
13
3
0
0
0
21
5
2
49
2
47
74
16
3
9
44
1
3
23
0
0
1
3
1
0
1
24
6
3
68
8
50
76
17
3
11
54
1
4
26
0
2
1
28
7
12
25
11
38
16
36
23
9
8
1
0
0
1
1
3
3
29
1
1
1
1
0
1
2
2
4
3
31
1
1
1
F1 = superior frontal gyrus.
sites were found 71 times, pure reading arrest 176 times,
paraphasias sites 39 times, ocular movement sites 22 times
and hesitations during reading 98 times. These different
types of reading interferences were analysed in relation with
their location:
(i) Articulatory interferences with visible contraction of the
face or tongue were mainly found in the lower part
of the left and right precentral and postcentral gyri
(P < 0.00001). In these cases, reading was stopped or
slowed (sometimes with slurred speech) when stimulation
was applied. These symptoms probably reflect the involvement of the articulatory process in the ‘reading aloud’
activity.
(ii) ‘Pure’ reading arrest sites are sites in which no visible
contraction of the face or tongue and no visible ocular movements are seen. No underlying mechanism leading to
reading arrest was found. They were frequently found either
in frontal or in temporo-parietal areas. Reading arrest was
often abrupt once stimulation was applied.
(iii) Of the different types of reading interference found, paraphasias were probably the most striking; the language
output remained fluent (though sometimes slowed) but
consisted of unintelligible pseudo-words or inappropriate
words. The patients continued reading as phonological
integration was impaired. Significant sites producing
paraphasia during reading were found in the supramarginal gyri, the posterior T1 and T2 gyri and sometimes in
the angular gyrus (P < 0.01).
(iv) Ocular movements impairing reading were found mainly
in frontal areas, especially in dominant and non-dominant
middle frontal gyri (P < 0.00001) (see Fig. 5). It must be
noted that reading interference sites without visible ocular
movements were also found in F2. Rarely (two patients),
ocular movements inducing reading interference were
also found in the inferior parietal lobe.
(v) Finally, other patterns of response were also found, including hesitation or perseveration. No specific location was
found for these miscellaneous effects. These sites were
usually located close to reading arrest or paraphasia sites.
1802
F.-E. Roux et al.
Table 5 Bilingual patients (n = 19); number of stimulations performed in each region and number of common (Lang. A, B,
C and D) and specific (Lang. A or B) reading interferences found (number of mappings = 41)
Region studied
Left hemisphere
F1
Upper precentral gyrus
Upper postcentral gyrus
Upper parietal
F2
Middle precentral gyrus
Middle postcentral gyrus
Supramarginal gyrus
Angular gyrus
F3
Lower precentral gyrus
Lower postcentral gyrus
Anterior T1
Middle T1
Posterior T1
Anterior T2
Middle T2
Posterior T2
Anterior T3
Middle T3
Posterior T3
Right hemisphere
F1
Upper parietal
Angular gyrus
Supramarginal gyrus
Middle precentral gyrus
F2
Lower postcentral gyrus
Lower precentral gyrus
F3
Posterior T1
Middle T1
Number of
areas
stimulated
in the region
studied
(n = 796)
Number of
reading
(including
Lang. A, B,
C and D)
interferences
found (n = 201)
Number of
cortical sites
producing
common (Lang.
A, B, C or D)
reading
interferences (n = 91)
Number of
specific
reading
interferences
in Lang. A
(French)
(n = 6)
Number of
specific
reading
interferences
in Lang. B*
(n = 7)
16
10
8
0
72
16
12
90
8
94
44
48
32
48
88
20
26
92
0
4
4
2
0
0
0
12
2
0
30
2
21
33
12
0
11
38
0
4
20
0
2
0
1
0
0
0
6
1
0
13
1
9
15
6
0
4
18
0
1
10
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
1
0
2
0
0
1
0
0
0
0
0
0
0
0
1
0
2
0
0
0
3
0
0
0
0
0
1
0
16
0
0
0
4
14
6
12
12
0
0
0
0
0
0
2
4
2
4
0
0
0
0
0
0
0
1
2
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Patients 7 and 8 had no specific reading areas for their Lang. C and D. Russian (patient 7) and Chinese Mandarin and German reading areas
(patient 8) were common to French reading sites. T3 = inferior temporal gyrus; F1 = superior frontal gyrus; Lang. = language.
Topographical specificity of reading sites:
‘reading-specific’ sites
During brain mappings, naming and reading interference sites
could be located within the same cortical area. Nevertheless,
reading interference sites could also be specific. In these cases,
no naming interference was detected in a cortical site evoking a
reading interference. We called these sites ‘reading-specific’
sites. Overall, 58 reading-specific sites were found, 32 in
the monolingual subgroup (n = 35) and 26 in the bilingual
subgroup (n = 19).
Studying these reading-specific sites, three facts were noted:
First, reading-specific sites were rarely found isolated but
rather adjacent to (distance of 1 cm or less) naming sites.
By ‘isolated’ reading site, we mean a reading site that is not
contiguous to a naming site (see Fig. 4 for illustration and
stimulations performed in the angular gyrus showing
two isolated reading sites). Over these 58 reading-specific
sites, an isolated reading site was found only 17 times (29%).
Secondly, reading-specificity depended on the nature of
language interference. Sites of speech arrest during naming
were almost always sites of sharp reading arrest. Overall,
speech arrest during naming was found 79 times. Seventyfive of these sites (95%) were also associated with reading
arrest; in other words, a speech arrest during naming was
almost always associated with reading arrest. But this was
different when we compared the sites producing typical
anomia during naming tasks with reading sites. Overall,
anomia sites were found 54 times. At these same sites, reading
interferences were observed only 35 times (65%).
Cortical sites of reading
1803
Fig. 2 Reading location in the left hemisphere in 44 mono- or multilingual patients (61 different brain mappings). Cortex has schematically
been divided into several regions limited by dotted lines. Circles = number of times this cortical region has been studied (over these
61 brain mappings). Squares = percentage of reading interference found over of the total number of stimulations performed in this
region (whatever the type of reading interference considered).
Fig. 3 Reading location in the right hemisphere in 10 mono- or multilingual patients (15 different brain mappings). Cortex has
schematically been divided into several regions limited by dotted lines. Circles = number of times this cortical region has been studied (over
these 15 brain mappings). Squares = percentage of reading interference found over of the total number of stimulations performed in
this region (whatever the type of reading interference considered).
1804
F.-E. Roux et al.
Fig. 4 Angular, T1 and supramarginal gyrus reading sites. A 44-year-old monolingual (French) woman with a left parietal high-grade
tumour (WHO grade III). (A) Axial and coronal MRI slices showing the lesion. (B) Pre-mapping cortex. (C) Reading sites. (D) Naming
sites. Black line = posterior part of the sylvian fissure; R = reading sites; L = naming sites; H = naming hesitation sites; *R = reading
site producing paraphasias when stimulated. To improve the understanding of the intra-operative pictures, cortical sites producing no
language impairment were not labelled by a sterile ticket. Reading interference sites were found within the angular gyrus and in the
posterior part of the sylvian fissure (T1 and supramarginal gyrus). In this case, reading sites were sometimes specific (i.e. different from
naming sites), especially in the angular region, where no naming sites were found.
Finally, with regard to reading/naming independence, we
assessed whether reading-specific sites were preferentially
found in a given region. Reading-specific interferences were
found more numerous then naming-specific interferences in
three zones, dominant posterior T1, dominant supramarginal
gyrus and dominant angular gyrus, although this did not reach
statistical significance.
The left-handed patients’ cases
Three left-handed patients were studied. Two of them had their
left hemisphere studied. Reading interferences were found in
these two left hemispheres (both in precentral and inferior frontal
gyri). Reading interference sites were also found in right frontal
gyri in one left-handed bilingual (French/English) patient.
Results of reading mapping in bilingual patients
Overall, 41 different language studies were performed in our 19
bi- or multilingual patients. Two of them (patients 3 and 11) had
their right hemisphere studied. All but one (patient 3) were
right-handed. Regarding naming interferences, a total of 81
common naming interference sites (motor mechanism, typical
anomia, speech arrest or hesitation) were found. We found 91
sites of common (Lang. A, B, C or D) reading interferences
(including all types of reading interferences). Including our
two patients (patients 7 and 8) in whom three and four languages were tested, a total of 201 reading interferences were
detected in these sites. Some sites (n = 91) were involved in
two, three or four languages. Others were language-specific
(n = 13). A subgroup of sites were not only language- but also
reading-specific (n = 9). No difference in terms of reading or
naming site location was found between monolingual and
bilingual patients (P > 0.05); in other words, bilingual patients
did not have a different distribution of naming and reading sites
compared with monolingual patients. Although languagespecific cortical sites were sometimes found in some patients,
no statistical difference between first and second language
Cortical sites of reading
1805
Fig. 5 ‘Ocular’ mechanisms impairing reading in right and left frontal region. (A) Right-handed patient with a low-grade astrocytoma in the
right superior frontal gyrus. Three sites, (denoted ‘O’, ocular site) were found impairing reading while stimulated, in the middle frontal (two
sites) and superior frontal (one site) gyri. Contralateral (left) eye deviation was noted. Another site, located in the superior frontal gyrus
produced repeated reading hesitation (H). (B) Right-handed patient with a high-grade astrocytoma (WHO grade III) in the left superior
frontal gyrus. Two sites impairing reading with ocular deviation (denoted ‘O’) to the right were found in the middle frontal gyrus. Note
the Rolandic area (H = hand area; T = thumb contraction) posterior to the lesion.
(French versus other languages) organization (for both naming
and reading tests) was found (P > 0.05).
‘Language-specific’ sites
In our group of bilingual patients we found 13 language-specific
reading sites (i.e. producing only reading interferences in one
language and not in any other, as illustrated in Fig. 6 and summarized in Table 6). Six of these specific sites were found in
French and seven in second languages. Language-specific sites
were found in dominant frontal, parietal or temporal regions.
In these 19 bilingual patients, eight of them had at least one
language-specific reading site (one had three language-specific
reading sites, and one had four language-specific reading sites).
All types of reading interference were found in second language mappings. No switch between languages during reading
was observed. During stimulation of one site, one patient said
that she could not find any words in her first language but she
was able to read in her second language. This phenomenon was
observed only once.
Language- and reading-specific sites
In our group of bilingual patients, naming and reading were
both tested. Whether we found language-specific sites in these
patients, we also found a sub-group of cortical sites which were
language- and reading-specific (i.e. producing only reading
interference in one language). Nine of the 13 language-specific
1806
F.-E. Roux et al.
Fig. 6 Bilingual cortical reading site specificity in left hemisphere. (A) A 37-year-old bilingual (French/Spanish) right-handed man. Spanish
was his mother tongue that he always spoke with his parents. He started learning French at age 3 years at kindergarten, continuing to the end
of primary school. He regularly speaks Spanish with his family in Spain. He had surgery for a left sylvian infiltrative lesion diagnosed to be
an atypical meningioma. He had no language deficit. R = reading sites; M = motor, articulatory sites (Rolandic region); H = Spanish reading
hesitation site; N = negative sites; R* = French reading site producing paraphasias when stimulated. T1 = superior temporal gyrus. Dotted
line = sylvian fissure. Up to the sylvian fissure, on the Rolandic region, motor sites were common to French and Spanish. Reading sites were
found at different locations for French and Spanish. (B) Naming sites (L) found in French and Spanish. (C) Summary of the reading and
naming mapping in French and Spanish on a 3D reconstruction of the patient’s brain. Reading specific sites were found in T1 and
supramarginal gyri (Spanish and French). In the Rolandic area, sites of interferences were common to Spanish and French languages
(articulatory mechanism).
reading sites found were not only language-specific but also
reading-specific. These language- and reading-specific sites
were found in patients 2, 15 and 16 (three patients with all
different levels of proficiency of language) in both T1 and
supramarginal gyri.
Finally, two cases of brain mapping were performed for
reading non-Latin alphabets, the first one being a Cyrillic alphabet (Russian), and the second involving logograms (Mandarin
Chinese). In these two patients, French reading interference
sites were common to Mandarin Chinese (in T2) and Russian
(in posterior F2, F3 and precentral areas) reading arrest.
Postoperative reading assessment
It was our policy in our patient population to spare the cortical
sites found during naming and reading tasks (by remaining at
least 1 cm from the site found). Among the 50 patients in whom
the reading sites were effectively spared, difficulties in reading
were found postoperatively in 13 patients (26%), probably due
to brain retraction or postoperative oedema. The difficulties
consisted of slow reading, reading hesitation or phonemic
errors. Of these 13 patients, seven were tested again a few
weeks after surgery and none of them had significant persistent
reading difficulties. For various reasons, the remaining six
patients were lost to follow-up and could not be tested again.
For oncological reasons, the reading sites found were not
preserved during surgery in four of our patients. None of these
sites were sites of typical reading arrest, but were sites of
reading paraphasias or hesitation (in the dominant and
non-dominant supramarginal gyri in two cases) and in the
dominant postcentral gyrus in the remaining two patients.
The latter patients had postoperative expressive readingaloud difficulties and slight facial palsy. Significant reading
Cortical sites of reading
Table 6 Summary of the language-specific reading sites
found according to language classification used in this
study for bilingual patients
Patient
Lang. A
Lang. B, Lang. C, Lang. D
1
2
3
4
5
6
7
French = L. I
French = L. I*
French = L. I
French = L. I
French = L. I
French = L. I
French = L. I
8
French = L. I
9
10
11
12
13
14
15
16
17
18
19
French = L. I
French = L. I
French = L. I
French = L. I
French = L. I
French = L. I
French = L. I*
French = L. I
French = L. I
French = L. I
French = L. I
English = L. II
English = L. II
English = L. II
English = L. III*
English = L. III
English = L. III
English = L. II,
Russian = L. III
English = L. II*,
German = L. III,
Mandarin Chinese = L. III
Occitan = L. I*
Occitan = L. I*
Occitan = L. I
Occitan = L. III
Occitan = L. IV
Spanish = L. I*
Spanish = L. I*
Spanish = L. III*
German = L. II
German = L. III
Arabic = L. I
*At least one language-specific reading site found. Lang. =
language.
difficulties were gradually resolved within 3 months, although
reading aloud remained slow. A bilingual patient (French/English) in whom the cortex of the dominant supramarginal gyrus
(containing only a French reading paraphasias site) was opened
to reach the tumour showed several phonemic substitutions
while reading the day after his operation (almost exclusively
in French). Interestingly, reading improved dramatically
within 24 h and no phonemic substitutions were noted 48 h
after the operation (pseudo-word reading was also strictly
normal). The last patient in whom tumour removal was
performed in the non-dominant right supramarginal gyrus
(where two reading sites were found) showed no postoperative
reading disturbance.
Discussion
Reading as a specific brain function
The study of the neural substrate of reading has been predicated
on three strategies: clinical studies based mainly on acquired
language deficits in selected patients, research studies on
volunteers using brain mapping tools or direct stimulation
studies on neurosurgical patients. Studies in patients with
neurological disorders allowed the first insights into the functional anatomy of reading and the specificity of this cerebral
function. Although impairments and subsequent recovery of
reading function have been known for centuries (the first
description dates from Pliny the Elder, AD 23–79, and
Valerius Maximus, AD c. 31), reports on this topic became
1807
more numerous at the end of the 19th century and throughout
the 20th century (Déjerine, 1891, 1892; Hinshelwood, 1902;
Dedic, 1926; Lyman et al., 1938; Stevens, 1957; Luria, 1960;
Yamadori, 1975; Albert and Obler, 1978; Wechsler, 1977;
Damasio and Damasio, 1983; Kawamura et al., 1987).
Déjerine, who published two different cases of alexia in
1891 and 1892, hypothesized that the visual information,
first treated by the right and left occipital lobes, reaches the
angular gyrus through white matter structures and the splenium
for visual information of the right hemisphere (Déjerine, 1891,
1892). In daily clinical practice, alexia or reading difficulties in
individuals more often consist of reading and writing disturbances in the context of predominantly aphasic symptomatology (Taylor Sarno, 1998). Among reports of acquired alexia,
the cases of Japanese aphasic patients are particularly interesting, as this language has two modalities of writing (Kanji,
derived from Chinese characters, and Kana, a phonological
syllabic system). Dissociated cases of selective or preferential
reading impairment in one or other modality have been
described (Yamadori, 1975; Kawamura et al., 1987; Sakurai
et al., 1994). These two systems of the Japanese writing could
involve different anatomical and/or functional modalities
(Kawamura et al., 1987). The involvement of distinct brain
circuits in different reading systems has also been evoked in
patients unable to read letters, but still able to read numbers
(Déjerine, 1891) or musical notes (Assal and Buttet 1983).
These observations led researchers to postulate that the cerebral organization of reading processes may be individualized
and may also depend on the code used for reading and writing.
Using various functional imaging techniques, modern
approaches to elucidate the neural basis of reading have differentiated two pathways for reading (Price et al., 1994): a lexical
mechanism, converting visual inputs to a whole word phonological representation, and a sublexical route used to convert
orthographic segments into the corresponding phonological
elements. Implication of the angular, supramarginal and superior temporal gyri, as well as basal temporal and Broca regions
in both presumed reading pathways, has been demonstrated
(Petersen et al., 1990; Simos et al., 2000). Some brain regions
could be preferentially dedicated to one mechanism. For
instance, the superior temporal and the supramarginal gyri
may be implicated in the analysis of phonological operations,
whereas lexical analysis may be more likely in the angular
gyrus (Rumsey et al., 1999). From a neurosurgical point of
view, only a few direct cortical stimulation studies have
explored the cortical areas (Schäffler et al., 1996; Ojemann,
1998) involved in reading. In a 1998 monograph, Ojemann
identified different modules involved in the reading process
in sites different or common to naming sites (Ojemann, 1998).
The sites producing reading interference also showed individual variability, as did naming sites, being localized in
temporo-parietal but also in frontal areas. In 45 patients
operated for epilepsy, Schäffler et al. (1996) identified cortical
sites in Broca’s and Wernicke’s areas, but also in basal temporal regions that could produce total reading arrest or significant slurring.
1808
F.-E. Roux et al.
Our results are comparable to those of previous direct stimulation studies which showed that the organization of language
was divided into multiple functional systems, called ‘essential
areas’ for language (Ojemann, 1991). According to its accuracy, direct electrical stimulation can produce impairment of
small cortical areas involved in specific tasks. This organization is variable according to the language skill studied,
although the specialization of the areas involved in different
language skills (such as comprehension, articulation or grammar) is relative (Mesulam, 1990). Indeed, in a recent direct
cortical stimulation study, it has been shown that verb generation tasks or naming tasks can involve shared but also different
cortical areas (Ojemann et al., 2002). This is also true when
reading is tested: we found shared areas for naming/reading,
but also reading-specific sites. In this study, specific cortical
areas involved in reading were most frequently found in
dominant supramarginal, posterior T1 and angular gyri.
Furthermore, reading sites were often found in the vicinity
of naming sites, thus creating ‘cortical patches’ of language
areas. The frequent impairment of phonological processing
(fluent speech with several paraphasias) during stimulations
of the supramarginal and posterior temporal gyri could be
related to activation studies that showed involvement of this
region in phonological tasks (Demonet et al., 1992, 1996) and
in the sublexical route for reading.
With regard to reading interferences related to ocular movements induced by stimulation, we found these in dominant and
non-dominant frontal areas (mainly over F2). This is not surprising, since reading requires coordinated ocular movements.
Moreover, it has long been known that ocular movements can
be induced by stimulation over frontal areas (Penfield and
Roberts, 1959). Finally, one-fifth of the brain mappings performed in this study were performed in the right hemisphere.
Most reading sites in the non-dominant hemisphere were found
in the pre- and postcentral gyri, but sometimes were found
elsewhere (in frontal or supramarginal gyri). The possible
involvement of the non-dominant hemisphere in reading had
already been hypothesized based on the findings of some residual reading in patients with large left hemispheric lesions
(Coslett and Monsul, 1994) or from data of brain activation
studies (Price et al., 1994; Fiez and Pertersen, 1998).
The reading process in bilinguals
While the technique of direct brain mapping is used frequently
in neurosurgery, many factors have not facilitated the study
of cortical organization in bilinguals (Roux and Trémoulet,
2002). Limitations have included the absence of a strict definition of a ‘bilingual’ person, and in particular, what level of
proficiency in languages a person must have to be considered as
bilingual. Bilingual individuals do not develop the same level
of competence or proficiency for their respective languages
when they use them on different occasions. Some estimate that
over 50% of the world population is bilingual (Cristal, 1997;
Fabro, 1999). This is not always acknowledged, however, limiting the number of bilinguals to a restricted number of persons
that speak ‘major’ languages and excluding languages considered dialects for social, practical or other reasons (Itziar
and Korostola, 2001).
One-third of the patients of this series were bi- or multilingual. One of the most striking features seen in bilinguals
with aphasia is the possibility of selective aphasia for one
language or differential impairment (Pitres, 1895). At the end
of the 19th century, after the first described cases of selective
aphasia in bilinguals, some physicians postulated that each
language could have a separate representation in the brain
(Scoresby-Jackson, 1867; Adler, 1983; Hinshelwood, 1902).
However, several authors (Pitres, 1895; Penfield and
Roberts, 1959; Minkowsky, 1965; Lhermitte et al., 1966;
Critchley, 1974) have disagreed with or even denied the
neuroanatomical basis of phenomena of selective aphasias
or recoveries.
The literature on bilingualism now reports numerous cases
of dissociation of languages during the acute phase of aphasia
or in its recovery (Pitres, 1895; Albert and Obler, 1978; Paradis,
1995). More selective forms of dissociations have been published, such as selective impairment of reading and writing in
only one language (Hinshelwood, 1902; Dedic, 1926;
Yamadori, 1975; Welscher, 1977), adding another level of
complexity to the understanding of the cortical organization
of bilinguals. Hinshelwood described a striking case of a
34-year-old polyglot patient with reading difficulties after
brain damage who first recovered his capacity to read a
‘dead’ language (ancient Greek), followed by Latin and subsequently, after several months, French and English (Hinshelwood, 1902). Another interesting case was described by Gleb
regarding a professor of ancient languages who sustained an
injury of the left frontal lobe, becoming aphasic in his mother
tongue, but maintaining ability to read silently and understand
books in Latin and ancient Greek (Gleb, 1983). Stevens
described a striking case of a bilingual English/Hebrew patient
who had epilepsy preferentially triggered by reading Hebrew
(Stevens, 1957) and documented EEG epileptiform changes
upon reading Hebrew. Other similar cases of selective impairment of reading in one language in bilinguals have been
reported (Lyman et al., 1938), sometimes without concomitant
deficit of writing (Luria, 1960; Welscher, 1977).
In a few recent studies, functional MRI (fMRI) has been used
to explore reading processes in bilinguals (Chee et al., 1999;
Nakada et al., 2001; Tan et al., 2003). Nakada et al. (2001)
studied 10 bilingual American/Japanese volunteers and found
that reading in the second language involved the same cortical
structures as those recruited for the first language. Using oral
word generation cued by written stimuli, Chee et al. (1999)
studied 24 Mandarin/English volunteers with fMRI and
showed that there were no significant differences between
the cortical representations of these languages at the single
word level. It has also been demonstrated by using fMRI
in bilinguals speaking very different languages (such as
Mandarin Chinese and English) that neural structures involved
in second language reading could be shaped by native language
reading (Tan et al., 2003).
Cortical sites of reading
In this series, we have found clear evidence that bilingual
patients use common but also different cortical areas for different languages. The possibility of finding different sites of
language in bilingual persons has also been noted by all other
cortical stimulation studies performed in bilinguals (Ojemann
and Whitaker, 1978; Ojemann et al., 2002; Walker et al., 2004)
and by some authors using fMRI experiments (Kim et al.,
1997). Why do some fMRI experiments on reading show overlap between two different languages in bilinguals, while electrical stimulation shows language-specific sites? The answer is
not straightforward, but the use of group analyses in most fMRI
studies could not show subtle anatomic differences in language
cortical organization in bilinguals. The possibility of finding
differential impairment in bilinguals is likely to depend
on multiple factors (neuroanatomical, linguistical, environmental, affective or age of acquisition). In this study we
found in bilinguals some language-specific areas for reading
in temporo-parietal but also in frontal regions. As has been
proposed for other language skills (Ojemann and Whitaker,
1978; Kim et al., 1997; Simos et al., 2001), and in view of our
results, we think that reported cases of selective or preferential
reading impairments in bilinguals could be explained, at least
in part, by different anatomical locations of some readingspecific cortical areas.
Limits of the study
Intra-operative mapping of reading has obvious limitations.
First, reading ability reflects complex processing evolving
with the reader’s performance and only imperfectly studied
by direct brain stimulation or by modern brain mapping tools.
Reading is not limited to ‘reading aloud’ but also, and most
importantly, accessing the meaning of a written message. The
‘reading aloud’ task that we chose for this study is useful
because it gives on-line access to the performance of the operated patient. The localization of small areas of the cortex
involved in reading becomes possible, but the whole process
of sentence or word understanding cannot really be assessed,
and must be far more complex.
Secondly, the question remains regarding whether reading
interference areas are truly necessary for language function and
need to be preserved. In almost all cases, no significant permanent postoperative reading deficits were seen in our patients
following the preservation of the reading areas identified by
direct stimulation. This does not prove, however, that these
areas are ‘essential’. The few cases in which the localized
reading areas were not spared do not allow one to draw
clear conclusions about this issue. Moreover, should some
‘essential’ areas be removed for surgical reasons, the consequences of these lesions could still be overcome by the functional reorganization and compensatory mechanisms that are
very likely to take place in such circumstances.
Finally, one of the main criticisms of the use of cortical
stimulation in brain mapping is that electrical stimulation of
the cortex is far from normal physiological brain activity. By
analogy, this criticism echoes the statement of Jackson on brain
lesions impairing speech: ‘to locate damage that destroys
1809
speech and to locate speech are two different things’ (Jackson,
1915). Stimulation brain mapping does not allow complete
exploration of all the neural structures involved in a complex
task such as reading. Moreover, although precisely localized
cortical stimulation is possible with this method, many cortical
areas remain inaccessible to direct stimulation either by sulcal
localization or by being beyond the surgical exposure. All these
factors are constraints of this method of brain mapping and
must be taken into account when analysing the results.
Conclusions
Intra-operative assessment of reading in some Indo-European
alphabetic languages has shown reading interference sites
mainly in dominant inferior and middle frontal, the supramarginal, the angular and the posterior portion of the superior and
middle temporal gyri. Not surprisingly, the inferior portions of
the pre- and postcentral gyri can also impair reading aloud
when stimulated (‘articulatory’ process interferences). Other
cortical areas can occasionally be implicated in reading processes, including the right hemisphere or frontal regions
controlling eye movements. Finally, in bilingual patients, if
common cortical areas were found, the reading process can
sometimes be localized to language- and reading-specific
areas. Although cognitive and neural processes involved in
reading are probably more complex than suggested by data
obtained from direct brain mapping, reading interference
sites can easily be identified by this method. Because there
was only partial overlap between naming and reading sites, we
think that it is important to use a reading task (in all languages
in which the subject is fluent) in addition to classic naming
tasks in patients that undergo direct brain mapping for brain
tumour or epilepsy surgery.
Acknowledgements
The authors wish to thank Denise Géblé for her help in supplying them with the references for this article, the speech therapists of the Federation of Neurology, Purpan Hospital, for
language evaluations, and Christiane Lubrano and Noëlle
Ricard for editorial assistance.
References
Adler A. Contributions to the classification and theory of aphasia. Beiträge zur
Kasuistik und Theorie der Aphasie, 1889. In: Paradis M, editor. Readings on
aphasia in bilinguals and polyglots. Montreal: Didier; 1983. p. 23.
Albert ML, Obler LK. The bilingual brain. New York: Academic Press; 1978.
Anderson SW, Damasio AR, Damasio H. Troubled letters but not numbers.
Brain 1990; 113: 749–66.
Assal G, Buttet J. Agraphia and preservation of music writing in a bilingual
piano teacher. [French]. Rev Neurol (Paris) 1983; 139: 569–74.
Beauvois MF, Derouesne J. Phonological alexia: three dissociations. J Neurol
Neurosurg Psychiatry 1979; 42: 1115–24.
Chee MW, Tan EW, Thiel T. Mandarin and English single word processing
studied with functional magnetic resonance imaging. J Neurosci 1999; 19:
3050–6.
Cohen L, Dehaene S, Naccache L, Lehericy S, Dehaene-Lambertz G, Henaff
MA, et al. The visual word form area: spatial and temporal characterization
1810
F.-E. Roux et al.
of an initial stage of reading in normal subjects and posterior split-brain
patients. Brain 2000; 123: 291–307.
Cohen L, Lehericy S, Chochon F, Lemer C, Rivaud S, Dehaene S. Languagespecific tuning of visual cortex? Functional properties of the Visual Word
Form Area. Brain 2002; 125: 1054–69.
Coslett HB, Saffran EM. Evidence for preserved reading in ‘pure alexia’. Brain
1989; 112: 327–59.
Coslett HB, Monsul N. Reading with the right hemisphere: evidence from
transcranial magnetic stimulation. Brain Lang 1994; 46: 198–211.
Cristal D. The Cambridge encyclopedia of language. 2nd ed. Cambridge:
Cambridge University Press; 1997.
Critchley M. Aphasia in polyglots and bilinguals. Brain Lang 1974; 1: 15–27.
Damasio AR, Damasio H. The anatomic basis of pure alexia. Neurology 1983;
33: 1573–83.
Dedic S. Zur Aphasiefrage. Z ges Neurol Psychiat 1926; 106: 208–13.
Déjerine J. Sur un cas de cécité verbale avec agraphie, suivi d’autopsie.
Mém Soc Biol 1891; 3: 197–201.
Déjerine J. Contribution à l’étude anatomo-pathologique et clinique
des différentes variétés de cécité verbales. Mém Soc Biol 1892; 3:
QJ;197–201.
Demonet JF, Chollet F, Ramsay S, Cardebat D, Nespoulous JD, Wise R, et al.
The anatomy of phonological and semantic processing in normal subjects.
Brain 1992; 115: 1753–68.
Demonet JF, Fiez JA, Paulesu E, Petersen SE, Zatorre RJ. PET studies of
phonological processing: a critical reply to Poeppel. Brain Lang 1996;
55: 352–79.
Fabro F. The neurolinguistics of bilingualism. Hove (UK): Psychology Press;
1999.
Fiez JA, Petersen SE. Neuroimaging studies of word reading. Proc Natl Acad
Sci USA 1998; 95: 914–21.
Gleb A. On medical psychology and philosophical anthropology. In:
Paradis M, editor. Readings on aphasia in bilinguals and polyglots.
Montreal: Didier; 1983. p. 383–4.
Hinshelwood J. Four cases of word-blindness. Lancet 1902; 1: 358–63.
Itziar L, Korostola LE. Aphasia manifestation in Basque. In: Paradis M, editor.
Manifestation of aphasia symptoms in different languages. Oxford:
Pergamon; 2001. p. 4–23.
Jackson JH. On the nature of the duality of the brain. Brain 1915; 38:
80–103.
Kawamura M, Hirayama K, Hasegawa K, Takahashi N, Yamaura A. Alexia
with agraphia of Kanji (Japanese morphograms). J Neurol Neurosurg
Psychiatry 1987; 50: 1125–9.
Kim KH, Relkin NR, Lee KM, Hirsch J. Distinct cortical areas associated with
native and second languages. Nature 1997; 388: 171–4.
Lhermitte R, Hécaen H, Dubois J, Culioli A, Tabouret-Keller A. Le problème
de l’aphasie des polyglottes: remarques sur quelques observations. Neuropsychologia 1966; 4: 315–29.
Luria AR. Differences between disturbances of speech and writing in Russian
and French. Int J Slavic Ling Poetics 1960; 3: 13–22.
Lyman R, Kwan S, Chao W. Left occipito-parietal brain tumour. Chinese Med J
1938; 54: 491–516.
Mesulam MM. Large scale neurocognitive networks and distributed
processing for attention, language, and memory. Ann Neurol 1990; 28:
597–613.
Minkowsky M. Considérations sur l’aphasie des polyglottes. Rev Neurol
(Paris) 1965; 112: 486–95.
Nakada T, Fujii Y, Kwee IL. Brain strategies for reading in the second language
are determined by the first language. Neurosci Res 2001; 40: 351–8.
Ojemann G. Cortical organization of language. J Neurosci 1991; 11:
2281–7.
Ojemann G. Intraoperative investigations of the neurobiology of reading. In:
von Euler C, Lundberg I, Llinas R, editors. Basic mechanisms in cognition
and language. Amsterdam: Elsevier; 1998.
Ojemann G, Whitaker HA. The bilingual brain. Arch Neurol 1978; 35: 409–12.
Ojemann G, Ojemann J, Lettich E, Berger M. Cortical language localization in
left, dominant hemisphere. J Neurosurg 1989; 71: 316–26.
Ojemann J, Ojemann G, Lettich E. Cortical stimulation mapping of language
cortex by using a verb generation task: effects of learning and comparison to
mapping based object naming. J Neurosurg 2002; 97: 33–8.
Oldfield RC. The assessment and analysis of handedness. The Edinburgh
Inventory. Neuropsychologia 1971; 9: 97–113.
Paradis M. Aspects of bilingual aphasia. Oxford: Pergamon; 1995.
Paulesu E, McCrory E, Fazio F, Menoncello L, Brunswick N, Cappa SF, et al.
A cultural effect on brain function. Nat Neurosci 2000; 3: 91–6.
Paulesu E, Demonet JF, Fazio F, McCrory E, Chanoine V, Brunswick N, et al.
Dyslexia: cultural diversity and biological unity. Science 2001; 291: 2165–7.
Penfield W, Roberts L. Speech and brain mechanisms. Princeton: Princeton
University Press; 1959.
Petersen S, Fox PT, Snyder AZ, Raichle ME. Activation of extrastriate and
frontal cortical areas by visual words and word-like stimuli. Science 1990;
249: 1041–4.
Pitres A. Etude sur l’aphasie des polyglottes. Rev Méd 1895; 15: 873–99.
Price CJ, Wise RJS, Watson JDS, Patterson K, Howard D, Frackowiak RSJ.
Brain activity during reading: the effects of exposure duration and task. Brain
1994; 117: 1255–69.
Roux FE, Trémoulet M. Organization of language areas in bilingual patients: a
cortical stimulation study. J Neurosurg 2002; 97: 857–64.
Rumsey JM, Horwitz B, Donohue BC, Nace KL, Maisog JM, Andreason P. A
functional lesion in developmental dyslexia: left angular gyral blood flow
predicts severity. Brain Lang 1999; 70: 187–204.
Sakurai Y, Sakai K, Sakuta M, Iwata M. Naming difficulties in alexia with
agraphia for Kanji after a left posterior temporal lesion. J Neurol Neurosurg
Psychiatry 1994; 57: 609–13.
Schäffler L, Lüders HO, Beck GJ. Quantitative comparison of language deficits
produced by extraoperative electrical stimulation of Broca’s, Wernicke’s,
and basal temporal language areas. Epilepsia 1996; 37: 463–75.
Scoresby-Jackson RE. Case of aphasia with right hemiplegia. Edinb Med J
1867; 12: 696–706.
Simos PG, Breier JI, Wheless JW, Maggio WW, Fletcher JM, Castillo EM, et al.
Brain mechanisms for reading: the role of the superior temporal gyrus in
word and pseudoword naming. Neuroreport 2000; 11: 2443–7.
Simos PG, Castillo EM, Fletcher JM, Francis DJ, Maestu F, Breier JI, et al.
Mapping of receptive language cortex in bilingual volunteers by using
magnetic source imaging. J Neurosurg 2001; 95: 76–81.
Stevens H. Reading epilepsy. N Engl J Med 1957; 257: 165–70.
Tan LH, Spinks JA, Feng CM, Siok WT, Perfetti CA, Xiong J, et al. Neural
systems of second language reading are shaped by native language.
Hum Brain Mapp 2003; 18: 158–66.
Taylor MD, Bernstein M. Awake craniotomy with brain mapping as the routine
surgical approach to treating patients with supratentorial intraaxial tumors:
a prospective trial of 200 cases. J Neurosurg 1999; 90: 35–41.
Taylor Sarno M, editor. Acquired aphasia. 3rd ed. San Diego: Academic Press;
1998.
Walker JA, Quinones-Hinojosa A, Berger MS. Intraoperative speech mapping
in 17 bilingual patients undergoing resection of a mass lesion. Neurosurgery
2004; 54: 113–8.
Wechsler A. Dissociative alexia. Arch Neurol 1977; 34: 257.
Yamadori A. Ideogram reading in alexia. Brain 1975; 98: 231–8.
Zar JH. Biostatistical analysis. 4th ed. Upper Saddler River (NJ): Prentice-Hall;
1999.