Brain Advance Access originally published online on July 7, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Brain, Vol. 127, No. 8, 1796-1810,
August 2004
© 2004 Guarantors of Brain
doi: 10.1093/brain/awh204
Intra-operative mapping of cortical areas involved in reading in mono- and bilingual patients
1 Institut National de la Santé et de la Recherche Médicale, Unité 455, 2 Federation of Neurosurgery, 3 Service d'Epidémiologie and 4 Federation 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: rouxfran{at}compuserve.com
Received August 7, 2003. Revised February 1, 2004. Second revision on March 23, 2004. Accepted March 29, 2004.
 |
Summary |
|---|
 Top Summary IntroductionMaterial and methodsResultsDiscussionReferences |
|---|
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 non-dominant 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.
Key Words: 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
|
Introduction |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
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 (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, 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 mono- and 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.
|
Material and methods |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Patients
Over a 5-year period (from December 1997 to December 2002), 76 language mappings were performed in our department using at least two standard tasks (naming and reading) in 54 patients (aged 13–76 years old, average 54 years; 23 male, 31 female patients) operated on for brain tumour or occasionally for other lesions (such as cortical dysplasia). Naming and reading sites were identified in order to spare them during surgery. Data from these successive brain mappings were prospectively 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, patients with brain tumours located in the right hemisphere were offered ‘awake surgery’ if mapping of the Rolandic area was required.
Hand dominance, language and reading testing
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):
- early acquisition (before the age of 7 years, considered as native language), high proficiency in language: L. I;
- late acquisition (after the age of 7 years), high proficiency in language: L. II;
- late acquisition (after the age of 7 years), low proficiency in language: L. III; or
- early acquisition (before the age of 7 years), low proficiency in 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), 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.
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; Radionics®, 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 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 pre-operatively 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 pre- and 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 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 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 |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
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 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:
- 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.
- ‘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.
- 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)n
- 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.
- 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.
|
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. Seventy-five 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%).
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 language-specific cortical sites were sometimes found in some patients, no statistical difference between first and second language (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 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 reading-aloud difficulties and slight facial palsy. Significant reading 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 |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
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 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.
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).
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 reading-specific 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 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 |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
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.[Medline]
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 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. Language-specific 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.[CrossRef][Web of Science][Medline]
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.[Medline]
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.[CrossRef][Medline]
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.[Medline]
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.[CrossRef][Web of Science][Medline]
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.[CrossRef][Medline]
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.[CrossRef][Web of Science]
Luria AR. Differences between disturbances of speech and writing in Russian and French. Int J Slavic Ling Poetics 1960; 3: 13–22.[Medline]
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.[CrossRef][Web of Science][Medline]
Minkowsky M. Considérations sur l'aphasie des polyglottes. Rev Neurol (Paris) 1965; 112: 486–95.[Medline]
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.[CrossRef][Web of Science][Medline]
Ojemann G. Cortical organization of language. J Neurosci 1991; 11: 2281–7.[Abstract]
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.[Web of Science][Medline]
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.[Web of Science][Medline]
Oldfield RC. The assessment and analysis of handedness. The Edinburgh Inventory. Neuropsychologia 1971; 9: 97–113.[CrossRef][Web of Science][Medline]
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.[CrossRef][Web of Science][Medline]
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.[Medline]
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.[Web of Science][Medline]
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.[CrossRef][Web of Science][Medline]
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.[CrossRef][Web of Science][Medline]
Scoresby-Jackson RE. Case of aphasia with right hemiplegia. Edinb Med J 1867; 12: 696–706.[Medline]
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.[Web of Science][Medline]
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.[Web of Science][Medline]
Stevens H. Reading epilepsy. N Engl J Med 1957; 257: 165–70.[Web of Science][Medline]
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.[CrossRef][Web of Science][Medline]
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.[Web of Science][Medline]
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.[CrossRef][Web of Science][Medline]
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.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  F. -E. Roux, V. Lubrano, V. Lauwers-Cances, C. Giussani, and J. -F. Demonet Cortical areas involved in Arabic number reading Neurology, January 15, 2008; 70(3): 210 - 217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G. Simos, J. M. Fletcher, S. Sarkari, R. Billingsley-Marshall, C. A. Denton, and A. C. Papanicolaou Intensive Instruction Affects Brain Magnetic Activity Associated with Oral Word Reading in Children with Persistent Reading Disabilities J Learn Disabil, February 1, 2007; 40(1): 37 - 48. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Crinion, R. Turner, A. Grogan, T. Hanakawa, U. Noppeney, J. T. Devlin, T. Aso, S. Urayama, H. Fukuyama, K. Stockton, et al. Language control in the bilingual brain. Science, June 9, 2006; 312(5779): 1537 - 1540. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||