One of the most important concerns for patients with brain tumors, who are a candidate for surgery, is to make sure that their brain function would be survived after the tumor resection. Brain tumors may affect different brain functions such as language, memory, sensory and motor functions, or other cognitive processes of the human. However, accurately predicting the degree of disability after surgical intervention is always a challenging task for the treatment team and of course the patient. In addition to the specific factors associated with the tumor features and the surgical process, the outcome depends on the complex and different aspects of each individual’s brain function. Surviving the patients’ speech is one of the most important challenges for a surgeon during brain tumor resection. Fortunately, today, the majority of the population can speak at least two languages, making the functional areas more insensible to local surgery inside the brain. Now, we know that the brain structures and functions have obvious differences in bilinguals compared to monolinguals [1-6]. Furthermore, brain lesions may alter expected language re-organization differently in bilingual patients than the monolinguals [7-9]. Previous studies have also shown significant changes in brain structure and function with second (L2) and third (L3) languages learning. It was shown that these changes are influenced by various factors such as; a) age of L2 learning, b) Level of proficiency in both languages, c) time of exposure to both languages, d) linguistic distance, e) frequency of language switching, f) modality of acquisition, and g) manner of acquisition [7,10].

Based on our observational experiences over the last few years in NIAG group, we met several bilingual patients who retained the ability to speak well after the surgery while their second language was mostly impaired. We assume that bilinguals and monolinguals do not follow a completely similar symptomatic and activation pattern during their illness or even after the surgical treatment. However, there is still a lack of evidence on how this process happens and what the extent of brain susceptibility for damage and/or repairment in bilinguals compared to monolinguals after a tumoral regression or resection.

We compared two groups of bilinguals and monolinguals having three sizes of lesional damage, known as non-invasion, partial-invasion, and severe- invasion, in order to find out different language reorganization capacities. This study aims to contribute to our understanding of brain function differences in response to brain tumors in monolingual and bilingual individuals. The results of this study would help us to understand the possible patterns of the first and second language brain networks in patients with brain tumors, and to predict the risk of damage to each patient during brain surgery and tumor resection.

Ethics

This study was performed by following the Ethics Statement of Tehran University of Medical Sciences. All participants declared their assent during the initial interview after being informed about the general aims of the study, and they signed their consent form on the test day.

Participants

We selected 26 right-handed subjects from patients referred to our neurosurgery department for four years. The handiness was assigned based on the Edinburgh Handedness Inventory assessment. All these patients were suffering from pre-Sylvian Intra axial lesions and were candidates for brain surgery. Initially, they were divided into two groups of 12 monolinguals and 14 bilinguals. The mean age of monolinguals was 38.3 and the mean age of bilinguals was 37.7. Each group of bilinguals and monolinguals have been classified into three subgroups, regarding the accurate location of the lesion relative to the Broca`s area, as a reference;

• Patients with a mass close to Broca`s area (It means that the tumor is located in the left frontal lobe, and the core or edematous area is not in attach with Broca)
• Patients whose Broca is partially invaded by the mass (It describes as a situation where Broca`s area is attached or surrounded by the edematous area, but the core of the tumor is still distinct from Broca)
• Patients whose Broca is completely invaded by the mass (it is defined as a Broca invaded by the core). We call these three groups “non-invasion”, “partial-invasion”, and “severe-invasion”, respectively, through this article. In each group, we had patients with mild to moderate aphasia, nearly equal (Table 1).

All the clinical examinations and para-clinical reports were performed by neurosurgeons, radiologists, or well-experienced Physicians in each related field. We examined pre-operative fMRI for all patients, from 48 hours to a week before brain surgery. The following inclusion and exclusion criteria were considered.

Inclusion Criteria and Conditions:

• Ages between 20 and 60 years old.
• All suffering from pre-Sylvian glioma, with different sizes and expansion, but classified into 3 groups, as explained above. However, we did not consider the subgroups of high, and low-grade tumors. Regarding the retrospective study we performed, all of the diagnoses could be confirmed by pathology reports, and patients with proven glial tumors were kept in this study.
• Pure Persian monolingual patients who were not able to speak another language or another dialect, for the first group.
• For the second group, bilingual patients speaking Turkish and Persian fluently, all with brain lesions characteristics similar to the Monolingual group. The native languages of bilingual patients were Turkish and their second language was Persian. All of them have learned their second language between 3 to 6 years old. It was very challenging to find a completely homogeneous group of bilinguals, moreover, it was more difficult to find well-matched monolinguals with confirmed similar glial tumor structures. However, regarding the considerable database of over 400 subjects, we were able to make a well-matched two groups in terms of factors influencing bilingual brain differences.
• All subjects had high proficiency in their languages, whereas bilinguals had learned L1 in their family and L2 in the pre-school period.

Exclusion Criteria:

• Other types of brain lesions (such as vascular lesions, extra-axial lesions, stroke, Infectious cystic lesions, and Metastasis). We had to remove 6 subjects after preparation of the histopathology report.
• Any evidence of previous brain surgery.
• Severe Aphasia.
• Having a third language at any period of life.
• History of substance abuse (smoking subjects were not excluded).
• Those who had a previous ability of secondary dialectic due to living in a multicultural country like Iran.

Those who had a weak performance in the task process, mainly due to the disabilities coming from brain lesions (eighteen patients were removed from initial selection).

Table 1: Bio-Imaging information of subjects with lesions in L.H.

Imaging

Magnetic Resonance Imaging of the brain was carried out using a SIEMENS 3 Tesla MRI scanner (MAGNETOM Trio; Siemens Healthcare GmbH, Federal Republic of Germany) with a 32-channel head coil at the Medical Imaging Center, Tehran University of Medical Science Hospital, and Tehran, Iran. During the scanning procedure, foam cushions were used to minimize movements of the head within the coil. Functional T2*-weighted images were collected using blood oxygen level-dependent (BOLD) contrast (TR= 3000ms, TE= 30ms, ?ip angle= 90degree, FOV= 192mm2, matrix size= 64×64, voxel size= 3×3×3mm, and slice gap= 0mm). Prior to the functional scan, a T1-weighted anatomical image was acquired, using a gradient echo pulse sequence (TR= 1800ms, TE= 3.44ms, ?ip angle= 7degree, voxel size= 1×1×1mm, FOV= 256mm2, matrix size= 256×256, and slice gap= 0mm).

FMRI

We have used the RWR (reverse word reading) task as a standard language task developed locally for clinical subjects in Farsi language [11]. Each Monolingual and Bilingual individual performed this task for Farsi letters. All functional and structural images were using DICOM format for offline workstations, and all images were transformed to NIfTI format to prepare for analysis (dcm2nii, MRIcron Ch. Rorden, http://www.nitrc.org/projects/mricon). Image Processing and statistical analysis were performed by FSL tools (FMRIB’s Software Library, www.fmrib.ox.ac.uk/fsl). First of all, pre-processing was done with the following steps: Using the BET (Brain Extraction Tool) for stripping the skull with a threshold of 0.25 [12], Motion correction, spatial- smoothing, high pass temporal filtering, and registration of functional images to the high-resolution structural images. First-level (single subject) analysis was done by FEAT (FMRI Expert Analysis Tool) Version 6.00, which used the GLM method for analysis. Double-gamma function was selected for HRF. And two first slices were removed as a part of motion cleaning data. A region with a threshold of a z-stat greater than 2.3 and a p-value less than 0.05 was been reckoned as an activated one. The next step of the investigation was second-level (between groups, bilingual vs. monolingual) analysis, and Contrasts of parameter estimates (COPs) from the previous stage were used for a mixed-effect method. This was completed by using the FMRIB Local Analysis of Mixed Effects Tool [13]. Statistical images were bounded using a cluster-corrected threshold of Z > 2.3 and p < 0.05 in the second level as well as the first one.

Laterality Index Estimation

The laterality index (LI) was calculated with a focus on the frontal lobe. The right and the left frontals’ masks were extracted by using MNI (Montreal Neurologic Institute) brain atlas. It has been shown previously that regional LI estimation methods would surpass other methods [14]. Numerous methods have been suggested to quantize language LI based on the activated brain region (i.e., the number of ‘‘active” voxels) or the detail of the fMRI signal change [15]. In this study, laterality indexes were calculated based on the active voxels within the ROI in each hemisphere. LI was approximated by the following formula: LI = (nL- nR) / (nL + nR), where nL and nR are the number of active voxels in the left and the right Frontal lobe, respectively. Reckoned LI with this expression lies between -1 and 1. As suggested by Briellmann values between - 0.2 and 0.2 were considered bilateral [16]. And also, LI value greater than 0.2 and less than -0.2 was thought-out as left-dominant and right-dominant language function, consecutively [16].

Both hemispheres were active during the task performance in every 6 subgroups, although left hemisphere (LH) was significantly dominant in 9 subjects, this was not unexpected for the right-handed patients. 15 patients had bilateral language function, and 2 subjects showed right hemisphere dominancy (Figure 1). Regarding invasive features of Glioma and probable decrease of bold signal around the core or in the edematous area, and neurovascular uncoupling, we have used the activation of the frontal lobe as a general eloquent motor area for estimation of LI. We observed higher LI in monolinguals, although it was accompanied by a sharp decline in LI within the 3 Monolinguals belonged groups, besides the mean LI decrease during these three sub-groups (non-invasion, partial-invasion, and severe- invasion) in both monolinguals and bilinguals. This finding suggests that due to the presence of a mass, mean LI decreased in all groups and this decrement continues in parallel as the tumor gets closer to the Broca`s, but Bilingual`s shifting to contralateral non-dominant language hemisphere was revealed from the non-invasion group, even when the mass is not yet quite infiltrate to the eloquent language area (Figure 1).

Figure 1: Lateralization indices alteration in the Bilinguals and Monolinguals. Left lateralization is reduced during tumor approaches to the language eloquent areas in both main groups but Monolinguals decrement is sharper (Close-Bi; non-invasion Bilinguals, Moderate-Bi; partially-invasion Bilinguals, Severe-Bi; severe-invasion Bilinguals, Close-Mono; non-invasion Monolinguals, Moderate-Mono; partially- invasion Monolinguals, Sever-Mono; sever-invasion Monolinguals).

In total, 34 activation areas were identified in subject groups including monolinguals and bilinguals. The occipital gyrus activation which is commonly stimulated (due to visual sense) during the RWR task performance was ignored (Table 2).

Table 2: Active regions during language task performance by fMRI study in bilingual and monolingual group analyses (Bi: Bilingual, Mono; Monolingual).

In an overview, bilinguals had more diversity in activated areas (Table 2). We observed that the pre-motor area, SMA, pre-central gyrus, superior parietal lobe, and inferior frontal gyrus were the most common active areas in all subjects. The supramarginal area, angular gyrus, superior temporal gyrus, and anterior cingulate cortex were activated with lesser frequency (Table 3).

Table 3: Activated regions based on clusters coming from the main feat/ fsl results in Bilingual, Monolingual, and Bilingual>Monolingual contrast groups analyses.

Our study revealed the frequency and priority of these language areas in terms of each eloquent group are considerably different, besides significant activation of cerebellum and ACC as the uncommon language areas in the bilingual group were notable. We found that basal ganglia and subcortical nuclei are significantly active in bilinguals and seem it is a potential ability in bilingual subjects who use it not only in double language speaking but also in the face of invasion since the subcortical areas present significant activation as the tumor gets closer to the Broca in bilinguals while we do not see this pattern in monolinguals (Figure 2). In the leading discussion, we will assess the main groups of bilinguals and monolinguals separately during their groups afterward to compare them together.

Figure 2: An initial model of language region recruitments in face of glial brain tumors infiltration in bilinguals and monolinguals. The green areas include: Left pre-central gyrus, left pre-motor cortex, left inferior frontal gyrus, and supplementary motor area. The yellow areas include; superior parietal lobe, supra-marginal gyrus, anterior cingulate cortex, and inferior parietal lobe. The blue areas include; sub-cortical areas, cerebellum, prefrontal cortex.

Although the left hemisphere is dominant in most populations even among left-handed subjects, we know today that hemisphere dominancy is a dynamic pattern. This fact implies that shifting between hemispheres could happen depending on a specific situation, during age, the presence of brain lesions, seizures, or even during the short time of learning a second language [7,17]. In our experience with glioma patients recruited in the current analyses, the left hemisphere was dominant in monolinguals, especially in the "non-invasion" group. Nevertheless, the laterality index began to decrease (toward bi-laterality) in the "partial-invasion" group and continues to more bi-laterality or even to right dominancy in the severe-invasion group (Figure 1). The mean LH index among monolinguals was higher than bilinguals which is in agreement with other reports [7]. However, laterality index values are significantly different between "non-invasion" and "severe- invasion" groups when comparing monolinguals with bilinguals.

Our finding implies that monolinguals shift language organization to a counter-lesion side just when at least the lesion starts to infiltrate the language eloquent area whereas, bilinguals` language areas shift initially to more ipsilateral pre-regulate areas during the lifetime, such as pre-frontal and even subcortical regions. This could be due to the fact that monolinguals are engaging higher activation in the left hemisphere compared to bilinguals. In other words, bilinguals have the ability to recruit different areas in both hemispheres for language organization [18,19]. However, we are not able to confirm the same pattern in the severe-group probably due to the small size of these groups (Figure.1). Bilinguals follow a balanced and continuous pattern of activation, and their reaction is seen even when the tumor has not completely attached the Language eloquent area. Stronger functional and structural connectivity in particular networks during the bilingual`s lifetime could be a substantial reason for this demonstration [20,21]. Our finding suggests that bilinguals tend to benefit from bilateral language organization. Moreover, they are able to do re-organization in the non- eloquent areas such as subcortical regions in case of invasion [22,23,8,9,18]. In contrast, monolinguals have higher localized activate areas in the left hemisphere they are more susceptible to losing their eloquent language area, especially in the eloquent language area when facing a severe invasion. These findings are in agreement with other recent studies [24,31].

Language Organization in Bilinguals and Monolinguals

Although there have been limited studies on bilinguals, the former studies focused more on the differences among L1, L2, L3, and L4 in bilingual and multi-lingual individuals. In this study, we have focused on differences between monolinguals and bilinguals in terms of re-organization in the presence and location of an invading lesion in the subsistence of different patterns of language areas. Different language activation areas have already been revealed by many fMRI or cortical stimulation studies for bilinguals and monolinguals [9,24-29] concurring with our tables (Table 3). The most common active language areas among all subjects are believed to be; left pre- central gyrus, left pre-motor cortex, left inferior frontal gyrus, supplementary motor area (SMA) [29-31]. The left and right superior parietal lobe, and left and right supra- marginal gyrus, and angular gyrus have also shown significant activities in most cases. However, active but less common anatomical sites among the subjects are believed to be areas such as the left and right anterior cingulate cortex (ACC), left inferior parietal lobe, cerebellum, left and right Insulars, and prefrontal cortex (DLPFC, VMPFC). Interestingly, subcortical areas are more significantly active in the bilingual group, especially when a lesion severely invaded them (as we found).

Non-Invasion Groups in Bilingual and Monolingual Patients

In Bilingual patients having no-invasion lesions not only the most activation of language eloquent areas was obtained, but also the most diverse language activity regions were seen since almost all areas that may be involved potentially in language organization were active. Furthermore, the significant activation of right hemisphere in these patients indicates the importance of the right hemisphere in bilingual patients. Significant cerebellum activation was revealed in almost all bilingual groups, regarding the initial role of the cerebellum in language verbal fluency, and grammatical processing, as well as memory function, especially in bilinguals [32,33]. A higher frequency of ACC activation was also found in this Bilingual group (compared to the same group in monolinguals), providing the lifelong competition of bilinguals’ brains for the stability of language based on the cognitive control development in higher activation and efficiency of ACC area [34]. The remarkable activity of subcortical areas is another noteworthy point in the bilingual group. In this regard, the obvious role of basal ganglia, in motor control, decision making, and a larger volume of subcortical gray matter caused by remodeling of activation in bilinguals proved by other studies in bilinguals [35-37,18] Our findings show the significant role of the subcortical area in language activity in parallel to former studies. In contrast, the monolinguals with similar non- invasion tumors mainly showed eloquent language area activations, especially in the left hemisphere, as well as limited activity in subcortical areas such as the left putamen and left insular, and right hemispheric non- frontal language areas such as right superior parietal lobe (Figure 3, Figure 4).

Figure 3: Axial views of the language activation areas in the mean bilingual and monolingual eloquent groups.

Partial-Invasion Groups in Bilingual and Monolingual Patients

In bilingual patients having partially invaded lesions, language eloquent areas are still shown to be dominant activations, besides the importance of non-eloquent areas in the left side such as the superior parietal lobe and, SMA. Moreover, parallel with switching to the right hemisphere, increasing activation in equivalent frontal areas on the right side is seen. In contrast, in the monolingual group with similar invading tumors, non-eloquent language areas showed a higher level of activity than the partially invaded eloquent areas (ie; BA44 and BA45). Right hemisphere switching is also seen in equivalent eloquent areas of the contralateral frontal lobe. Putamen activation as expected evidence of sensory-motor activation was revealed in non-invading and partially invading lesions [37].

Figure 4: Examples of three–dimensional slices showing the overlay of fMRI results in monolingual patients with non-invaded language eloquent areas by glioma in the left frontal lobe and bilingual patients with severely invaded eloquent language areas by glioma in the left frontal lobe.

Sever Groups in Monolingual and Bilingual

Given the small sample size in these groups, we explain the considerable evidence in each group individually. The highest shift to the right hemisphere was found in the severe-invasion bilinguals. Regions such as subcortical areas, right orbitofrontal cortex (OFC), right angular gyrus, right inferior parietal lobe, left ventromedial prefrontal cortex (VMPFC), left OFC, and left ACC were active (as shown in table 3, and figure 4). Significant structural changes in the frontostriatal network as well as a greater volume of subcortical area activation during the lifetime in bilinguals on one hand and neuroplasticity due to the lesion location in sever-invasion bilinguals, on the other hand, may be reasons for significant activation of subcortical and prefrontal areas in this group. The severe monolingual group continues to shift sharply to the right in the non-eloquent areas, especially in the parietal lobe, and then in the frontal areas without considerable activation in the subcortical areas.

Providentially, the subjects of our experience were those patients referring to our clinics at a certain point in time for pre-surgical planning using fMRI evaluations. Although these patients had some demographic heterogeneity, various lesion pathology, and different neuropsychological backgrounds, we could include quite well-matched subjects from the clinical environment in this study.

In summary, the priorities and possibilities of linguistic re-organization in the face of glioma infiltration were revealed in this study. It was shown that bilingual and monolingual individuals follow different, and potentially predictable patterns of language activation. These two groups of subjects, each with and without an invading lesion demonstrate different models of brain area recruitment. Bilinguals, as subjects with a more dynamic ability of brain re-organization, can modify language re-organization in tumor infiltration conditions. These patients rely mainly on subcortical and pre- frontal regions during brain activation. In comparison, monolingual individuals mainly keep classic language organizations as much as possible, but they rely mainly on the upper parietal areas with a shift to the right hemisphere where their language areas are more closely attacked by a tumor.

It is still unclear how much these facts would benefit the patients in clinics, especially for surgical intervention. A prospective pre- and post-operative, large sample size, and inter-clinic language mapping studies, seem to be essential. This is currently being organized by our team, aiming to justify various patterns of activation in bilinguals and monolinguals having different sizes and locations of lesions, helping with pre-surgical planning strategies.

The authors declare that have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the National Institute for Medical Research Development (NIMAD), [Grant number 964780].

1. Klein D, Mok K, Chen JK, Watkins KE. Age of language learning shapes brain structure: A cortical thickness study of bilingual and monolingual individuals. Brain and Language. 2014; 131: 20-24.
2. Abutalebi J, Canini M, Della Rosa PA, Sheung LP, Green DW, Weekes BS. Bilingualism protects anterior temporal lobe integrity in aging. Neurobiology of Aging. 2014; 35: 2126-2133.
3. Gold BT. Lifelong bilingualism and neural reserve against Alzheimer’s disease: A review of findings and potential mechanisms. Behavioural Brain       2015; 281: 9-15.
4. Gold CE. Exploring the Resting State Neural Activity of Monolinguals and Late and Early Bilinguals. 2018.
5. Rossi E, Cheng H, Kroll JF, Diaz MT, Newman SD. Changes in white-matter connectivity in late second language learners: Evidence from diffusion tensor imaging. Frontiers in Psychol. 2017; 8: 1-15.
6. Berken JA, Chai X, Chen JK, Gracco VL, Klein D. Effects of early and late bilingualism on resting-state functional connectivity. J Neurosci. 2016; 36: 1165-1172.
7. Po?czynska MM, Japardi K, Bookheimer SY. Lateralizing language function with pre-operative functional magnetic resonance imaging in early proficient bilingual patients. Brain and Language. 2017; 170: 1-11.
8. Krieg SM, Sollmann N, Hauck T, Ille S, Foerschler A, Meyer B, et al. Functional Language Shift to the Right Hemisphere in Patients with Language-Eloquent Brain Tumors. PLoS ONE. 2013; 8.
9. Roux FE, Lubrano V, Lauwers-Cances V, Tremoulet M, Mascott CR, Demonet JF. Intra-operative mapping of cortical areas involved in reading in mono- and bilingual patients. Brain. 2004; 127: 1796-1810.
10. Liu H, Cao F. L1 and L2 processing in the bilingual brain: A meta-analysis of neuroimaging studies. Brain and Language. 2016; 159: 60-73.
11. Mahdavi A, Houshmand S, Oghabian MA, Zarei M, Mahdavi A, Shoar MH, et al. Developing optimized fMRI protocol for clinical use: Comparison of different language paradigms. J Magnetic Resonance Imaging. 2011; 34: 413-419.
12. Smith SM. Fast robust automated brain extraction. Human Brain Mapping. 2002; 17: 143-155.
13. Beckmann CF, Jenkinson M, Smith SM. General multilevel linear modeling for group analysis in FMRI. NeuroImage. 2003; 20: 1052-1063.
14. Arora J, Pugh K, Westerveld M, Spencer S, Spencer DD, Todd Constable R. Language lateralization in epilepsy patients: FMRI validated with the Wada procedure. Epilepsia. 2009; 50: 2225-2241.
15. Jansen A, Menke R, Sommer J, Forster AF, Bruchmann S, Hempleman J, et al. The assessment of hemispheric lateralization in functional MRI-Robustness and reproducibility. NeuroImage. 2006; 33: 204-217.
16. Briellmann RS, Labate A, Harvey AS, Saling MM, Sveller C, Lillywhite L, Jackson GD. Is language lateralization in temporal lobe epilepsy patients related to the nature of the epileptogenic lesion?. Epilepsia. 2006; 47: 916-920.
17. Xiang H, Van Leeuwen TM, Dediu D, Roberts L, Norris DG, Hagoort P. L2-Proficiency-Dependent Laterality Shift in Structural Connectivity of Brain Language Pathways. Brain Connectivity. 2015; 5: 349-361.
18. Hull R, Vaid J. Bilingual language lateralization: A meta-analytic tale of two hemispheres. Neuropsychologia. 2007; 45: 1987-2008.
19. Roman P, Gonzalez J, Ventura-Campos N, Rodriguez-Pujadas A, Sanjuan A, Avila C. Neural differences between monolinguals and early bilinguals in their native language during comprehension. Brain and Language. 2015; 150: 80-89.
20. Grady CL, Luk G, Craik FIM, Bialystok E. Brain network activity in monolingual and bilingual older adults. Neuropsychologia. 2015; 66: 170-181.
21. Luk G, Bialystok E, Craik FIM, Grady CL. Lifelong bilingualism maintains white matter integrity in older adults. J Neuroscience. 2011; 31: 16808-16813.
22. Ille S, Engel L, Albers L, Schroeder A, Kelm A, Meyer B, et al. Functional reorganization of cortical language function in glioma patients-a preliminary study. Frontiers in Oncol. 2019.
23. Ruff IM, Petrovich Brennan NM, Peck KK, Hou BL, Tabar V, Brennan CW, et al. Assessment of the language laterality index in patients with brain tumor using functional MR imaging: Effects of thresholding, task selection, and prior surgery. American J Neuroradiol. 2008; 29: 528-535.
24. Leung LWL, Unadkat P, Bertotti MM, Bi WL, Essayed WI, Bunevicius A, et al. Clinical Utility of Preoperative Bilingual Language fMRI Mapping in Patients with Brain Tumors. J Neuroimaging. 2020; 30: 175-183.
25. Fernandez-Coello A, Havas V, Juncadella M, Sierpowska J, Rodriguez-Fornells A, Gabarros, A. Age of language acquisition and cortical language organization in multilingual patients undergoing awake brain mapping. J Neurosurg. 2017; 126: 1912-1923.
26. Po?czynska MM, Benjamin CFA, Japardi K, Frew A, Bookheimer SY. Language system organization in a quadrilingual with a brain tumor: Implications for understanding of the language network. Neuropsychologia. 2016; 86: 167-175.
27. Roux FE, Tretmoulet M. Organization of language areas in bilingual patients: A cortical stimulation study. J Neurosurg. 2002; 97: 857-864.
28. Sierpowska J, Fernandez-Coello A, Gomez-Andres A, Camins A, Castaner S, Juncadella M, et al. Involvement of the middle frontal gyrus in language switching as revealed by electrical stimulation mapping and functional magnetic resonance imaging in bilingual brain tumor patients. Cortex. 2018; 99: 78-92.
29. Li Q, Pasquini L, Del Ferraro G, Gene M, Peck KK, Makse HA, et al. Monolingual and bilingual language networks in healthy subjects using functional MRI and graph theory. 2019; 1-14.
30. Uebelacker LA. HHS Public Access. Physiology and Behavior. 2017; 176: 139-148.
31. Benjamin CF, Walshaw PD, Hale K, Gaillard WD, Baxter LC, Berl MM, Bookheimer SY. Presurgical language fMRI: Mapping of six critical regions. Human Brain Mapping. 2017; 38; 4239-4255.
32. Pliatsikas C, Johnstone T, Marinis T. Grey matter volume in the cerebellum is related to the processing of grammatical rules in a second language: A structural voxel-based morphometry study. Cerebellum. 2014; 13: 55-63.
33. Grogan A, Green DW, Ali N, Crinion JT, Price CJ. Structural correlates of semantic and phonemic fluency ability in first and second languages. Cerebral Cortex. 2009; 19: 2690-2698.
34. Kroll JF, Bobb SC, Hoshino N. Two Languages in Mind: Bilingualism as a Tool to Investigate Language, Cognition, and the Brain. Current Directions in Psychological Science. 2014; 23: 159-163.
35. Frank MJ. Hold your horses: A dynamic computational role for the subthalamic nucleus in decision making. Neural Networks. 2006; 19: 1120-1136.
36. Burgaleta M, Sanjuan A, Ventura-Campos N, Sebastian-Galles N, Avila C. Bilingualism at the core of the brain. Structural differences between bilinguals and monolinguals revealed by subcortical shape analysis. NeuroImage. 2016; 125: 437-445.
37. Stocco A, Prat CS. Bilingualism trains specific brain circuits involved in flexible rule selection and application. Brain and Language. 2014; 137: 50-61.