Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 1.
Published in final edited form as: Brain Lang. 2024 Jan 31;250:105380. doi: 10.1016/j.bandl.2024.105380

Neural specificity for semantic and syntactic processing in Spanish-English bilingual children

Neelima Wagley 1,2,*, Xiaosu Hu 2, Teresa Satterfield 3, Lisa M Bedore 4, James R Booth 1, Ioulia Kovelman 2
PMCID: PMC10947424  NIHMSID: NIHMS1963626  PMID: 38301503

Abstract

Brain development for language processing is associated with neural specialization of left perisylvian pathways, but this has not been investigated in young bilinguals. We examined specificity for syntax and semantics in early exposed Spanish-English speaking children (N=65, ages 7–11) using an auditory sentence judgement task in English, their dominant language of use. During functional near infrared spectroscopy (fNIRS), the morphosyntax task elicited activation in left inferior frontal gyrus (IFG) and the semantic task elicited activation in left posterior middle temporal gyrus (MTG). Task comparisons revealed specialization in left superior temporal (STG) for morphosyntax and left MTG and angular gyrus for semantics. Although skills in neither language were uniquely related to specialization, skills in both languages were related to engagement of the left MTG for semantics and left IFG for syntax. These results are consistent with models suggesting a positive cross-linguistic interaction in those with higher language proficiency.

Keywords: Brain, language, bilingualism, syntax, semantics

1. Introduction

Comprehending a sentence requires integrating both syntactic and semantic information and relies on a distributed network of brain regions involved in these processes (Enge et al., 2020). The connections between network regions are strengthened throughout development and activations narrow as individual brain areas begin to specialize for different language subfunctions (see Interactive Specialization account by Johnson, 2011). Variations in language experiences, such as early exposure to two languages, influence this dynamic process (Claussenius-Kalman et al., 2021; Werker & Hensch, 2015). Yet, much of our understanding regarding language specialization in the brain has been guided by research with monolingual speakers (Frederenko et al., 2020; Hagoort & Indefrey, 2014; Skeide et al., 2014; Wang, Rice, & Booth, 2020; Wang et al., 2021; Zaccarella et al., 2017). The current study is the first to complement and expand upon prior monolingual work by investigating bilingual children’s emerging functional brain organization for syntactic and semantic language processes in English, their dominant language of academic instruction and use.

Friederici’s (2012) language comprehension model proposes that the neural bases of syntax and semantics gradually separates, or specializes, over development (Brauer & Friederici, 2007; Hagoort & Indefrey, 2014; Skeide et al., 2014). Based on this perspective, the superior temporal gyrus (STG) is the region most engaged in both semantic and syntactic processing roughly until age seven. Around 9- to 10-years-old, specificity for semantic processes is thought to emerge in the left middle temporal gyrus (MTG) and in the left inferior frontal gyrus pars triangularis (IFGtri; Binder et al., 2009) while specificity for syntactic processes is thought to emerge in the left IFG pars opercularis (IFGop; Goucha & Friederici, 2015). Alternative perspectives argue for no such selectivity, but rather an integration between syntactic/combinatorial and lexico-semantic processing during comprehension (Frederenko et al., 2020; Matchin & Hickok, 2020; Pylkkänen, 2019). As such, there is mixed evidence in support of a semantic and syntactic separation within the literature focused on monolingual adults, and limited research addressing this issue in monolingual developmental work.

Wang and colleagues systematically examined functional specialization in monolingual English-speaking children ages 5 to 6 (Wang et al., 2020), 7 to 8 (Wang et al., 2021), and 9 to 10 (Wang et al., in prep.) using fMRI. All three studies used a double dissociation paradigm that contrasted semantic plausibility and morphosyntax sentential grammaticality judgement tasks, stimuli nearly identical to those used in the current study. At ages 5 to 6, results indicate evidence in support of semantic specialization in the left MTG and sensitivity for both syntax and semantics in the left STG (Wang et al., 2020). At ages 7 to 8, there was evidence in support of semantic specialization in the left MTG as well as the left IFGtri (Wang et al., 2021). At ages 9 to 10, results indicate evidence in support of syntactic specialization in the IFGop and sensitivity to both syntax and semantics in the left STG, MTG, and IFGtri (Wang et al., in prep.). These results are generally consistent with Friederici’s (2012) functional specialization model, with specificity for semantics progressing from the temporal lobe to the frontal lobe around age seven and specificity for syntax emerging in the frontal lobe around age nine. Building upon this model and the monolingual findings, the present study investigates syntactic and semantic specialization in the left frontotemporal language network in a group of Spanish-English bilinguals exposed to both languages early in development.

Variations in dual-language experiences and proficiency are thought to contribute to bilingual individuals’ neural organization for language function in both languages (Costa & Sebastián-Galles, 2014; Claussenius-Kalman et al., 2021; Li et al., 2014). Bilingual dual first language learners are children who acquire two languages in early life in immersive contexts. Heritage language learners are dual first language learners exposed to one language in the home, typically from birth, and a different language in the community, such as English in the U.S. (Ortiz et al., in press). As their bilingual experience falls within the early periods of brain development for language function, heritage language bilinguals are often believed to have the possibility of developing native-like proficiency and neural organization for each of their languages (Sulpizio et al., 2020; Weber-Fox & Neville, 2001). Nevertheless, language and academic development in bilingual immigrant environments can also be challenging (Hoff, 2018). Our work examines neural organization for language function in young bilingual heritage language speakers during the important periods of language and brain development.

Bilingual speakers’ language proficiency is related to the neural organization for each of their two languages. For instance, research has shown that bilingual Spanish-English infants raised in the U.S. who have greater experiences in English, show a more specialized neural response to phonetic contrasts in English and vice-versa for Spanish (Garcia-Sierra et al., 2016). On the other hand, experiences across bilinguals’ two languages may also have an additive effect (Chung et al., 2019). Sun et al. (2022) examined functional connectivity in bilingual Spanish-English and Chinese-English children educated in the U.S. and showed that children with better English language skills had stronger functional connectivity during English word recognition tasks. Bilinguals with stronger heritage language skills, either in Spanish or Chinese, also showed stronger functional connectivity during the English language task after accounting for English language proficiency. These neuroimaging findings are consistent with theoretical models of bilingual development suggesting that bilinguals’ two languages interact, yielding both language-specific as well as cumulative influences on children’s acquisition and processing of each of their languages (Chung et al, 2019). In the current study, we examine the relation between young heritage language speakers’ proficiency in each of their languages, the heritage language (Spanish) and the language of academic instruction (English), on the neural organization for language function in English.

The brain’s early organization for phonological processing lays the foundation for subsequent language development, such as lexical and morphosyntactic growth (Kuhl et al., 2008; Petitto et al., 2012; Zhao & Kuhl, 2022). Certain morphemes such as past-tense -ed or third-person present singular verb form -s are low in phonetic salience and are structures of difficulty for English-speaking children. Different structures low in phonetic salience, such as plural markers or adjective agreement, are also difficult for Spanish-speaking children (Bedore & Leonard, 2001; Cooperson et al., 2013). The acquisition of inflectional morphology in Spanish-English bilinguals is language-specific and dependent on experiences in each of the languages (Peña et al., 2015). At the same time, prior work has shown that Spanish-English bilinguals benefit from the semantic knowledge of shared Latin cognates (Kuo et al., 2017; Ramírez et al., 2013) and can transfer elements of their syntactic competence between the two languages (Dijkstra et al., 2019; Kroll et al., 2015). Bi-directional language transfer effects are more likely to influence bilinguals with early exposure and relatively balanced proficiency (in contrast, unidirectional L1 to L2 transfer is more likely to influence later exposed and less balanced bilinguals; DeLuca et al., 2019; Román et al., 2015; Sulpizio et al., 2020). As a result of cross-linguistic influences, more balanced bilinguals may develop greater automaticity for the processing of shared lexical elements and greater sensitivity to those word structures that are unique to each language (Sun et al., 2022). Based on this previous research, greater dual-language proficiency may be related to differences in the neural specialization for semantics and syntax. By examining early exposed Spanish-English heritage language bilinguals, the current study aims to investigate how proficiency in either Spanish or English influences specialization for semantics and syntax.

Developmental bilingual neuroimaging research on sentence processing has primarily focused on activation pattern differences between bilinguals and monolinguals. Jasinska and Petitto (2013) asked children ages 7–10, and adults, to complete an English sentence judgment task with varied syntactic complexity using fNIRS. Consistent with findings suggesting temporo-frontal trajectory of brain development for language function (Enge et al., 2020), both bilingual and monolingual children exhibited a stronger response to the more complex relative to the simpler sentences in the temporal lobe, whereas only adults showed a stronger response in the left IFG. Ding et al. (2021) examined brain activation in monolingual and Chinese-English bilingual children ages 9–11 during an English sentence processing task using fNIRS. Sentences were controlled for semantic cues such that participants had to judge the sentence based on syntactic knowledge. They found an interaction effect between language group and sentence structure type in the inferior frontal cortex. Bilingual children had greater hemodynamic responses to object-relative clauses (dominant sentence type in Chinese) than to subject-relative sentences (canonical sentence type in English), suggesting that their Chinese knowledge influenced sentence processing in English. Arredondo et al. (2018) also used fNIRS to assess specific engagement of the left IFG during an English morphosyntax judgment task in Spanish-English bilingual children. Seven to 11-year-old bilinguals showed stronger and more restricted left IFG activation in comparison to skill-matched monolingual peers, suggesting a possible link between the localized activation for syntax and the enriched linguistic experiences with two languages from early in development. These overall group differences have been attributed to various factors that are unique to bilinguals, including dual-language co-activation and language-switching (Hernandez, 2009; Sulpizio et al., 2020).

The attention of prior research has primarily been on bilingual versus monolingual group comparisons, and therefore we need to know how dual-language proficiency influences semantic versus syntactic specialization within bilinguals. Bilingual and monolingual groups typically vary in socio-cultural and socio-linguistic contexts of development. To better understand the effects of dual language proficiency and use on neural specialization for language function, we examine these effects on a continuum within one group of early exposed Latinx bilingual Spanish-English speakers in the U.S. (DeLuca et al., 2019; Luk & Bialystok, 2013; Rothman et al., 2023). Based on the prior literature, we hypothesized that bilingual children would exhibit semantic specialization in the left MTG and left IFGtri and syntactic specialization in the left IFGop. We also expected the STG to be sensitive to both semantic and syntactic processing. Although previous literature did not allow for principled predictions, we expected that the degree of semantic and syntactic differentiation in specialized areas may be related to children’s dual-language proficiency.

2. Method

Sixty-five Spanish-English speaking bilingual children are included in the final analyses (Mage= 8.6, range = 7–11.83 years; Table 1). Data from those who met the following selection criteria were analyzed from an initial sample of 123 who participated: exposure to Spanish from birth and English before age five, at least one native Spanish-speaking parent with only Spanish or English spoken at home, right-handed, normal hearing and normal or corrected vision, no known history of head injury, no developmental delays, not being treated by psychotropic medications at the time of testing. Participants were also excluded for missing data and poor task performance as described below.

Table 1.

Participant demographics and task performance

M (SD)
Age (years) 8.6 (1.2)
Age of first words in English (years) 2.4 (1.3)
Grade 3.0 (1.4)
Average Spanish/English Use1 28/72%
Attended Heritage Language School N=13
Free/Reduced Lunch Program N=19
Bilingual English-Spanish Assessment (BESA-ME) 2
 English Semantic Knowledge (out of 23) 83% (12)
 Spanish Semantic Knowledge (out of 26) 66% (20)
 English Morphosyntax Knowledge (out of 46) 94% (05)
 Spanish Morphosyntax Knowledge (out of 51) 70% (20)
Semantic Plausibility fNIRS Task 3 91 (6)
 Strongly Congruent 90 (9)
 Weakly Congruent 89 (10)
 Semantically Inflectious 95 (6)
Syntax Grammaticality fNIRS Task 3 89 (5)
 -ED&S Omission 78 (12)
 -ING Omission 96 (5)
 Gramatically Correct 93 (8)
Open in a new tab
1

Bilingual Input Output Survey (BIOS); Peña et al. (2018)

2

Percent correct calculated from raw scores

3

Task accuracy in percentage; conditions of interest are bolded.

Families were recruited in Southeast Michigan, USA by a community liaison. This geographical region of the country is composed of majority White and English-dominant communities. English was the primary language of instruction at school for all participants. However, 20% of the participants (N=13) regularly attended a heritage language school on the weekends which exposed them to academic instruction in Spanish and assigned Spanish language and literacy homework. Participants came from middle-class homes with a median household income on par with the surrounding county-level and national-level norms (U.S. Census Bureau, 2019). Most of the participants (~70%) had at least one parent who held a bachelor’s degree or higher, indicating relatively high educational attainment.

2.1. Procedures

Parents completed an eligibility form and the Bilingual Input Output Survey, which asked parents to describe the quantity of their child’s language use to the best of their ability (details reported in Baron, Wagley, Hu, & Kovleman, 2023; Peña, et al., 2018; see Supplementary Figure 1). Eligible participants were invited for a lab visit which included language and literacy assessments in English and Spanish, experimental tasks during functional near infrared spectroscopy (fNIRS) imaging, and parent questionnaires (see Wagley et al., 2022). Participants were compensated $30 and a toy.

2.2. Bilingual English Spanish Assessment – Middle Extension (BESA-ME)

Children’s language comprehension skills were measured using two subtests of the BESA-ME (Peña et al., 2016) in each language: Semantic Knowledge assessing category generation, similarities and differences, and analogies and Morphosyntax Knowledge examining knowledge of grammatical morphemes and sentence structure using cloze task and sentence repetition items see Supplementary Figure 2).

2.3. fNIRS Experimental Tasks

In both sentence tasks, participants heard 60 sentences (20 per condition). Detailed description of task development and example stimuli can be found at https://osf.io/bh9n8/.

2.3.1. Semantics Plausibility Task

There were three conditions of the semantic sentence stimuli: strongly congruent (SCon; sing-song), weakly congruent (WCon; catch-fish), and semantically infelicitous (InCorr; bounce-paper). Each experimental condition varied in the degree of semantic association between the verb and object pairing embedded within the sentence (adapted from the sentence stimuli in Wang et al., 2022). These verb-object pairs are embedded within a sentence context that naturally elicits a biased interpretation based on experiential knowledge (e.g., She is singing a song; He did not catch any fish; They are bouncing the paper).

2.3.2. Morphosyntax Grammaticality Task

There were three conditions of the syntax sentence stimuli: omission of the -ed or -s morpheme (-ED&S; Laura score_ a winning goal; Yesterday, they finish_ all of the homework), omission of the -ing inflectional morpheme (-ING; Right now, he is walk_ his dog), and grammatically correct (Corr). Ten sentences in the ED&S condition are missing the –s person ending (e.g. Nicholas bite_ into a pizza) and ten sentences are missing the –ed past tense ending (e.g. Last week, they laugh_ with grandma). Of the 20 correct sentences, ten sentences include –ing endings, five sentences include –ed endings, and five sentences include verb –s endings.

Participants with complete imaging data for both sentence tasks and those who met the following criteria were included in the analysis: (1) at least 50% accuracy on both the strongly congruent and grammatically correct conditions and (2) no greater than 50% difference in accuracy between the strongly congruent and semantically infelicitous conditions for the semantic task, and the grammatically correct and -ED&S omission conditions for the syntax task.

2.4. fNIRS Data Acquisition & Analysis

We utilize fNIRS as it puts fewer constraints on natural movements (e.g., head tilts), is quiet as compared to fMRI, is lower in cost, and has been gaining traction in studying pediatric and bilingual populations (Arredondo, 2023; Nickerson & Kovelman, 2022). A TechEN-CW6 system with 690 and 830 nm wavelengths was used for data acquisition. The fNIRS cap included 5 emitters of near-infrared light and 8 detectors spaced ~2.7 cm apart, yielding 16 data channels per hemisphere (see Figure 1). Sources and detectors were mounted onto two custom-built head caps, one 54 cm and one 56 cm. The average head circumference of participants was 53.8 cm (SD=1.3, range=51–57) and most participants used a cap size of 54 cm (N=57 participants). To establish neuroanatomical precision of probe placements using our fNIRS cap, we conducted a separate study using MRI and photogrammetry-based stereoscopic optode registration described in (Hu et al., 2020).

Figure 1.

Figure 1.

Open in a new tab

A shows the probeset with light sources in red and detectors in blue. Patterns of brain activation for the within-tasks contrasts are shown in C and D. Patterns of brain activation for the between-tasks contrast is shown in B. Results shown are at p<.05 corrected.

We used the NIRS Toolbox (Santosa et al., 2018) and a set of customized scripts based on Hu et al., 2010) to analyze the data. At the subject-level, the following preprocessing steps were applied: trimming of raw data file to keep 5 seconds of pre- and post- experimental task baseline data, resampling the data from 50Hz to 5Hz, optical density and hemoglobin concentration change data conversion using the modified Beer-Lambert law. Data was analyzed using a fixed-effects GLM, assuming the dual-gamma canonical hemodynamic response function peaking 6-seconds after trial onset. This yielded estimated HbO (oxygenated) and HbR (deoxygenated) hemoglobin beta values for each participant, each condition, and each channel. We used a pre-whitening autoregressive filter combined with a weighted least square estimation approach to eliminate the non-spherical noise structure caused by physiological and motion artifacts in the time series. We analyzed both HbO and HbR time series, but primarily report on HbO as it accounts for a greater portion (~75%) of the total fNIRS signal (Gagnon et al., 2012) and several studies have found that HbR signals are particularly susceptible to noise when using fNIRS (Hoshi, 2007; Strangman et al., 2002). The signal quality acquired by the TechEN system used for this study was experimentally established to be superior for the HbO relative to the HbR. Thus, HbR results are reported in the Supplementary Table 1.

Group-level analyses used a linear mixed-effects model for each data channel. The group-level model included the 3 conditions for each experimental task as fixed effects, participants as a random effect variable, and hemoglobin beta values as the dependent variables.

Estimated group-level beta values were extracted for each channel for each of the following contrasts: ED&S>CORR for the syntax task, InCorr > SCon for the semantics task, and ED&S > InCorr as a between-task comparison. These contrasts we chosen based on prior work using sentences with semantic and syntactic errors or non-canonical syntactic structure (e.g., Brauer & Friederici, 2007; Skeide et al., 2014; Skeide & Friederici, 2016; Wang et al., 2021; Wang et al., in prep.). Specifically, the -ED&S omission errors are of relevance as bilingual children from different language backgrounds (e.g., Spanish, Catalan, Italian) commonly omit tense and agreement as developmentally appropriate for their respective grammars (e.g., Tsakali & Wexler, 2004; Gavarró, Torrens, & Wexler, 2010). The semantically infelicitous sentences were chosen to be methodologically parallel to the -ED&S omission sentences as they are ungrammatical or semantically implausible sentences that require a ‘no’ judgment response. All correct and incorrect response trials were included in the analysis. Group-level results (unstandardized beta) for each contrast were plotted on to the MNI 152 brain template using the specified MNI coordinates in Hu et al. (2020). All secondary analyses were computed in R (R Core Team, 2020).

3. Results

Performance on the bilingual language assessments and neuroimaging tasks are reported in Table 1. Participants had developmentally age-appropriate language scores in at least one of their languages, with the majority at or above average in both English and Spanish. As expected, English assessment scores were higher than Spanish for the semantic (t(63)=8.29, p<.0001) and morphosyntax (t(62)=9.45, p<.0001) knowledge subtests of the BESA-ME. For both experimental tasks, percent accuracies across all three experimental conditions were significantly greater than chance (p<.001).

Within-task brain activation results are presented in Figure 1C and Figure 1D and corresponding statistics are reported in Table 2. Sentences requiring a grammaticality judgment (ED&S> CORR) elicited significant activation in the left IFG, including the opercular, triangular, and orbital regions, as well as the supramarginal gyrus (SMG). Sentences requiring a semantic plausibility judgment (InCorr>SCon) elicited significant activation in the posterior left MTG and the left IFG pars opercularis.

Table 2.

Group-level HbO result statistics for within- and between- task contrasts

ROI Ch. β SE t-stat df p (corr.)
Within Tasks
SYN (ED&S > Correct) IFG orbitalis 1 1.28 0.23 5.49 381 0.000
IFG triangularis 2 1.29 0.23 5.55 381 0.000
IFG triangularis 3 0.92 0.32 2.85 381 0.017
IFG opercularis 4 1.39 0.24 5.84 381 0.000
SMG 8 1.04 0.27 3.82 381 0.001
SEM (InCorr > SCon) IFG opercularis 4 0.82 0.25 3.22 381 0.006
MTG 9 3.60 0.33 11.05 381 0.000
MTG 11 1.09 0.28 3.96 381 0.001
MTG 13 1.03 0.33 3.13 381 0.007
Between Tasks
ED&S > InCorr STG (syn>sem) 5 2.01 0.54 3.70 381 0.004
STG (syn>sem) 7 1.76 0.52 3.38 381 0.006
STG/MTG (sem>syn) 9 −2.16 0.65 −3.34 381 0.006
STG/MTG (sem>syn) 13 −3.17 0.64 −4.94 381 0.000
AG (sem>syn) 14 −1.96 0.74 −2.65 381 0.034
Open in a new tab

To assess brain-behavior relations for each construct, we correlated the within-task brain activation in the IFG for the syntax task and the MTG for the semantics task with the corresponding morphosyntax or semantic behavioral raw scores on the BESA-ME, while controlling for age. Brain activation in the pars opercularis region of the IFG (Ch 4) for syntax correlated with Spanish (r=−0.35, p=.006) but not English (r=−0.14, p=.281) morphosyntax knowledge at p<.006 correcting for multiple comparisons). All other regions in the IFG cluster for syntax were non-significant. Brain activation in the posterior MTG (Ch 11) for semantics correlated with both Spanish (r=−0.24, p=.058) and English (r=−0.33, p=.010) semantic knowledge, but did not meet the p<.008 threshold for multiple comparisons. All other regions in the MTG cluster for semantics were non-significant.

Between-task specialization results are presented in Figure 1B, and corresponding statistics are reported in Table 2. As predicted, there was significantly greater activation for syntax than semantics in the more anterior left STG and greater activation for semantics than syntax in the more posterior the STG adjacent to the MTG. Semantic specialization was also evident in the left posterior MTG and the angular gyrus (AG). Contrary to our hypothesis, there was no evidence in support of either semantic or syntactic specialization in the frontal lobe.

To assess the effects of bilingual proficiency on children’s patterns of neural specificity, we correlated children’s language skills with between-task activations. First, we created three clusters of interest by averaging the significant regions showing specialization: STG for syntax (Ch 5 & 7), STG/MTG for semantics (Ch 9), and MTG/AG for semantics (Ch 13 & 14). Activations in these clusters were then correlated with Spanish morphosyntax knowledge, while controlling for English morphosyntax knowledge and age. The same was also computed with Spanish semantic knowledge, while controlling for English semantic knowledge and age. The STG cluster showing syntactic specialization was not related to Spanish semantic (r=0.04, p=.754) or morphosyntax (r=0.06, p=.631) knowledge, while controlling for English knowledge and age. The STG/MTG cluster showing semantic specialization was not related to Spanish semantic (r=0.07, p=.573) or morphosyntax (r=0.14, p=.268) knowledge, while controlling for English knowledge and age. The MTG/AG cluster showing semantic specialization was not related to Spanish semantic (r=−0.03, p=.842) or morphosyntax (r=0.09, p=.474) knowledge, while controlling for English knowledge and age.

4. Discussion

The present study examined semantic and syntactic specialization during English sentence comprehension in early exposed Spanish-English speaking bilingual children ages 7 to 11 years old. Across the whole group, we found specificity for semantic processing in the left MTG and specificity for syntactic processing in the left STG. These results are largely consistent with the prior monolingual findings in children of similar age ranges (Skeide et al., 2014; Wang et al., 2021; Wang et al., in prep.). Contrary to the language model proposed by Friederici (2012), we did not observe semantic or syntactic specialization in the frontal lobe, but developmental models argue that language processing in the temporal lobe is earlier maturing (Enge et al., 2020). In addition, children’s bilingual language skills were not uniquely related to areas of specialization. However, bilingual language skills were related to overall engagement of the left MTG for semantics and engagement of the left IFGop for syntax. Together, our findings show evidence for specialization of the temporal cortex for semantic and syntactic processing, and that language skills in both languages are related to how robustly this network is activated.

During the semantic plausibility sentence judgement task, bilingual children showed activation in the left MTG as well as the left IFG pars opercularis. This pattern of activation is in line with the prior literature which suggests localization of lexical-semantic processes during sentence comprehension in the MTG and IFG including the pars triangularis, orbitalis, and opercularis (Binder, et al., 2009; Brauer & Friederici, 2007; Hagoort & Indefrey, 2014; Hogdson et al., 2021; Wang et al., 2020; 2021). Early developing temporal lobe structures support direct mapping between sounds to meaning in the MTG whereas the IFG supports controlled processes involved in meaning judgments or plausibility categorization (Thompson-Schill et al., 1997; Binder et al., 2009). Our analysis compared functional activation for the semantically infelicitous with strongly congruent sentences, while maintaining the same sentence structure. Activation observed in the IFG during this plausibility judgement comparison may reflect accessing stored semantic knowledge in the temporal cortex (Thompson-Schill et al., 1997; Lau et al., 2008).

Activation in the posterior MTG was negatively related to children’s Spanish and English semantic knowledge assessed behaviorally. Thus, children with better semantic knowledge in Spanish and English showed less activation in the left MTG during the semantic judgment task. The MTG correlations did not reach significance when correcting for multiple comparisons, thus further investigations are necessary to replicate these effects. Nevertheless, similar effects have been previously reported during single-word lexical morphology tasks in English in comparable populations of young Chinese-English bilingual children growing up in the United States (Ip et al., 2017). It is therefore possible that reduced MTG activation relates to better lexical processing in young bilinguals.

During the morphosyntax grammaticality judgment task, bilingual children showed broad activation across the opercularis, triangularis, and orbitalis regions of the left IFG, like previous research on morphosyntactic processing in monolinguals (e.g., Nuñez et al., 2011; Schneider & Maguire, 2019; Wu et al., 2016) and Spanish-English bilinguals (e.g., Arredondo et al., 2019; Baron et al., 2023; Kovelman et al., 2008). A recent meta-analysis on language comprehension found consistent activation in the left ventral IFG (pars triangularis) in children, whereas activation peaked in the left dorsal IFG (pas opercularis) in adults (Enge et al., 2020). The shift from more ventral to dorsal IFG activation is thought to be related to the increase in sensitivity to syntactic information in the developing brain (Enge et al., 2020). Our analysis included sentences with -ed and -s inflectional morpheme errors in which a “no” response was required for the sentence judgement, a relatively difficult condition with the lowest task accuracy. These sentences were contrasted with grammatically correct sentences. Processing grammatically incorrect sentences has shown to place additional demands on the cognitive system, particularly in children (Schneider et al., 2016). Thus, the ventral IFG activation observed in these bilingual children may reflect more general top-down comprehension processes during language processing (Hogdson et al., 2021; Skeide & Friederici, 2016).

We also observed activation in the dorsal IFG, a key region that supports syntactic computations (Goucha & Friederici, 2015; Nuñez et al., 2011; Zaccarella et al., 2017; Walenski et al., 2019). Activation in the IFGop was negatively related to bilingual children’s Spanish and English morphosyntax knowledge, though only the former was significant. Thus, children with better morphosyntax knowledge, possibly in both languages, showed less activation in the IFGop when processing English sentences. Bilingual children in this study had relatively high English morphosyntax knowledge but had greater variability in their Spanish morphosyntax knowledge (see Table 1). Thus, significant brain-behavioral correlations with Spanish, but not English, morphosyntax knowledge during English morphosyntactic processing may indicate some evidence of cross-linguistic influence.

It was hypothesized that bilingual children proficient in both Spanish and English would show semantic specialization in the left MTG and the left IFG triangularis, syntactic specialization in the left IFG opercularis, and that the STG would be sensitive to both processes. In support of our hypotheses, we observed semantic and syntactic specialization in the left STG and semantic specialization in the left MTG. However, we did not find evidence in support of either semantic or syntactic specialization in the frontal lobe. Our results partially converge with Friederici’s (2012) language model and the previous literature in monolingual children, suggesting that specialization in the frontal lobes develops later (Brauer & Friederici, 2007; Skeide et al., 2014; Wu, et al., 2016; Wang et al., 2020; Wang et al., 2021).

We observed semantic specialization in the posterior MTG in bilingual children ages 7–11 years old. This result complements and extends the prior findings based on monolingual children (Skeide et al., 2014; Wang et al., 2021). Both Skeide et al. (2014) and Wang et al., (2020, 2021) found that children ages 6–8 years old demonstrated distinct effects for semantics in the temporal lobe. However, semantic specialization was no longer observed in the temporal lobe in older age groups of 9–10 in Skeide et al. (2014) or in Wang et al. (in prep.). Results from the current study, which controlled for age, indicate that bilingual children show specialization for semantics in the MTG until about age eleven. We also observed evidence for semantic specificity in the angular gyrus even though we did not explicitly predict the difference in this region. This finding is not surprising given the AG’s prominent role in semantic processing (Binder et al., 2009), although its specific function in semantics is a matter of debate in the literature (Hogdson et al., 2021). Together, our results support the idea that the MTG is a central semantic processing hub implicated in both word- and sentence-level processes (Skeide & Friederici, 2016) and that the AG may additionally contribute to semantic processing during comprehension in young bilinguals (Li et al., 2014).

We also observed syntactic specialization in the STG. This result complements and extends the prior findings by Skeide et al. (2014) showing distinct effects for syntax in the temporal lobe in children ages 6–7, but not in children 9–10 years old. Across the three sets of studies, Wang and colleagues did not find evidence for syntactic specialization in the temporal lobe in children from age 5- to 10-years old. Results from the current study, which controlled for age, indicate that bilingual children show specialization for syntax in the STG until about age eleven. Friederici’s (2012) model suggests that the posterior STG is central to the integration of syntactic and semantic information. Support for this is evident in syntactic and semantic interaction effects at ages 3–4 and 6–7 (Skeide et al., 2014). The temporal lobe’s sensitivity to both syntax and semantics is also demonstrated in the MVPA findings by Wang and colleagues. They found that children ages 5–6 showed sensitivity to both syntax and semantics in the STG (Wang et al., 2020) whereas children ages 9–10 showed sensitivity to both in the STG as well as the MTG (Wang et al., in prep.). Our results closely parallel this pattern of activation. That is, the observed specificity for syntactic processing in the STG with adjacent specificity for semantic processing in the nearby MTG channel, could indicate syntax-semantic integration in the temporal lobe. However, multichannel comparisons are required to test this hypothesis.

There was no evidence for semantic or syntactic specialization in the frontal lobe, contrary to our hypothesis and the prior literature on monolingual individuals. We expected that the IFG pars triangularis would show greater activation for semantics than syntax (Wu et al., 2016; Wang et al., 2021), whereas the IFG pars opercularis would show the opposite (Skeide et al. 2014; Wang et al., in prep.). The lack of specialization we observed in the frontal lobe may be due to a limitation of our analytical approach. The current study used a univariate, single dissociation, analyses that captures overall brain activation per channel (i.e., per-voxel). However, Wang and colleagues conducted both univariate and multivariate analyses and only found evidence of semantic specialization in the IFGtri and syntactic specialization in the IFGop using the multivariate approach (Wang et al., 2021; Wang et al., in prep.). Although our method parallels other prior studies on language specialization (Brauer & Friederici, 2007; Wu et al., 2016; Skeide et al., 2014), results from Wang et al. suggests that a more sensitive measure of pattern analysis may be necessary to examine language specialization during sentence processing. Future studies could employ different approaches, such as multichannel decoding for fNIRS (Emberson et al., 2017), which has the potential to capture substantially greater detail about the patterns of activations within and across tasks (Haynes et al., 2015).

It is possible that a different syntactic manipulation may yield different results. Previous literature examining syntactic specialization have used a variety of stimulus types including subject-relative versus object-relative sentences (Ding et al., 2021; Jasinska and Petitto, 2013), phrase structure errors (Brauer & Friederici, 2007), and finiteness violations (Wang et al., 2021). Additionally, prior fMRI studies with monolingual children indicate evidence for specialization even when contrasting grammatically correct and semantically plausible sentences with no explicit syntactic or semantic violation present (Wang, Rice, & Booth, 2019; Wang et al., 2021). This suggests that functional specialization may be observed using sentence stimuli with an absence of an obvious syntactic or semantic violation. We included sentences with a clear morphosyntactic or semantic manipulation, thus perhaps the null effects observed here may be due to other language or cognitive processes or the lack of sensitivity in our analytical approach to detect such differences.

5. Conclusions

Theories of bilingualism posit that children’s experiences with the two languages make a combined contribution to brain development for language function. Our findings advance these perspectives by showing comparable patterns of brain-behavior associations between children’s proficiency in each of their languages and patterns of neural activation measured for their dominant and primary language of academic instruction. Moreover, our findings showcase that dual-language experiences afford normative language and brain development processes as evidenced in patterns of neural specificity that are generally commensurate with those previously reported for monolingual children.

Supplementary Material

1

Highlights.

  • 65 Spanish-English speaking children ages 7–11 completed functional near infrared spectroscopy imaging

  • Skills in both languages were related to brain engagement for syntax and semantics

  • Left superior temporal gyrus showed syntactic specialization

  • Left middle temporal gyrus showed semantic specialization

  • Results consistent with models suggesting a positive cross-linguistic interaction

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

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

References

  1. Arredondo MM, Hu XS, Seifert E, Satterfield T, & Kovelman I (2019). Bilingual exposure enhances left IFG specialization for language in children. Bilingualism: Language and Cognition, 22(4), 783–801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Baron A, Wagley N, Hu X, & Kovelman I (2023). Neural Correlates of Morphosyntactic Processing in Spanish–English Bilingual Children: A Functional Near-Infrared Spectroscopy Study. Journal of Speech, Language, and Hearing Research. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bedore LM, & Leonard LB (2001). Grammatical morphology deficits in Spanish-speaking children with specific language impairment. Journal of Speech, Language, and Hearing Research, 44(4), 905–924. [DOI] [PubMed] [Google Scholar]
  4. Binder JR, Desai RH, Graves WW, & Conant LL (2009). Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cerebral Cortex, 19(12), 2767–2796. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brauer J, & Friederici AD (2007). Functional neural networks of semantic and syntactic processes in the developing brain. Journal of Cognitive Neuroscience, 19(10), 1609–1623. [DOI] [PubMed] [Google Scholar]
  6. Claussenius-Kalman H, Hernandez AE, & Li P (2021). Expertise, ecosystem, and emergentism: Dynamic developmental bilingualism. Brain and Language, 222, 105013. [DOI] [PubMed] [Google Scholar]
  7. Cooperson SJ, Bedore LM, & Peña ED (2013). The relationship of phonological skills to language skills in Spanish-English-speaking bilingual children. Clinical Linguistics & Phonetics, 27(5), 371–389. [DOI] [PubMed] [Google Scholar]
  8. Costa A, & Sebastián-Gallés N (2014). How does the bilingual experience sculpt the brain? Nature Reviews Neuroscience, 15(5), 336–345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dijkstra TON, Wahl A, Buytenhuijs F, Van Halem N, Al-Jibouri Z, De Korte M, & Rekké S (2019). Multilink: a computational model for bilingual word recognition and word translation. Bilingualism: Language and Cognition, 22(4), 657–679. [Google Scholar]
  10. Ding G, Mohr KA, Orellana CI, Hancock AS, Juth S, Wada R, & Gillam RB (2021). Use of Functional Near Infrared Spectroscopy to Assess Syntactic Processing by Monolingual and Bilingual Adults and Children. Frontiers in Human Neuroscience, 15, 621025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Enge A, Friederici AD, & Skeide MA (2020). A meta-analysis of fMRI studies of language comprehension in children. NeuroImage, 215, 116858. [DOI] [PubMed] [Google Scholar]
  12. Emberson LL, Zinszer BD, Raizada RD, & Aslin RN (2017). Decoding the infant mind: Multivariate pattern analysis (MVPA) using fNIRS. PloS One, 12(4), e0172500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fedorenko E, Blank IA, Siegelman M, & Mineroff Z (2020). Lack of selectivity for syntax relative to word meanings throughout the language network. Cognition, 203, 104348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Friederici AD (2012). The cortical language circuit: from auditory perception to sentence comprehension. Trends in Cognitive Sciences, 16(5), 262–268. [DOI] [PubMed] [Google Scholar]
  15. Gavarró A, Torrens V, & Wexler K (2010). Object clitic omission: Two language types. Language acquisition, 17(4), 192–219. [Google Scholar]
  16. Goucha T, & Friederici AD (2015). The language skeleton after dissecting meaning: A functional segregation within Broca’s area. NeuroImage, 114, 294–302. [DOI] [PubMed] [Google Scholar]
  17. Hagoort P, & Indefrey P (2014). The neurobiology of language beyond single words. Annual Review of Neuroscience, 37, 347–362. [DOI] [PubMed] [Google Scholar]
  18. Haynes JD (2015). A primer on pattern-based approaches to fMRI: principles, pitfalls, and perspectives. Neuron, 87(2), 257–270. [DOI] [PubMed] [Google Scholar]
  19. Hernandez AE (2009). Language switching in the bilingual brain: What’s next? Brain and Language, 109(2–3), 133–140. [DOI] [PubMed] [Google Scholar]
  20. Hodgson VJ, Ralph MAL, & Jackson RL (2021). Multiple dimensions underlying the functional organization of the language network. NeuroImage, 241, 118444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hu XS, Hong KS, Ge SS, & Jeong MY (2010). Kalman estimator-and general linear model-based on-line brain activation mapping by near-infrared spectroscopy. Biomedical Engineering Online, 9(1), 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hu XS, Wagley N, Rioboo AT, DaSilva AF, & Kovelman I (2020). Photogrammetry-based stereoscopic optode registration method for functional near-infrared spectroscopy. Journal of Biomedical Optics, 25(9), 095001–095001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Ip KI, Hsu LSJ, Arredondo MM, Tardif T, & Kovelman I (2017). Brain bases of morphological processing in Chinese‐English bilingual children. Developmental Science, 20(5), e12449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jasinska KK, & Petitto LA (2013). How age of bilingual exposure can change the neural systems for language in the developing brain: A functional near infrared spectroscopy investigation of syntactic processing in monolingual and bilingual children. Developmental Cognitive Neuroscience, 6, 87–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Johnson MH (2011). Interactive specialization: a domain-general framework for human functional brain development? Developmental Cognitive Neuroscience, 1(1), 7–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kovelman I, Baker SA, & Petitto LA (2008). Bilingual and monolingual brains compared: a functional magnetic resonance imaging investigation of syntactic processing and a possible “neural signature” of bilingualism. Journal of cognitive neuroscience, 20(1), 153–169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kuo LJ, Ramirez G, de Marin S, Kim TJ, & Unal-Gezer M (2017). Bilingualism and morphological awareness: a study with children from general education and Spanish-English dual language programs. Educational Psychology, 37(2), 94–111. [Google Scholar]
  28. Lau EF, Phillips C, & Poeppel D (2008). A cortical network for semantics:(de) constructing the N400. Nature Reviews Neuroscience, 9(12), 920–933. [DOI] [PubMed] [Google Scholar]
  29. Li P, Legault J, & Litcofsky KA (2014). Neuroplasticity as a function of second language learning: Anatomical changes in the human brain. Cortex, 58, 301–324. [DOI] [PubMed] [Google Scholar]
  30. Matchin W, & Hickok G (2020). The cortical organization of syntax. Cerebral Cortex, 30(3), 1481–1498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Nuñez SC, Dapretto M, Katzir T, Starr A, Bramen J, Kan E, … & Sowell ER (2011). fMRI of syntactic processing in typically developing children: structural correlates in the inferior frontal gyrus. Developmental Cognitive Neuroscience, 1(3), 313–323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Peña ED, Bedore LM, & Kester ES (2015). Discriminant accuracy of a semantics measure with Latino English-speaking, Spanish-speaking, and English–Spanish bilingual children. Journal of Communication Disorders, 53, 30–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Peña ED, Gutérrez-Clellen VF, Iglesias A, Goldstein BA, & Bedore LM (2018). Bilingual English Spanish Assessment (BESA). Brookes. [Google Scholar]
  34. Peña ED, Bedore LM, Gutiérrez-Clellen VF, Iglesias A, & Goldstein BA (2016). Bilingual English-Spanish Assessment-Middle Extension. Field Test Version (BESA-ME). Unpublished manuscript. [Google Scholar]
  35. Petitto LA, Berens MS, Kovelman I, Dubins MH, Jasinska K, & Shalinsky M (2012). The “Perceptual Wedge Hypothesis” as the basis for bilingual babies’ phonetic processing advantage: New insights from fNIRS brain imaging. Brain and Language, 121(2), 130–143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Pylkkänen L (2019). The neural basis of combinatory syntax and semantics. Science, 366(6461), 62–66. [DOI] [PubMed] [Google Scholar]
  37. Ramirez G, Chen X, & Pasquarella A (2013). Cross-linguistic transfer of morphological awareness in Spanish-speaking English language learners: The facilitating effect of cognate knowledge. Topics in Language Disorders, 33(1), 73–92. [Google Scholar]
  38. Román P, González J, Ventura-Campos N, Rodríguez-Pujadas A, Sanjuán A, & Ávila C (2015). Neural differences between monolinguals and early bilinguals in their native language during comprehension. Brain and Language, 150, 80–89. [DOI] [PubMed] [Google Scholar]
  39. Rothman J, Bayram F, DeLuca V, Di Pisa G, Dunabeitia JA, Gharibi K, … & Wulff S (2023). Monolingual comparative normativity in bilingualism research is out of “control”: Arguments and alternatives. Applied Psycholinguistics, 44(3), 316–329. [Google Scholar]
  40. Santosa H, Zhai X, Fishburn F, & Huppert T (2018). The NIRS brain AnalyzIR toolbox. Algorithms, 11(5), 73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Schneider JM, & Maguire MJ (2019). Developmental differences in the neural correlates supporting semantics and syntax during sentence processing. Developmental science, 22(4), e12782. [DOI] [PubMed] [Google Scholar]
  42. Skeide MA, Brauer J, & Friederici AD (2014). Syntax gradually segregates from semantics in the developing brain. NeuroImage, 100, 106–111. [DOI] [PubMed] [Google Scholar]
  43. Skeide MA, & Friederici AD (2016). The ontogeny of the cortical language network. Nature Reviews Neuroscience, 17(5), 323–332. [DOI] [PubMed] [Google Scholar]
  44. Sulpizio S, Del Maschio N, Fedeli D, & Abutalebi J (2020). Bilingual language processing: A meta-analysis of functional neuroimaging studies. Neuroscience & Biobehavioral Reviews, 108, 834–853. [DOI] [PubMed] [Google Scholar]
  45. Sun X, Zhang K, Marks RA, Nickerson N, Eggleston RL, Yu CL, … & Kovelman I (2022). What’s in a word? Cross‐linguistic influences on Spanish–English and Chinese–English bilingual children’s word reading development. Child Development, 93(1), 84–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Tsakali V, & Wexler K (2004). Why children omit clitics in some languages but not in others: new evidence from Greek. LOT Occasional Series, 3, 493–504. [Google Scholar]
  47. Thompson-Schill SL, D’Esposito M, Aguirre GK, & Farah MJ (1997). Role of left inferior prefrontal cortex in retrieval of semantic knowledge: a reevaluation. Proceedings of the National Academy of Sciences, 94(26), 14792–14797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wagley N, Marks RA, Bedore LM, & Kovelman I (2022). Contributions of bilingual home environment and language proficiency on children’s Spanish–English reading outcomes. Child Development, 93(4), 881–899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Walenski M, Europa E, Caplan D, & Thompson CK (2019). Neural networks for sentence comprehension and production: An ALE‐based meta‐analysis of neuroimaging studies. Human Brain Mapping, 40(8), 2275–2304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wang J, Rice ML, & Booth JR (2020). Syntactic and semantic specialization and integration in 5-to 6-year-old children during auditory sentence processing. Journal of Cognitive Neuroscience, 32(1), 36–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wang J, Wagley N, Rice ML, & Booth JR (2021). Semantic and syntactic specialization during auditory sentence processing in 7–8-year-old children. Cortex, 145, 169–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Werker JF, & Hensch TK (2015). Critical periods in speech perception: new directions. Annual Review of Psychology, 66, 173–196. [DOI] [PubMed] [Google Scholar]
  53. Zaccarella E, Schell M, & Friederici AD (2017). Reviewing the functional basis of the syntactic Merge mechanism for language: A coordinate-based activation likelihood estimation meta-analysis. Neuroscience & Biobehavioral Reviews, 80, 646–656. [DOI] [PubMed] [Google Scholar]
  54. Zhao TC, & Kuhl PK (2022). Development of infants’ neural speech processing and its relation to later language skills: A MEG study. NeuroImage, 256, 119242. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1963626-supplement-1.pdf (229.6KB, pdf)

RESOURCES