Neural signature of bilingualism- A magnetoencephalographic study on content and language integrated learning

Resumo

Beneficial effects of bilingualism on cognitive development have been well proved using behavioral and neuroimaging methods. Classifying bilinguals into different categories is a challenging task but a fruitful avenue of research. Our participants were classified based on the amount of exposure to second language (L2) during schooling hours per week. By integrating our theoretical, behavioral and brain imaging results we tried to hold a great potential for the clarification of the true nature of European version of bilingualism, which is known as Content and Language Integrated Learning (CLIL). Although there are numerous success records of CLIL implementation in educational settings and its principles have been found effective, the impact of the whole educational methodology in terms of behavior and neural correlates remains to be addressed. In the first step, we conducted a comprehensive search and a meta-analysis on 102 published papers on CLIL to revisit the effectiveness of the method from theoretical and statistical point of view. The Chi-squared significance test shows that the distribution of effect sizes has heterogeneity (Q = 5, P < .001). Likewise, I-squared statistic quantifies the heterogeneity on the data. In general, results from this analysis suggest that no strong causal claims about the efficacy of CLIL may be derived from these studies due to the publication bias, insignificant weighted effect size, and small sample size. In this dissertation we assessed the impact of exposure to second language by conducting a set of empirical studies (behavioral and magnetoencephalography (MEG) studies). First, we investigated the unique contribution of language proficiency and the amount of exposure to second language in response inhibition and Stroop interference in congruent trials and incongruent trials. We enrolled 45 students (age range of 11.4 to 12) from different monolingual, semi-bilingual (CLIL) and bilingual schools. The results of word reading, color naming and color-word trials showed that the bilinguals outperformed their semi-bilingual and monolingual peers and semi-bilingual had a better performance than their monolingual peers (p < .00001), likely due to their cognitive abilities augmented by the control on two different languages. Two additional studies, using magnetoencephalography (MEG) were conducted – a picture naming and a semantic task. In the first MEG study we aimed to gain insight into the dynamics of word production in picture naming tasks in three groups of monolinguals, semibilingual and bilingual learners (N = 5 per group). We applied a picture naming task under magnetoencephalography (MEG) to provide more insights into the language production processes and to track temporal information with a precision of millisecond range, allowing investigation of the temporal flow and scalp distributions of well-established behavioral effects. Our results showed that participants from the CLIL group were found to be qualitatively different from their proficient counterparts in the different stages of picture naming in terms of brain mechanism associated with word production. We observed larger prominent activity area with increasing language proficiency. The activations were less scattered and bilateral for monolinguals (superior parietal) than CLIL and bilinguals, which had similarly scattered and left-hemisphere dominant activations (supramarginal, superior and inferior parietal). CLIL participants showed left frontal activations in Broca’s area (including BA44 L and BA45 L) that were absent in the other groups. Due to the level of proficiency, CLIL and proficient bilinguals showed a different evoked responses in the Broca’s area (N300 vs P600 accordingly). In the second MEG experiment we have examined the brain response to linguistic anomalies in the same group of participants (monolingual, semi-bilingual and bilingual learners; N = 5 per group) when the participants made acceptability judgments to correct semantic errors in syntactic sentences. The time-frequency analysis of the MEG data and multi-resolution signal decomposition technique allowed detecting changes in brain activity induced by lexical semantic processing and the functional interpretation of the N400 through congruent versus incongruent sentence endings which elicit N400. We observed that both the N400 onset latency and its amplitude vary between the three groups and did not correlate with proficiency. Our results also suggested that proficiency could modulate the event-related spectral perturbation effects attributed to greater linguistic and cognitive control gifted by schooling systems and the amount of exposure to L2. The results from this dissertation suggest that the brain response to bilingualism vary according to L2 exposure. CLIL elicits a different brain network showing a different pattern of electrical activity at the scalp especially N400 in terms of frequency and latency and variation in evoked brain activation patterns during language processing task. Moreover, the CLIL learners displayed lower executive control in comparison with their proficient bilingual counterparts. Our observations call into mind that CLIL exposure is associated with a different brain and behavior response than monolingual and bilingual education, namely recruiting a different brain response in different stages of word retrieval and word processing. Moreover, our results indicate that higher proficiency in L2 equips the learner with higher levels of interference suppression. In conclusion, the schooling system and the degree of bilingualism, which is decided by the amount of exposure to L2 can modify both function and neural structure of the brain.