Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Exploring neural mechanisms of bilingual language control: an fMRI study using functional localizer approach
Poster Session B, Friday, October 25, 10:00 - 11:30 am, Great Hall 3 and 4
Keng-Yu Lin1, Agata Wolna2, Jakub Szewczyk1, Kalinka Timmer3, Michele Diaz4, Zofia Wodniecka1; 1Jagiellonian University, 2Massachusetts Institute of Technology, 3University of Warsaw, 4Pennsylvania State University
Bilinguals usually need to navigate between their first language (L1) and second language (L2). When they need to use one of the languages, speaking in L2 is usually harder than speaking in L1, as reflected by longer reaction times and increased neural activity in L2 (vs. L1). This difficulty is argued to reflect the greater engagement of the bilingual language control (BLC) mechanisms. The demands become even more challenging when bilinguals have to switch between L1 and L2. One of the most commonly used paradigms to test BLC is the language-switching task. On the behavioral level, it typically demonstrates that switching to L1 is related to bigger costs than switching to L2 (a.k.a. language switch costs asymmetry) and that the requirement to switch between languages results in L1 becoming overall slower than L2 (a.k.a. reverse language dominance). However, it remains unclear (i) whether the asymmetric switch costs and the reverse language dominance reflect the same language control mechanism, and (ii) whether the behavioral effects in language switching reflect the increased difficulty for language-specific processing, or the engagement of domain-general control mechanisms. We conducted an fMRI study to identify brain networks involved in BLC during language switching. To explore whether the control draws on domain-general or language-specific mechanisms, we examined brain responses within two well-established functional networks identified at the individual-subject level: the domain-general Multiple Demand (MD) Network (Duncan et al., 2010) and the Language Network (Fedorenko et al., 2010). Forty-five Polish-English bilinguals completed a cued language-switching task where they named pictures in L1Polish and L2English, as well as two functional localizer tasks: a language localizer task, which contrasts listening to intact vs. degraded speech (Malik-Moraleda et al., 2022), and a domain-general (MD) localizer, which contrasts hard vs. easy working memory tasks (Fedorenko et al., 2013). Using the localizers, at the level of individual subjects, we identified functional regions of interest (corresponding to 10% of the most active voxels for a localizer contrast), from which we drew the estimates of brain responses to the language switching task. Our results revealed a language switch cost asymmetry in both the language and the MD network, whereas the reverse language dominance effect was reflected only in the MD network. The dissociation between the reverse language dominance effect and the asymmetry of switch costs within these two networks suggests that these phenomena are supported by at least partially separate mechanisms. On the one hand, both effects are reflected in the MD network. This suggests a domain-general nature of the mechanisms supporting the bilingual language control which is in line with previous research postulating that BLC relies largely on the domain-general mechanisms. On the other hand, however, the asymmetry of switch costs was also reflected in the language network, suggesting that the effect is also underlain by language-specific processes. We speculate that while the reverse dominance reflects a general and large-scale, proactive control adjustment, the asymmetry of switch costs partially reflects short-term, local and reactive changes in the activation of the two languages.
Topic Areas: Multilingualism, Language Production