Slide Slam E7
The Effect of L2 Proficiency in Grammatical Processing: An fMRI Study
Haining Cui1, Hyeonjeong Jeong1, Yuichi Suzuki2, Kiyo Okamoto1, Ryuta Kawashima1, Motoaki Sugiura1; 1Tohoku University, Japan, 2Kanagawa University, Japan
The extent to which grammatical representation and processing in second language (L2) rely on the same brain networks as first language (L1) is a fundamental issue. Previous neuroimaging studies showed that L1 grammatical processing recruits several key brain regions, such as the left inferior frontal gyrus (LIFG) and posterior superior temporal gyrus, which have been reported as critical areas for detecting grammatical errors. In contrast, grammatical processing by adult L2 learners is less efficient and automatic, recruiting the additional regions related to cognitive control (e.g., anterior cingulate cortex) and semantic analysis (e.g., angular gyrus) for compensating processing difficulties (e.g., Roncaglia-denissen & Kotz, 2016). While L2 proficiency presumably contributes to the development of L2 neural systems, it remains unclear whether increased L2 proficiency would result in L2 grammatical processing engaged in a native-like neural mechanism. To further clarify the role of L2 proficiency in grammatical processing, we conducted an fMRI experiment by recruiting L2 learners and L1 speakers to perform an Auditory Grammatical Judgment Task (AGJT) with L2 proficiency level as a factor of interest. Twenty-one healthy, right-handed L2 learners of Japanese with L1 Chinese background (mean age: 24.05±1.72; 15 females) and 21 native Japanese speakers (mean age: 21.57±1.62, 7 females) were asked to perform the AGJT. The AGJT contains 64 Grammatical [G] and Ungrammatical [U] sentences with syntactic structures that do not exist in L2 participants’ L1 Chinese. Participants were instructed to judge the grammaticality of each sentence within 10 seconds during an fMRI scanning. L2 learners’ general proficiency level was measured by using Japanese C-tests (a passage-based fill-in-the-gap test). For the brain data analysis, first, to investigate the group differences in processing grammatical error, we analyzed the brain activation in the contrast [U > G] for both L1 and L2 groups, respectively. We then compared the brain activation between L1 and L2 groups for the contrast [U > G] with two-sample T-tests implemented in SPM12 (corrected to P < 0.05 at cluster level). Second, to investigate the effect of L2 proficiency, correlation analysis using the proficiency scores of each L2 learner was performed at the whole-brain level for the [U > G] contrast. The results for group differences showed that the L1 speakers engaged higher activation of the LIFG than L2 learners in response to grammatical errors [U > G]. In contrast, the L2 learners recruited greater activation in the anterior cingulate cortex and the left angular gyrus than the L1 speakers. Notably, a significant positive correlation between activation in L2 [U > G] and proficiency level was found in the LIFG. Taken together, these findings suggest that L2 grammatical processing heavily demands cognitive control and sentence meaning analysis. However, with increased L2 proficiency, L2 learners can efficiently process grammatical information engaged in the same brain mechanism as native speakers.