Poster C52, Friday, August 17, 10:30 am – 12:15 pm, Room 2000AB

Neural correlates of sematic ambiguity in second language learners

Kiyo Okamoto1, Hyeonjeong Jeong2,3, Haining Cui3, Ryuta Kawashima4, Motoaki Sugiura2;1Graduate School of Medicine, Tohoku University, Sendai, Japan, 2Department of Human Brain Science, IDAC, Tohoku University, Sendai, Japan, 3Graduate School of International Cultural Studies, Tohoku University, Sendai, Japan, 4Department of Advanced Brain Science, IDAC, Tohoku University, Sendai, Japan

Learning ambiguous words that have more than one meaning (i.e., polysemy) is challenging for most second language (L2) learners. It is true that bilinguals including L2 learners face difficulty in processing ambiguity in each of their languages as well as between languages (Degani & Tokowicz, 2010). Previous neuroimaging studies on first language (L1) have reported that semantic-related areas (i.e., the left inferior frontal gyrus) were activated during the reading of ambiguous words (e.g., Bitan et al., 2017). However, little is known about how L2 learners process ambiguous L2 words in the brain. We hypothesized (a) that L2 learners may need resolution of semantic processing with a similar way to process their L1, and (b) that L2 learners may process L2 ambiguous words in cognitive control areas such as the anterior cingulate cortex and caudate nucleus (Abutalebi & Green, 2016). Furthermore, learner’s L2 proficiency levels may influence cognitive mechanisms of resolving ambiguous process of L2 words. The current fMRI study attempted to investigate the brain mechanism involved in resolution of ambiguity across languages (L2: English - L1: Japanese) and the mediating role of L2 proficiency level on its brain mechanism by controlled within-language ambiguity (L1: Japanese). Participants were 18 healthy right-handed native speakers of Japanese who had learned English as L2 (mean age: 21.6, female 7). They performed cross-language (L2-L1, hereafter referred to as L2 task) and within-language (L1-L1, hereafter referred to as L1 task) semantic relatedness judgment tasks during MRI scanning. In both tasks, we presented either ambiguous words or unambiguous words with semantically related or unrelated words in a random order. For L2 (cross-language) task, we prepared 25 ambiguous and unambiguous English words for each task. The same numbers of Japanese ambiguous (i.e., homonyms) and unambiguous words were prepared for the L1 (within-language) task as a control. Four types of conditions in each task (L2 and L1) were created: related ambiguous words (RA), unrelated ambiguous words (UA), related unambiguous words (RN), and unrelated unambiguous words (UN). We focused on only RA and RN conditions in L2 and L1 for the data analysis. Participants’ standardized language proficiency score of L2 (i.e., TOEIC) were used to examine the proficiency effect. The correlation analysis was conducted in the contrast L2 [RA-RN] and the proficiency scores in the whole brain analysis. Statistical analyses were performed with SPM12 using a random effects model (voxel-level p<0.001 uncorrected and then corrected to p<0.05 by cluster-level). Two major findings are noteworthy. First, the main effect of ambiguity in the L2 and L1 tasks [RA_L2+RA_L1]-[RN_L2+RN_L1] revealed that both language similarly induced activation in the inferior frontal gyri. However, there was no significant interaction effect [RA_L2-RN_L2]-[RA_L1-RN_L1]. Second, the higher L2 proficiency was related to greater activation in the left caudate nucleus (small volume correction, FWE p<0.05). This result suggests that advanced L2 learners may recruit more language control areas to process ambiguous L2 words than lower-level L2 learners.

Topic Area: Language Development