Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks

From shared representations to individual variations in non-native speakers: shared activity patterns predict naturalistic reading performance

Poster B128 in Poster Session B, Tuesday, October 24, 3:30 - 5:15 pm CEST, Espace Vieux-Port

Chanyuan Gu1, Samuel. A Nastase2, Ping Li1; 1The Hong Kong Polytechnic University, 2Princeton University

Learning a second language is an important part of daily life for many people. As second language learning plays a significant role in education and work, it is essential to understand the neural underpinnings of second language reading. Despite similar neural mechanisms involved in first and second language reading at the level of single words or sentences, it remains unclear how non-native speakers successfully comprehend real-world texts in naturalistic settings. In this study, we propose to examine whether the neural activities in non-native speakers are aligned to each other and also aligned to native speakers during naturalistic science-text reading comprehension. We then compare the degree of alignment to other assessments of linguistic abilities. We recruited 56 Chinese-English speakers as non-native speakers and 52 native English speakers as native speakers to silently read scientific texts in English. Using a fixation-related fMRI paradigm, we simultaneously recorded participants’ eye movements and neural activity during reading. We hypothesized that shared patterns of neural activity among participants will be detected in brain regions involved in reading comprehension due to similar language background and efficiency. By leveraging the spatial intersubject correlation (pISC) framework, we measured the shared activity patterns among non-native speakers and between non-native and native speakers. Finally, we adopted intersubject representation similarity analysis and constructed multiple regression models to examine the effect of linguistic abilities on pISC. We found strong pISC in non-native speakers in brain regions including language and default-mode network (DMN) regions, multiple-demand (MD) regions, and visual regions. The alignment among non-native speakers in those brain regions was correlated with reading performance. Further, a subject-by-subject vocabulary size matrix, where non-native subjects with similar vocabulary size were more similar, was positively correlated with the alignment in DMN, MD, and visual regions, suggesting stronger alignment in non-native speakers with similar vocabulary size. In addition to the alignment among non-native speakers, we found significant alignment between non-native and native speakers in DMN, language, MD and visual regions, which was also positively correlated with reading performance. Finally, the regression models comprising vocabulary size, general reading ability, and their interaction term significantly predicted neural alignment in the right language and left DMN regions, indicating that linguistic abilities jointly predict the alignment of brain activity patterns in non-native speakers to those of native speakers. To conclude, by leveraging a naturalistic reading paradigm, our study provides novel fMRI evidence about the neural underpinnings of naturalistic scientific reading comprehension in non-native speakers. Specifically, non-native speakers engaged distributed lower- and higher-order brain regions during naturalistic reading comprehension, revealed by exhibiting shared activity among non-native speakers and between non-native and native speakers, both of which predict reading performance. Further, individual linguistic abilities significantly impacted shared spatial representations in language and DMN brain regions, respectively. Our findings thus provide the first systematic data in this respect to shed light on how shared activity patterns can capture naturalistic language comprehension and how individual variations impact shared representations in second language reading.

Topic Areas: Reading, Language Development/Acquisition

SNL Account Login

Forgot Password?
Create an Account