Slide Slam E6
How bilinguals dynamically comprehend words in the text: Evidence from functional connectivity and network science
Chanyuan Gu1, Ping Li1; 1The Hong Kong Polytechnic University
Introduction: Numerous studies have investigated neurocognitive differences between first (L1) and second (L2) languages in reading, but most of this literature has focused on single-word reading. Reading of whole sentences and paragraphs involves more component processes such as inference and integration than single-word reading. Further, naturalist reading involves backward saccades that are not possible within a RSVP paradigm in which single words are presented one by one. These differences point to potential different neural mechanisms involved in text vs. single-word reading. In this study, we use functional connectivity (FC) and network science-based graph metrics to study large-scale networks supporting L1 and L2 naturalistic reading of text. Methods: Fifty-two native English speakers (L1, 18-40 years old) and 56 Chinses-English (L2, 19-38 years old) speakers were recruited to read expository texts in English saliently. Fixation-related fMRI method was employed to record eye-movements and blood-oxygen-level-dependent (BOLD) signals simultaneously. First-pass reading time measures the sum of all fixations counting from first-time eye landings on targets until leaving. Second-pass reading time measures the sum of all fixations on targets after the first-pass reading time. Both were employed as events to convolve BOLD signals for words that were re-read. FC during reading was measured with large-scale networks, and network-based statistical analysis was adopted. In addition, topologies of subnetworks were further measured with graph metrics. Results: We found significantly different patterns for first-pass and second-pass reading. (1) During the first-pass reading time, one subnetwork was identified for the contrast “L1 > L2”, mainly encompassing connections between occipital and prefrontal areas, between occipital and temporal areas, between prefrontal and parietal areas, and within the occipital cortex. Six hubs (e.g., the left precentral gyrus and the right middle frontal gyrus) were identified. Another subnetwork was detected for the contrast “L2 > L1”, mostly involving connections between prefrontal and other cortical areas, between occipital and parietal areas, and within the prefrontal cortex. Six hubs (e.g., the right dorsal-lateral superior frontal gyrus (DLSFG) and the left middle temporal gyrus) were identified; (2) During the second-pass reading time, one subnetwork was found for the contrast “L1 > L2”, mainly consisting of connections between parietal and other areas and within the occipital cortex. Seven hubs (e.g., the bilateral posterior cingulate gyrus (PCG)) were identified. Another subnetwork was reported for the contrast “L2 > L1”, primarily comprising connections between prefrontal and temporal areas, parietal and occipital areas, and within the prefrontal cortex. Six hubs (e.g., the right DLSFG and superior temporal gyrus) were identified. Conclusion: Our study suggested great engagement of a distributed network for L1 during the first-pass reading time, while L2 relied more on the prefrontal system. During the second-pass reading time, there was more engagement of the parietal system for L1, while L2 mainly utilized the prefrontal-temporal system. These data differ from previous neuroimaging studies of single-word reading, showing that distinct neural cooperative systems were dynamically recruited in L1 and L2 reading of text, systems that are necessary for additional cognitive resources employed by L2 readers in text comprehension.