Poster D63, Friday, August 17, 4:45 – 6:30 pm, Room 2000AB

The relationship between resting state networks and bilingual reading: A preliminary investigation

Jason Gullifer1,4, Xiaoqian Chai2,4, Veronica Whitford3,4, Irina Pivneva1,4, Shari Baum1,4, Denise Klein1,4, Debra Titone1,4;1McGill University, 2Johns Hopkins University, 3University of Texas at El Paso, 4Centre for Research on Brain, Language & Music

Bilinguals coactivate both languages when reading text in a single language, evidenced by words that share lexical overlap across languages. For example, eye-movement measures have shown that cognates, which share form and meaning across languages (e.g., piano in English and French), result in facilitation relative to language-unique words, whereas interlingual homographs, which share form but not meaning (e.g., chat is a casual conversation in English vs. a cat in French), result in interference. Yet, the magnitude of these cross-language effects varies across individuals as a function of language background and executive control abilities. Bilinguals with higher second-language proficiency experience less cognate facilitation, and bilinguals with higher executive control abilities experience less homograph interference, suggesting that independent mechanisms are responsible for language control depending on the type of coactivation (Pivneva, et al., 2014). Neurocognitive research employing resting state functional connectivity analysis has identified at least eight independent brain networks composed of areas that show synchronous temporal activity. At a broad level, these include the visual, language, sensorimotor, default mode, salience, frontoparietal control, dorsal attention, and cerebellar networks. Connectivity within these networks is observable even when the brain is at rest (Lee, et al, 2013). Importantly, networks involved in executive control (salience, frontoparietal, and dorsal attention) also play important roles in bilingualism (Kousaie et al., 2017) and are assumed to be highly integrated with the language network for bilinguals (Abutalebi & Green, 2016). To date, no studies have combined these two threads of research to investigate how cross-language activation during bilingual reading maps onto connectivity within resting state networks. Here, we report data from 27 French-English bilinguals who read English sentences containing target French-English cognate, interlingual homograph, and control words while their eye-movements were monitored. In another session, the participants underwent a resting state fMRI scan. In a preliminary analysis, we extracted coarse reading measures (total reading time) for target words and computed difference scores for cognate facilitation and homograph interference. With connectivity data, we conducted an independent component analysis and used spatial mapping to identify likely networks for each component (Calhoun et al., 2001). We then associated the coactivation effects with the connectivity within each network in a second-level analysis. Preliminary results showed that both types of coactivation were related to connectivity in the frontoparietal control network with similar patterns. Facilitation (irrespective of cognates or homographs) was associated with greater connectivity within this network and inhibition was associated with less. These preliminary results suggest a general entrainment between the magnitude of cross-language activation, irrespective of word type, and connectivity within the frontoparietal control network at an individual level. They also support neurocognitive models of bilingualism which predict that language and cognitive systems are highly integrated (Abutalebi & Green, 2016). Future analyses will involve eye-tracking measures that are more sensitive earlier stages of lexical processing (e.g., first fixation and gaze duration) vs. the late stage measure used here (total reading time), as previous studies have shown differences between the two stages in terms resolution of language coactivation.

Topic Area: Multilingualism