Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Language Entropy and White Matter Connectivity: dimensional approach to bilingual experience
Poster Session E, Sunday, September 14, 11:00 am - 12:30 pm, Field House
Tatiana Davydova1, Valeria Mosqueda Cáceres1, Jaume Usó Cubertorer1, Eric Rivero Zaragoza1, Jesús Adrián Ventura2, Víctor Costumero1; 1Jaume I University, 2Universidad de Zaragoza
Introduction: Bilingualism has been linked to cognitive effects that extend beyond the language domain (Bialystok et al., 2012), with evidence also pointing to structural brain changes associated with the use of more than one language (Tao et al., 2021). For instance, Pliatsikas et al. (2014) reported effects in white matter (WM) integrity among late bilinguals compared to monolinguals, with higher fractional anisotropy values in the corpus callosum, in the inferior fronto-occipital fasciculus (IFOF), superior longitudinal fasciculus and uncinate fasciculus. However, most previous research has approached these neurocognitive adaptations within a categorical framework (bilingual vs monolingual), potentially oversimplifying the complex bilingual experience. Adopting a dimensional perspective may provide an enhanced comprehension of its impact on the brain. Therefore, we aim to explore the relationship among WM and language entropy, a measure based on Information Theory that quantifies the uncertainty and diversity in bilinguals´ use language across different communicative contexts (Gullifer & Titone, 2021). Method: Data from 164 healthy right-handed participants (mean age = 22.16 ± 2.69) who had completed a language use questionnaire and underwent an MRI diffusion Tensor Imaging (DTI) were analysed. The number of streamlines was extracted following the automatic tractography procedure outlined by Radwan et al. (2019). As measures of multilingual use, we calculated an average of self-rated proficiency across all languages known and a global language entropy for each participant as the average of entropy estimates from family and social contexts items regarding most recent years. Results: Bivariate correlations were conducted to examine the relationship between language entropy and potential covariates. A significant positive correlation was found between education level and language entropy (Spearman’s r = 0.185, p < 0.05). Generalized linear models were then constructed, controlling for sex, gender, and fluid intelligence (measured with matrix reasoning test), with a focus on brain regions previously identified in Pliatsikas’s study. Among these regions, only the left IFOF showed a significant improvement in model fit when measures of multilingual use were included. Furthermore, the inclusion of language entropy alongside multilingual proficiency further improved model fit, suggesting that greater structural connectivity in the left IFOF, as indicated by the number of streamlines, may be associated with both higher multilingual proficiency (z = 2.024; p = .044) and increased diversity in multilingual context (z = 2.089; p = 0.038). Given the IFOF's role in semantic processing (Nogueira et al., 2025), these results may reflect a neural adaptation to optimize semantic processing, which could be linked to the need to enhance cognitive flexibility and semantic control in response to linguistic unpredictability in multilingual context. *Valeria Mosqueda Cáceres and Tatiana Davydova contributed equally to this work.
Topic Areas: Multilingualism,