| Language processing involves the collaborative efforts of multiple brain regions.Early studies have shown that the structure and function of the brains of bilinguals(those who are proficient in two or more languages)would be changed.Despite numerous studies on the effects of bilingualism on the brain,research on the impact of the age of second language acquisition on brain functional organization and dynamics remains scarce.With the development of technology and the deepening of brain research,functional magnetic resonance imaging(f MRI)has become a reliable tool for studying the brain.Based on network science,studies have found that functional connections between different brain regions can form a network.Complex network analysis methods based on graph theory have played a crucial role in quantitatively analyzing the topological structure,dynamic behavior,and robustness of networks,and can be used to predict network behavior under different conditions,as well as reveal key nodes or communities of networks.This method has been successfully used to explore the characteristics and behavioral problems of many practical systems.Against this background,this study used f MRI and complex network analysis methods to construct and compare the static and dynamic brain networks of early bilinguals(EB),late bilinguals(LB),and monolingual controls(MC)(In this study,EB refers to bilinguals who acquired a second language before the age of 10,LB refers to those who acquired a second language after the age of 14,and MC refers to those who can only proficiently use one language).The study aimed to investigate the impact of bilingual experience on the structure and dynamics of brain functional networks:1.In exploring functional network structures,this study evaluated the individual brain network topological structures of participants based on graph theory methods and explored the degree to which they were influenced by different periods of second language acquisition.Specifically,graph theory analysis was used to analyze the network topological properties and rich-and diverse-club organizations simultaneously and calculated the functional connections and topological properties of member nodes of these organizations.The results showed that the network efficiency and rich-club functional connections in the EB group were significantly higher than those in the MC group.Furthermore,some nodes’ functional connections in the rich-club organizations of the EB group were significantly higher than those of the other two groups.However,there were no significant differences between the LB and MC groups.The increase in the strength of these nodes’ functional connections was positively correlated with the number of languages learned by the participants.In addition,as one of the overlapping nodes of rich-club and diverse-club organizations,A24rv＿L in the left cingulate gyrus showed a higher betweenness centrality in EB compared to the other two groups.This suggests that the structure of brain functional networks may undergo positive changes during the process of second language learning,and these changes may occur primarily in early bilinguals.2.In exploring brain dynamics,this study used a sliding window method to construct dynamic multi-layer functional networks and investigated,for the first time,the relationship between bilingual experience and the large-scale brain network dynamics during rest.Specifically,this study first constructed dynamic multi-layer functional networks and explored differences in neural flexibility and temporal variability among the three groups at the node and network levels.Classification features were extracted from the neural flexibility and temporal variability using LASSO(Least Absolute Shrinkage and Selection Operator)regression.These features were then used to train classification models with support vector machines(SVM)for both monolingual and bilingual individuals.The results showed that significantly increased neural flexibility was observed in the EB group at A9l＿L of the left superior frontal gyrus,A7m＿R of the right precuneus,the limbic network,and the default network compared to the MC group.Additionally,the neural flexibility in A7m＿R was higher in the LB group compared to the MC group.The temporal variability of basal ganglia in the EB group was significantly higher than that in the LB and MC groups.Furthermore,the number of languages learned by participants showed a positive correlation with the neural flexibility in A9l＿L,the default network,and the limbic network,as well as the temporal variability in A7c＿R of the right superior parietal lobule.The model classification accuracy reached 72% using the neural flexibility of A9l＿L,A7m＿R,and the limbic network,all of which were extracted using LASSO regression as features for training in SVM.This means that second language acquisition may enhance brain flexibility,and more language experience may play a greater role in shaping brain flexibility.This may be because bilingual individuals frequently switch between languages,requiring stronger language processing,cognitive control,and attention,factors that make the brain more flexible in adapting to constantly changing cognitive demands. |
