Neural Specificity Of Advanced Skilled Chinese - English Learners Grammatical Features Of Processing

In the cognitive neurolinguistic field, it’s a basic question whether the first language (L1) and the second language (L2) engage only a single brain mechanism or two distinct neural networks in language processing. Research findings indicate that early bilingual learners process semantics by neural networks underlying the native language; late bilingual learners are negatively affected by declined neural plasticity and display non-overlapped brain activation in processing grammatical information. Even though it is found that increased language proficiency leads to narrowed activation disparity of semantic tasks so that neural substrates of two semantics in the late bilingual brain can overlap, few researches have reported that neural activation disparity of grammatical features can be modulated by intensive practices. Until today, advocates of decisiveness of language proficiency fail to explain the overlapping of cortical regions underlying two semantics, but rest on the description of the overlapped findings. On the other hand, it was still unknown why increased proficiency seldom bridged the gap between brain activation of the two languages’grammatical features. Recent bilingual researches focus on language features. They find that cortical regions underlying grammatical information in the early bilingual brain may be independent and extents of activation differences are modulated by language differences between languages.On condition of above findings, we proposed a dual mechanism of brain networks underlying idiosyncratic grammatical information, suggesting that the development of neural pathways is modulated by language features, and neural development efficiency is affected by language acquisition time and language proficiency. Late bilingual learners can develop a native-like neural substrates in processing grammatical features, which is independent from their first language brain networks. We then conducted three experiments to demonstrate the feasibility of such a dual mechanism model.Due to linguistic features of the analytic language, Chinese-English bilinguals are least influenced by their native language in learning English inflections. Such a contrast makes it possible to test if late Chinese English bilinguals would have to develop specialized neural networks for distinctive language features, which are similar to those of English native speakers. Through an event-related fMRI study by the repetition priming paradigm, we found that proficient late bilingual learners processed English inflected verbs the same way as English monolingual learners both behaviorally and neurophysiologically. As cortical regions underlying Chinese grammatical information overlap with that of semantic information and the phonological information is processed by the medial frontal gyrus and the inferior parietal lobe, it denies the possibility that processing English inflections rely on cortical areas of the native language. It has successfully demonstrated the decisive role of language features in building neural basis for new linguistic functions and the more vital importance of language proficiency than the language acquisition time in forming a native-like brain network. However, serial experiments I did not locate activation disparities by direct contrast of the decomposable words and the regularly inflected verbs. The results cannot reject the connectionist hypothesis proposing that activation differences between regular and irregular past tense are caused by phonological consistency of a regular past tense form and its stem.Serial experiments Ⅱ made use of updated data analysis in functional brain imaging field, and retested conclusions of serial experiment I. It aimed to investigate working mechanisms in processing regular and irregular past tense through effective connectivity analysis, hoping to reject the connectionist hypothesis that attributes activation disparity of regular and irregular verbs to different degrees of phonological relevance. We found different effectively connected networks in processing the regular past tense, irregular past tense, nonce verb, and decomposable word. In detail, the inferior frontal gyrus and basal ganglia areas (Brodmann44/45, putamen, and caudate) were hubs of the regular past tense network, with Brodmann39in the post area of the superior temporal gyrus as another; irregular past tense had three hubs subserving semantic access (Brodmann45, Brodmann37, and caudate); nonce verbs showed a only hub in Brodmann21; hubs of decomposable words were the temporal pole, the pars orbitalis, the pars opercularis and caudate (Brodmann38,,Brodmann47, Brodmann44, and caudate) taking charge of decomposition. Results of serial experiments II lent support to serial experiments I. It clarifies the processing of regular past tense, which contains verb decomposition, grammatical information recognition, phonological parsing, and semantic access. We also found that the response time had significantly negative correlations with clustering density and the in-out degrees of important nodes. It indicates that the effectively connected networks we found are trustful, and proposed the advanced role of functionally connected brain networks in settling hot debates in the brain imaging area of languages.To specify the independence of working mechanisms underlying language features, serial experiment III modulated degrees of English and Chinese simple present SAVO sentences. We found that in both affirmative and negative levels, even though English first person singular simple present sentences had the same structures with Chinese first person singular simple present sentences, English first person singular simple present sentences (1sg) and English third person singular simple present sentences (3sg) consistently displayed similar activation contrasts with that of Chinese. It demonstrates that working mechanisms of grammatical features of English and Chinese are independent; language information is processed consistently by the same neural substrates and don’t show mixed processing model even though parts of the information may be structurally the same across languages. Chinese simple present sentences are characterized by expressing ideas through word orders and activated the pars opercularis in the IFG. English simple present sentences need subject-verb agreement check, which involved the activation of phonological parsing area (Brodmann39) and the decomposition area (Brodmann38/47). Through mutual contrasts of affirmative and negative sentences, serial experiment Ⅲ also demonstrates different mechanisms in dealing with the two basic sentences:affirmative sentences rely on the information integration center of the pars opercularis, and negative sentences rely on the information integration center of area Spt in the parietal-temporal conjunction area. This finding can answer the debate between "truth reversion hypothesis" and the "reduced acceptability of false proposal hypothesis".Proceeding from the empirical evidence that processing grammatical features of the second language relies on native-like specialized brain networks, the current research proposed a dynamic model illustrating how the language neural network changes during second language acquisition. It insists a dominant role of linguistic features as the source of non-overlapping activations between L1and L2. Language proficiency plays a crucial role in practicing cerebral integration and segregation processes, during which the brain assimilates common information and accommodate foreign inputs. Given that reduced neural plasticity affects both the acquisition of syntax and semantics, it is the integrative expansion of the assimilation process and the segregated specificity of the accommodation process that cause the maturity asynchrony between syntax and semantics. Under the theoretical framework, we discussed main findings in the bilingual cognitive neurolinguistic research field. At last, we summarized main findings and conclusions of the current research, and suggested that research subjects can be extended to early Chinese English bilingual learners and lowly proficient bilingual learners to better prove the feasibility of the dual mechanism view on grammatical information.