Neurophysiological correlates of executive function and theory of mind in autistic and neurotypical children : EEG and machine learning

Abstract

The Bilingual Interactive Activation Model Plus (BIA+; Dijkstra & van Heuven, 2002) and the Revised Hierarchical Model (RHM; Kroll & Stewart, 1994), differ in claims regarding bilingual lexical non-selectivity, the idea of automatic co-activation of two languages in a bilingual when input of only one language is provided. The BIA+ model argues that lexical and meaning entries of L1 and L2 are integrated in one lexicon to have automatic and parallel access when similar phonemes or pronunciations are presented. The RHM claims for separate L1 and L2 lexicons, and the input of one language would not automatically activate the lexical entries of another language. Both models assume that L1 and L2 relationship is asymmetric, but the RHM assumes greater L2→L1 priming effects due to strength differences between L1 and L2. Conversely, the BIA+ model assumes that greater exposure of L1 representations than L2 contributes to greater L1 influence. The current event-related potential (ERP) study investigated how lexical representations of two different-script languages are organized in the bilingual lexicon. A lexical decision task was used to test whether phonological and semantic cues from L1 non-selectively activate L2 representations and vice versa prior lexical access. Prime-target pairs that varied in phonological and semantic similarity in a factorial design were used to compare Korean-English bilinguals with English monolinguals, and to contrast directions of priming (L1→L2 vs. L2→L1) within bilinguals. Four types of phonological and semantic prime-target language relationships were Phonologically similar-Semantically similar (P+S+) [Team –팀 / team/ (Team)], Phonologically similar-Semantically dissimilar (P+S-) [Soup –숲 /soup/ (Forest)], Phonologically dissimilar-Semantically similar (P-S+) [Bear –곰/gom/ (Bear)] and Phonologically dissimilar-Semantically dissimilar (P-S-) [Pain –쌀 /ssal/ (Rice)]. Comparison of bilinguals and monolinguals found that phonologically similar prime-target pairs were more prone to errors and evoked a greater positivity at the early central-right P2 component, but reduced negativity at the lexical N400 than phonologically dissimilar pairs in bilinguals. Words with semantically similar primes also led to more accurate and faster response, with earlier peak latencies and reduced activation at the late central N400 in bilinguals. Thus, shared phonology and meaning between Korean and English facilitate the ease of lexical-semantic retrieval and post-lexical processing during L2 word recognition. As greater P2 positivity and N400 negativity represent greater efforts required for the associated processing, the interaction of L1 and L2 phonology is inhibitory (greater efforts) at the sub-lexical level and becomes facilitatory (reduced efforts) as competition is resolved at the lexical level. Similarly, comparison of priming directions within bilinguals also showed phonological inhibition and semantic facilitation in behavioural results, in addition to ERP findings that initial sub-lexical phonological interference at middle P2 was followed by lexical phonological facilitation at the mid centro-parietal N400, and semantic facilitation at the middle N400. Critically, sub-lexical phonological interference, together with lexical phonological facilitation for semantically dissimilar words, was confined to the L1→L2 direction. This indicated that L1 influence on L2 is more dominant and the asymmetric phonological priming effect occurs at both sub-lexical and lexical level. The automatic and non-selective activation of phonological and semantic representations from two scripts with distinct visual forms as well as the asymmetric influence of L1 and L2 at sub-lexical phonological processing collectively give support to the BIA+ model. However, sub-lexical phonological interference effects and how orthographically distinct scripts are processed are not clearly explained in the current BIA+ model. Hence, accommodation of additional links to the language membership nodes from sub-lexical representations and including the degree of overlap between orthography and phonology are proposed.

Keywords: bilingualism, lexical non-selectivity, semantic priming, phonological priming, N400, P2.