Neural mechanisms of structural inference: an EEG investigation of linguistic phrase structure categorization

Language comprehension is a complex cognitive process of understanding spoken or written language, during which the brain derives meanings and interprets messages by constructing hierarchical structures from word sequences. A key component underlying this process is structural inference. This computation determines the syntactic category ("head") of phrases, such as noun phrases (NP) and verb phrases (VP). Categorization confers distinct semantic and distributional properties. However, isolating neural correlates of structural inference is challenging, as modulations of syntactic category usually alter lexical content. Here, we collected the electroencephalography (EEG) data while participants read Mandarin NPs and VPs, which differ in their syntactic "headedness" but with lexical-semantic, lexical-syntactic categories and word order strictly conserved. We found significant theta power increases at left central-parietal scalp sites in NPs relative to VPs around the presentation (0-210 ms) of the nouns, where only the nouns in NPs project a head. Both NPs and VPs relative to one-word phrases exhibited spatiotemporally consistent low-frequency power increases at the point of linguistic composition, consistent with previous studies. Overall, our findings offer novel constraints on neurocomputational accounts of linguistic structure-building through these previously undocumented signatures of syntactic inference.