Disentangling semantic composition and semantic association in the left temporal lobe

Abstract

Although composing two words into a complex representation is conceptually different from forming associations between a pair of words, the brain regions supporting semantic composition have also been implicated for associative encoding. Prior work has never directly compared the neural substrates of composition and association. Here, we adopted a two-word magnetoencephalography (MEG) paradigm which varies compositionality ("French/Korean cheese" vs. "France/Korea cheese") and strength of association ("France/French cheese" vs. "Korea/Korean cheese") between the words. We applied both univariate regression analyses and multivariate pattern classification to the source-localized MEG data in a bilateral language network. We show that the left anterior and middle temporal lobe are distinctly modulated by semantic composition and semantic association. Pattern-based directed connectivity analysis further revealed a continuous information flow from the anterior to the middle temporal region. These findings shed light into a functional dissociation within the left temporal lobe for compositional and distributional semantic processing.

Neural correlates of semantic number: A cross-linguistic investigation

“…It is interesting that we only identify the left ATL in the English results, but not the Chinese or French results, given its importance in the literature on semantic composition (e.g., Bemis & Pylkkänen, 2011Li et al, 2020). The anterior temporal lobes can be susceptible to signal loss during fMRI imaging (Devlin et al, 2000), however, and we perform full-brain analyses, which reduces statistical power.…”