Nouns, verbs, objects, actions, and abstractions: Local fMRI activity indexes semantics, not lexical categories

Highlights

• •

  Concrete nouns and verbs elicit different brain signatures in frontocentral cortex.

• •

  Abstract nouns and verbs fail to elicit different brain activation patterns.

• •

  Concrete verbs activate motor and premotor cortex more strongly than concrete nouns.

• •

  Concrete nouns activate inferior frontal areas more strongly than concrete verbs.

• •

  Lexical category distinctions in middle temporal cortex cannot be confirmed.

Abstract

Noun/verb dissociations in the literature defy interpretation due to the confound between lexical category and semantic meaning; nouns and verbs typically describe concrete objects and actions. Abstract words, pertaining to neither, are a critical test case: dissociations along lexical-grammatical lines would support models purporting lexical category as the principle governing brain organisation, whilst semantic models predict dissociation between concrete words but not abstract items. During fMRI scanning, participants read orthogonalised word categories of nouns and verbs, with or without concrete, sensorimotor meaning. Analysis of inferior frontal/insula, precentral and central areas revealed an interaction between lexical class and semantic factors with clear category differences between concrete nouns and verbs but not abstract ones. Though the brain stores the combinatorial and lexical-grammatical properties of words, our data show that topographical differences in brain activation, especially in the motor system and inferior frontal cortex, are driven by semantics and not by lexical class.