Detection of semantic inconsistencies of motor actions: From language to praxis

Abstract

Semantics of actions includes three types of knowledge: a) of the function of objects and tools, b) of the actions independently of the tools, and c) of the organization of simple actions in sequences. These types of knowledge might be structured as thematic roles into the semantics of actions. To test these hypothesis, 125 illustrations, divided into five conditions: i) Congruent (C), ii) Agent Inconsistent (AI), iii) Instrument Inconsistent (II), iv) Patient Inconsistent (PI), and v) Location Inconsistent (LI), were presented to 23 volunteers (50 % women), aged 20–25, who were asked to respond whether the image was congruent or incongruent. Electrical brain activity was recorded through 20 channels to obtain the event-related potentials (ERP) associated. Lower reaction times for II and PI than C, and a greater number of incorrect trials for C were found. A N300/N400 effect appeared for AI and LI conditions with respect to C. Finally, II and LI conditions present a deflected P600 in reference to C. These findings suggest semantic of actions is sensitive to thematic role manipulations and constitute evidence in favor of a semantic processing shared between visually observed praxis and words.