Influence of Object Affordance on Action Language Understanding: Evidence from Dynamic Causal Modeling Analysis

TL;DR

This study demonstrates that affordance-related motor regions causally influence semantic areas during action language comprehension, supporting grounded cognition theories through EEG and dynamic causal modeling analysis.

Contribution

It provides novel evidence of causal neural pathways from premotor to parietal and temporal regions in action language understanding.

Findings

Premotor cortex causally influences semantic regions during language comprehension.

EEG and DCM reveal a feedforward neural architecture involving motor and semantic areas.

Affordance processing in premotor regions drives downstream semantic activation.

Abstract

This study investigates the causal neural dynamics by which affordance representations influence action language comprehension. In this study, 18 participants observed stimuli displayed in two conditions during the experiment: text-only (e.g., `Hit with a hammer') and video+text (visual clips with matching phrases). EEG data were recorded from 32 channels and analyzed for event-related potentials and source localization using LORETA, which identified four left-hemisphere regions of interest: the Lateral Occipital Cortex (LOC), Posterior Superior Temporal Gyrus (pSTG), Ventral Premotor Cortex (PMv), and Inferior Parietal Lobule (IPL). A space of dynamic causal modeling (DCM) was constructed with driving inputs to LOC and pSTG, and multiple connectivity configurations were tested. Bayesian Model Selection revealed a dominant model in which PMv causally influenced IPL and pSTG, reflecting a feedforward architecture from affordance-related motor regions to semantic hubs. Bayesian Model Averaging further confirmed strong endogenous connections from LOC to PMv and IPL, and significant modulation from PMv to IPL. These findings provide direct evidence that affordance processing in premotor regions drives action language understanding by engaging downstream parietal and temporal areas. The results support grounded cognition theories and offer a mechanistic account of how sensorimotor information contributes to linguistic comprehension.