top of page

Action concepts in the brain [New paper]

Our interaction with the external world influences the way we learn concepts and gain knowledge. However, little is known about how the brain processes and represents the notions we acquire. According to the theory of Semantic Embodiment, concepts are embedded in the same neural systems that were involved in their acquisition. Indeed, over recent years the sensorimotor system has been linked to the representation of action-related concepts. This means that, for instance, processing the concept “to kick” involves to some extent those motor areas that specifically contribute to the physical action of kicking. There is, however, contrasting evidence as some studies rather show a contribution of sensory visual areas, while other evidence suggests that the motor system may only be involved in service of certain task goals.

In our recent study, published in Scientific Reports, we hypothesised that, consistent with the theory of Semantic Embodiment, we would observe rapid activation in the sensorimotor system when neurotypical individuals access the meaning of action concepts. To this end, we visually presented motor and non-motor verbs to participants while recording their brain activity via electroencephalography (EEG). We asked participants to carefully read each word silently and, on a subset of trials, to monitor their engagement in the task we asked whether a specific word definition matched with the previously presented word.

In the EEG signal, we found an early differentiation (164-203ms) of the event related potential (ERP) between the processing of motor and non-motor concepts (see Figure below). Although this effect happened early in the processing stream and is therefore in line with previous findings on semantic embodiment, our source estimation failed to support the hypothesis of a specific role of cortical sensorimotor regions. Alongside this early differentiation in the ERP, we observed a late differential modulation (555-785ms) in the alpha frequency band (8-12Hz) when comparing motor and non-motor concept processing. This difference in alpha power could be localised to the left superior parietal lobule and, due to its timing, likely indicates post-lexical processing.

While our results support the hypothesis that the human brain differentiates very early in the processing stream between motor and non-motor concepts, we did not find evidence for involvement of sensorimotor cortices in that process. Our findings therefore do not provide new evidence for the contribution of sensorimotor cortices in the comprehension of action related words, as predicted by semantic embodiment. However, our data may offer valuable insight to the specific task goals and contexts in which such rapid sensorimotor activation may be elicited, thus refining theories of embodied cognition.

All data, materials, and analysis scripts for this study are available here.

[The Figure shows ERP scalp topographies from 164–203 ms post-stimulus. (A) Grand average topographies separated according to condition. (B, left) Grand average topography of the difference between conditions. Electrodes contributing to the two clusters are marked. (B, right) Tukey boxplots and individual subject mean voltages within the two significant clusters.]

bottom of page