What does neuropsychology say about the role of sensorimotor processes in conceptual knowledge and abstract concepts

Comprehension of words is an important part of the language faculty, involving the joint activity of frontal and temporo-parietal brain regions. Transcranial Magnetic Stimulation (TMS) enables the controlled perturbation of brain activity, and thus offers a unique tool to test specific predictions about the causal relationship between brain regions and language understanding. This potential has been exploited to better define the role of regions that are classically accepted as part of the language-semantic network. For instance, TMS has contributed to establish the semantic relevance of the left anterior temporal lobe, or to solve the ambiguity between the semantic vs. phonological function assigned to the left inferior frontal gyrus (LIFG). We consider, more closely, the results from studies where the same technique, similar paradigms (lexical-semantic tasks) and materials (words) have been used to assess the relevance of regions outside the classically-defined language-semantic network—i.e., precentral motor regions—for the semantic analysis of words. This research shows that different aspects of the left precentral gyrus (primary motor and premotor sites) are sensitive to the action-non action distinction of words' meanings. However, the behavioral changes due to TMS over these sites are incongruent with what is expected after perturbation of a task-relevant brain region. Thus, the relationship between motor activity and language-semantic behavior remains far from clear. A better understanding of this issue could be guaranteed by investigating functional interactions between motor sites and semantically-relevant regions.

Theories of embodied cognition propose that language comprehension is based on perceptual and motor processes. More specifically, it is hypothesized that neurons processing verbs describing bodily actions, and those that process the corresponding physical actions, fire simultaneously during action verb learning. Thus the concept and motor activation become strongly linked. According to this view, the language-induced activation of the neural substrates for action is automatic. By contrast, a weak view of embodied cognition proposes that activation of these motor regions is modulated by context. In recent studies it was found that action verbs in literal sentences activate the motor system, while mixed results were observed for action verbs in non-literal sentences. Thus, whether the recruitment of motor regions is automatic or context dependent remains a question. We investigated functional magnetic resonance imaging activation in response to non-literal and literal sentences including arm and leg related actions. The sentence structure was such that the action verb was the last word in the subordinate clause. Thus, the constraining context was presented well before the verb. Region of interest analyses showed that action verbs in literal context engage the motor regions to a greater extent than non-literal action verbs. There was no evidence for a semantic somatotopic organization of the motor cortex. Taken together, these results indicate that during comprehension, the degree to which motor regions are recruited is context dependent, supporting the weak view of embodied cognition.