Grounding language in sensorimotor systems revealed by fMRI and TMS

International audienceOne of the claims of embodiment theory is that linguistic meaning is grounded in sensorimotor systems. For example, the simulation approach to language comprehension asserts that language becomes meaningful by simulating the content of a sentence using the same sensory and motor systems that might be used in dealing with the real events described by the sentence. This claim has been supported by behavioural work demonstrating interactions between language-based tasks and perception or action tasks. The work that I will describe provides fine-grained support for the claim by demonstrating interactions at the level of neural systems. In one experiment using fMRI (see Rueschemeyer, Glenberg, Kaschak, & Friederici, submitted), we tested the fine-grain prediction that comprehending sentences describing visual motion (compared to static visual scenes) would selectively activate an area of cortex known to be used in analyzing visual motion, namely V5/MT. We observed this effect with a twist: modification of activation in V5/MT was observed only for sentences describing action toward the listener (e.g., “The car drives toward you”) and not for sentences describing motion away from the listener (e.g., “The car drives away from you”). In a second experiment (see Glenberg, Sato, Cattaneo, Riggio, Palumbo, & Buccino, in press), we used transcranial magnetic stimulation to probe the motor system during language comprehension. Previous work has demonstrated that single-pulse TMS over motor cortex activates the corticospinal pathway, and that this activation can be measured in muscles in the form of motor evoked potentials (MEPs). Pairing the TMS pulse with other stimuli (e.g., sentences) may produce significant variation in the size of the MEPs which can then be attributed to an effect of those other stimuli on motor processes. In this experiment, participants read sentences describing transfer (e.g., “Tony gives you the pen”) or no transfer (e.g., “Tony and you look at the pen”), and we measured MEPs in the opponens pollicis muscle that participates in thumb movements. If understanding a transfer sentence involves simulating action using the motor system (and those systems that control the thumb, in particular), then we should observe greater modulation of the MEPs for the transfer sentences than the no-transfer sentences. We observed this effect for sentences describing the transfer of concrete objects as well as sentences describing transfer of abstract information (e.g., “Tony delegates the responsibilities to you.”)