MOTOR ACT CATEGORIZATION IN HUMAN PARIETAL CORTEX

Functional imaging studies in humans and macaques have shown that observation of others’ actions leads to reliable activation in occipito-temporal, intraparietal and premotor cortex. However, much less is known about the specific roles of these areas in the process of action observation. Previous work has indicated that parietal and premotor cortex might be organized based on two different principles. While in the premotor cortex, the observed actions were clustered according to the effector used, the intraparietal cortex appeared to be organized on the basis of motor act categorization (Jastorff et al., Soc Neurosci Abs, 2007). Yet, in the original study, we had to equate the mean activation across conditions to correct for overall differences in stimulus saliency. In a new experiment, we tested 15 subjects with video sequences, showing four different hand actions (dragging, dropping, grasping and pushing) together with non-biological motion control stimuli, where the hand movement was replaced by a moving object superimposed on a single frame of the original movie. All actions were presented within the same run and subjects performed an attentionally demanding high-acuity task during the scanning, to equate the level of attention. Following the lead from the previous study, we grouped the actions according to their relation between the object and the agent into positive actions (moving the object towards the agent; dragging and grasping) and negative actions (moving the object away from the agent; dropping and pushing). Random effects analysis identified two significant interaction sites (p < 0.001 uncorr.) in the left parietal cortex, in close proximity to putative human AIP. A ventro-rostal one showing significantly stronger activation for positive actions and a caudo-dorsal one more strongly activated for negative actions. Most importantly,the analysis of the fMRI signal for the two interaction sites defined in our previous study, revealed a significant interaction between these a priory ROIs and the sign of the motor acts in this new experiment. In an additional control experiment, we scanned again 7 of the 15 subjects of the previous experiment showing them the same actions, now carried out with the foot and with the mouth. Probing the fMRI signal changes within the two ROIs defined for positive and negative hand actions, we found that the organization based on the action sign was independent of the effector used to perform the action. Therefore, we hypothesize that the parietal organization, by generalizing the behavioral meaning of the motor act across effectors, allows a unified understanding of the observed behavior in terms of its positive or negative signs.