Abstract
Humans are able to identify self from body movements presented in the point-light display, despite little visual experience of seeing own actions. However, the neural mechanisms supporting visual self-recognition from body kinematics in actions remain unknown. Here, we utilized functional neuroimaging to examine which brain networks may support visual self-recognition from body movements.
Eleven participants performed six simple actions (e.g., jump) and six complex actions (e.g., get attention), which were recorded by a motion capture system. One of each participant's friends (gender-matched) was also recruited to record the motion capture data for performing the same actions. After a delay period of 12 – 20 days, the participants ran a fMRI session in which they viewed point-light actions either performed by the participant self, or by the gender-matched friend or stranger. Participants were asked to make a self-recognition judgment among three choices (i.e., self, friend, or stranger). Each point-light action stimulus was present for 5 seconds. The experiment consisted of 144 trials.
Accuracy in self-recognition from point-light displays (M= 0.48) was significantly above-chance. Brain activity was compared between self-action and stranger-action. We found that own-body movements activated bilateral IPL (key area of mirror neuron system, important for action goal identification) and bilateral TPJ (key area of higher-level mentalizing system, important for the explicit self-other distinction). Furthermore, we found that familiarity modulated right TPJ activity, with greatest neural activity for self, moderate activity for friend, and little activity for stranger. These results provide converging evidence that motor experience is critical for the construction of the bodily self. Specifically, the more effortful and controlled reasoning of the mentalizing system builds upon the automatically and reflexively recruited mirroring system for the construction of the hierarchical, multimodal self.