UCLA

Humans, as social creatures, are especially adept in perceiving others’ actions and inferring mental states such as intention based on non-verbal cues such as body movements. In recent years, substantial advances have been achieved in our understanding of biological motion perception and its relation to other cognitive and motorsensory processes, as well as the underlying neural correlates and computational components. In my thesis, I first focused on the top-down influence of meaningful interactive actions on stimulus visibility. We found that meaningful interaction boosted the visibility of single actor in binocular rivalry paradigm. Second, I investigated the action prediction modulated by action interaction (i.e. either single actor or two interactive actors) with temporal occluder paradigm. We found that observers obtained higher temporal precision in single actor condition but better sensitivity to posture sequence or rotation in two interactive actor condition. Third, we designed novel paradigm by briefly flashing joints near the moving limb to determine how action representation interacts with generic mechanisms for localization of moving features to bias perceived positions of moving joints. We found that global posture information (inversion or upside-down) modulated the effect size of well-established visual illusion (i.e. flash lag effect). Last but not least, I used this paradigm to dissociate local and global mechanisms that may underlie action prediction. By briefly flashing the stationary skeleton during the action viewing, either aligned or misaligned with the walking posture, we measured the motion-induced posture change. We demonstrated the existence of a global mechanism for anticipating changes in whole-body posture over time. However, although inversion of body orientation weakens global processing in biological motion perception, it does not eliminate the use of the global mechanism in predicting future posture even when the action is presented upside-down. This finding is consistent with previous studies of the inversion effect in biological motion perception (Pavlova & Sokolov, 2000; Sumi, 1984), and highlights the robust use of the global mechanism in predicting future actions.