Parsing activity into meaningful events

People segment ongoing activity into meaningful events, which has important implications for memory and planning. Movement is one basis for segmentation. Perceptual experiments show that movement features are correlated with participants' segmentation of simple events. Neuroimaging data show that regions motion processing regions transiently increase activity at natural event boundaries. The strong relationships between movement and event segmentation may prove helpful for human-robot interaction. It may be possible for computational vision systems to automatically segment ongoing activity into meaningful units, such that those units can be used as the input to algorithms for recognizing the significance of a partner's actions.