Abstract
From about one year of age humans are able to reliably follow gaze cues from others. These cues have been shown to covertly direct attention in adults (Langton & Bruce, Visual Cognition, 1999). Additionally, action towards an object is often preceded by gaze toward that object (Land & Hayhoe, Vision Research, 2001). Because gaze cues can direct attention and reliably precede future action, we conjectured that they might be used in the prediction of human action. To investigate this hypothesis we examined whether the amount of representational momentum caused by brief videos of reaching actions depended on gaze cues. Participants watched videos that contained an initial grasping action, but that stopped just prior to the start of a second grasp. Three different kinds of gaze cue preceded the stopping point: a shift in gaze from looking down to looking at a to-be-grasped object (Object Gaze Cue), a shift in gaze from looking at the camera to looking at the object (Communicative Gaze Cue), or no gaze cue (Control). Subjects indicated whether a probe image was the same as the last frame in the video. Critically, the examination of representational momentum was based on gaze cue alone, before any action of the hand and arm was visible. We found that communicative gaze cues produced the most representational momentum, as revealed by a relative predominance of false alarms in recognition of frames advanced in time relative to the stopping point (that is, frames depicting the beginning of the second reach), relative to recognition of the stopping-point frame. Both the object gaze cue and the control conditions produced less representational momentum, and were not different from each other. This finding suggests that communicative gaze invokes perceptual predictions about the short term course of upcoming actions.
Michael Mack, Stephen Killingsworth