Abstract
Eye movements often precede the motor behavior towards an object: we tend to first gaze at the object with which we intend to interact. Eye movements can thus reveal an individual's focus of attention and predict subsequent actions. Determining the direction of seen gaze toward a particular object in space (triadic eye gaze) is especially crucial in establishing joint attention. When an individual points to an object after having gazed at it, the availability of the cues indicating the chosen object dynamically changes over time. Initially, gaze direction is the only accessible cue, but as soon as the arm starts moving also the kinematic cues are available.
To investigate the temporal involvement of different cues we examined the eye movements of observers while they were watching videos of an actor performing gazing and pointing movements towards several target objects. Observers had to look at the target object as soon as possible. Two factors were manipulated: the presence of gazing behavior and the visibility of the target objects. The gaze triggered a rapid and accurate response on the target object. Observers were able to identify the target objects when the arm was still at the beginning of its trajectory, with fewer saccades and more accurately. When gaze information was not available, observers' gaze still led the hand movements of the actor, but was comparatively slower in identifying the target objects. The visibility of the target objects had an ameliorating effect on the spatial accuracy.
These findings suggest that other's gaze direction is an essential predictive cue about the final location of a pointing movement. Observers thus activate action plans based on the actor's gaze direction and kinematic cues to produce proactive eye movements.
Research supported by EU grant (FP7-ICT-217077-Eyeshots).