Abstract
Previous studies of eye fixations during language encoding have found that fixation attention allocated to either the eyes or the mouth depends on factors such as task demands and language familiarity. These studies have only considered cases in which the observer is viewing a prerecorded video, where the observer's fixation behavior cannot be viewed by their interlocutor, in contrast with real-world social interactions in which the interlocutor can observe said fixation behavior. Under such real-world conditions, the fixation behavior of the observer may also serve to communicate with the interlocutor, for example by signalling interest via gaze locking or communicating emotional expression. This absence of bidirectional communication in previous studies may lead to altered fixation behavior. To test this, we compared eye fixation patterns for 'Live' vs. 'Prerecorded' viewing of a talking face. In order to achieve maximal similarity across conditions, both used the same two recorded video sequences of a presented article, each with instructional segments and story segments. One video (Live condition) used deception to encourage observers to believe that they were interacting with a live person who was able to see and hear them through online remote video communication, and in the other (Prerecorded condition) observers were informed that they were watching a previously recorded video. We found that eye fixations were more heavily focused on the eyes in the Live condition than in the Prerecorded condition. Additionally, we found that the type of information (instructional vs story) resulted in differing patterns of facial eye fixations, with instructions resulting in more fixations of the eyes than the story portion. The findings demonstrate that the belief that an interlocutor can observe gaze direction modulates the allocation of attention towards facial regions during language encoding, and suggest that previous studies may lack ecological validity.
Meeting abstract presented at VSS 2016