Visual Scanning Patterns during Remote Diagnosis

Abstract

Through audio-video-data linkage, remotely situated experts can potentially provide supervisory advice to on-going events and activities and participate in dynamic decision making. Previous studies showed, however, that domain experts often missed critical cues necessary for assessing remote situations and thereby were unable to anticipate needed interventions and to provide effective supervision. Questions arose about the process whereby remote decision makers extracted information from audio-video-data linkage. This paper reports a study of the visual scanning patterns of domain experts during remote diagnosis of trauma patient resuscitation through audio-video-data linkages. Eye-gaze data were collected from decision makers of different domain expertise as well as from novices. The results indicated that (1) experts tended to focus on regions of the video scene corresponding to their respective real-life roles in a multidisciplinary team; (2) the majority of eye fixations were on the face or the head of the people in the video, regardless of viewers' experience backgrounds; (3) large individual differences existed in how systematically the subjects scanned various parts of the video scene which included vitals signs of the patient.