Abstract
The coordination of spatial attention between observers is an element of many tasks. Facilitating this coordination requires that spatial information be communicated between observers, typically through verbal dialogue. However, verbal descriptions of spatial information can be ambiguous. In contrast, shared gaze has been shown to be highly efficient at focusing joint attention between partners in complex collaborative search tasks. We examined the utility of shared gaze in a dynamic helicopter search and rescue task. Paired observers searched for a hiker in a simulated low altitude fly-over of Yellowstone National Park. Both observers were required to locate the hiker before it left the screen. Each observer wore an eye/head tracker, was seated in a separate room, and viewed an independent display. Performance was assessed in three communication conditions: shared gaze (SG), shared verbal (SV), and shared gaze with verbal (SG+V). Surprisingly, observers were slowest to converge on the target in the SG condition (∼1.7s); performance in the SV (1s) and SV+G (960ms) conditions was roughly equivalent. To assess whether our findings could be attributed to distraction from the gaze cursor, or ease of the task, new observers performed the same task in both shared gaze onset (gaze cursor appeared once one partner had located the target; SGO) and static display experiments. In the SGO experiment observers were as quick to co-locate the target as in SV condition of the first experiment in which gaze information was constant; in the static display experiment observers were faster at co-locating the target in the SG condition compared to the SV condition. Our data suggest that shared gaze is useful for coordinating spatial attention, but this utility exists on a continuum. In easier search tasks gaze information is wasted, and perhaps even distracting; in difficult search tasks shared gaze provides clear benefits over verbal communication.