Abstract
An observer will often pay greater attention to an object that is the target of another person’s gaze than to other objects in a scene, a phenomenon known as gaze cuing. What about objects that were the target of another person’s gaze in the recent past? We used the flicker paradigm to address this question. In this paradigm, two versions of a scene are cyclically presented with a brief blank in between each presentation. Under such conditions, observers often fail to see the change until attention is directed to the changing region. In experiment 1, participants (N = 21) searched for changes in pictures of complex, real-world scenes. For each participant, half the scenes contained a valid a gaze cue (i.e. a person looking at and reaching for the changing object), and the other half did not (i.e. no person was present). The stimulus set was created such that each scene could be used in each condition across participants, thus controlling for low-level differences between the cued and un-cued changes. As expected, reaction times were faster when a gaze cue was present (M = 7.34 seconds) than not (M = 11.22 seconds; t(20) = 2.55, p = 0.019). In experiment 2, on each trial participants (N = 50) viewed a one-second preview the scene before the flicker sequence began. Half of the previews contained a valid gaze cue, and the other half did not. The scenes presented in the flicker sequence did not contain people at all. Reaction times were faster when the preview contained a gaze cue (M = 4.11 seconds) than when it did not (M = 11.56 seconds; t(49) = 7.18, p < .001). These results suggest that memory for gaze can be as effective as perception of gaze at directing attention in the flicker paradigm.