Given limitations in the amount of visual information that a person can simultaneously process through to conscious perception, selective visual attention is necessary. Visual signals in the environment aid this selection process by triggering reflexive shifts of covert attention to locations of potential importance. One such signal appears to be others' eye gaze. Indeed, a gaze-cueing effect, whereby healthy adults respond faster to targets that are presented at locations cued rather than miscued by eye gaze has been consistently observed in the empirical literature. Critically though, the influences of task and cue features on this effect are not well understood. To address this gap, we report a meta-analytic integration of 423 gaze-cueing effects using a multilevel approach. A gaze-cueing effect emerged across all levels of the assessed task and cue features, indicating that others' eye gaze reliably directs observers' attention. We found that the magnitude of the gaze-cueing effect was moderated by whether direct gaze cues preceded directional gaze cues or not; the cue-target stimulus onset asynchrony (SOA), whether participants had to detect, localize, or categorize targets; and the cue's facial expression. Whether or not the gaze cue remained on screen after the target appeared, and whether schematic faces, computer-generated faces, or images of real faces were used as cues, did not appear to reliably function as moderators. The theoretical implications of these findings are discussed, particularly in relation to the social attention system. Public Significance StatementThe findings from this meta-analysis provide the strongest evidence to date that others' eye gaze is a powerful social cue that reliably directs observers to reflexively shift their own visual attention such that they engage in joint attention with the gazing person. The findings also provide insights into the specific circumstances in which this reflexive gaze-cued attention is enhanced.

Objectives The present study provides a meta-analytic assessment of how gaze-cued attention – a core social-cognitive process – is influenced by normal adult ageing. Method A multi-level meta-analysis of standardized mean changes was conducted on gaze-cueing effects. Age effects were quantified as standardized mean differences in gaze-cueing effect sizes between young and older adult samples. Results We identified 82 gaze-cueing effects (k = 26, N = 919 participants). Of these, 37 were associated with young adults (k = 12, n = 438) and 45 with older adults (k = 14, n = 481). Relative to younger adults, older adults had a reduced gaze-cueing effect overall, g = -0.59, with this age effect greater when the cues were predictive, g = -3.24, rather than nonpredictive, g = -0.78. Discussion These results provide the clearest evidence to date that adult ageing is associated with a reduction in gaze-cued attention. The results also speak to potential mechanisms of this age effect. In line with cognitive decline models of ageing, it was demonstrated that when gaze cues were predictive, only younger adults seem to benefit, suggesting that older adults exhibit a particularly reduced capacity to use gaze cues volitionally.

No abstract

Given limitations in the amount of visual information that a person can simultaneously process through to conscious perception, selective visual attention is necessary. Visual signals in the environment aid this selection process by triggering reflexive shifts of covert attention to locations of potential importance. One such signal appears to be others’ eye gaze. Indeed, a gaze-cueing effect, whereby healthy adults respond faster to targets presented at locations cued rather than miscued by eye gaze, has been consistently observed in the empirical literature. Critically though, the influences of task and cue features on this effect are not well understood. To address this gap, we report a meta-analytic integration of 423 gaze-cueing effects using a multi-level approach. A gaze-cueing effect emerged across all levels of all assessed task and cue features, indicating that others’ eye gaze reliably directs observers’ attention. We found that the magnitude of the gaze-cueing effect was moderated by whether direct gaze cues preceded directional gaze cues or not; the cue-target Stimulus Onset Asynchrony, whether participants had to detect, localize, or categorize targets; and the cue’s facial expression. Whether or not the gaze cue remained on screen after the target appeared, and whether schematic faces, computer-generated faces, or images of real faces were used as cues, did not appear to reliably function as moderators. The theoretical implications of these findings are discussed, particularly in relation to the social attention system.