Joint attention supports working memory when gaze cues are reliable and task-relevant.

This study investigated whether attentional orienting in response to gaze cues enhances visual working memory (WM) automatically, or whether engagement of top-down processes is necessary for such effects to emerge. Building on an existing gaze-cueing paradigm, we tested whether joint attention supports WM under two conditions. In Experiment 1, participants viewed centrally presented static images of human faces displaying directional gaze cues without any instruction to use gaze direction, and gaze validity was set at 50%, making the cue spatially uninformative of stimuli location. Following the cue, a memory array was presented, followed by a retention interval and a single-probe recall. Participants had to indicate whether the probe had appeared in the initial memory set. No WM advantage was found for validly cued items. In Experiment 2, we increased cue validity to 75% and explicitly informed participants that gaze direction was highly predictive of stimuli location. Under this condition, which presumably elicited higher engagement of top-down processes, valid gaze cues significantly enhanced WM performance relative to invalid cues. Interestingly, as cognitive load increased, the limited capacity of WM slightly constrained the extent to which this strategic orienting could translate into improved memory sensitivity. These results highlight the interplay between cue reliability, attentional control, and WM capacity in determining the efficacy of gaze cues. Our findings clarify the conditions under which joint attention facilitates WM and contribute to a growing literature showing that social attention effects on higher-level cognition are context-sensitive and cognitively mediated.