Abstract
Introduction: Gaze, head, and body postures are socially important cues to indicate future actions and intentions (Emery, 2000) and orient attention reflexively (Driver et al., 1999; Friesen et al., 2004). Most previous studies used static drawings of heads/bodies and simple targets (Azarian et al., 2017; Bayliss et al., 2004). Here, we investigated how people’s dynamic looking behaviors that include their head and body movements contribute to the orienting of attention in an ecologically relevant person search task. Method: Thirty subjects searched for a target person (yes/no task; present in 50 % of trials) in sixty videos of real-world scenes while maintaining central fixation. One through three individuals in the videos dynamically looked towards the same location. After a random onset ( 200ms or 500ms), additional individuals appeared in the video, including at the “looked at location”. In half of the target-present trials, individuals in the videos oriented their gaze to the target’s location (valid-look trials), while in the remaining 50 % they looked at a distractor person (invalid-look trials). Videos were manipulated so that looking individuals (but not targets and distractors) contained: entire silhouettes, floating heads, and headless bodies. Results: The cueing effect (d’valid – d’invalid) was strongest with the joint presence of heads and bodies and temporally sustained (Δd’ 200ms: .34, p<.05; 500ms: .45, p<.01). The cueing effects for the floating heads and headless bodies were not significant for the 500ms delay (Δd’= .06, p =. 66; Δd’ = .13, p = .62m, respectively). We also found a sustained bias in microsaccades direction toward the validly cued direction for the silhouettes (200ms :70.34%, 500ms: 64.46%, both p<.05) but not for the floating heads and headless bodies conditions. Conclusion: The joint presence of heads and bodies is critical for engaging sustained covert attention and microsaccades during search in dynamic scenes.