A Bayesian Model to Describe the Relationship between Gaze Pattern, Task Familiarity, and Search Efficiency in a Visual Foraging Task

Foraging is common in everyday life and studying search for multiple targets can shed light on underlying mechanisms of multiple visual processes. The current study had participants forage in virtual reality to allow for more naturalist eye and head movement, shooting targets amongst distractors. Eye-tracking metrics were used to unpack what strategies were related to search efficiency in earlier versus later trials where performance improved with task familiarity. A Bayesian hierarchical model showed that, consistent with previous findings, saccade frequency and fixation duration of targets were related to more efficient foraging. The relationship with saccade frequency was also stronger in later trials compared to earlier. Interestingly, longer fixation durations on distractors also related to more efficient foraging, but this pattern disappeared or was reversed when individual distractor types were examined. Our results highlight how gaze patterns may be altered to coordinate with other body movements in unconstrained search.