Visual search strategies are known to capitalise on the relationship
between the target and the distractors – i.e., how the target differs from
the distractors and where the target is likely to be amongst the
distractors. Additionally, it has been proposed that statistical structure
amongst distractors can also benefit search, enabling observers to parse
the scene more efficiently by reducing the number of distractors to contend
with. Evidence for this hypothesis comes from studies using everyday object
stimuli, showing that search efficiency improves when distractor objects
are positioned in familiar arrangements (e.g., lamp above table; Kaiser et
al. 2014). Here we asked whether the benefit of distractor structure
extends to novel shapes, whose spatial relationships (groups) must be
learned implicitly during visual search. By using novel shapes, potential
low-level visual differences and semantic relationship differences between
conditions can be ruled out. Participants searched for pre-cued target
items in arrays of shapes that comprised either four fixed pairs of shapes
(structured displays) or eight shapes randomly partitioned into four pairs
on each trial (unstructured displays). The distractor positions varied
across trials. Across three online experiments (N=700), we found that after
a period of search training, participants were more efficient when
searching for targets in structured displays, even though post-experiment
they could not identify which pairs were fixed in a two-alternative
forced-choice task. These results show that implicitly learned statistical
regularities between distractor shapes increase search efficiency.
[Funding source: Horizon
2020 Framework Programme (725970)]