Abstract: PURPOSE: The ability to ignore distracting information is crucial for improving visual search. Recent studies have shown that selection history can bias attention and 2D locations that may contain salient distractors can be suppressed through statistical learning. However, depth information has been shown to function differently from 2D spatial information in various visual tasks. This study aimed to investigate whether statistical regularities based on depth information could influence attention capture in 3D settings.
METHODS: We manipulated the probabilistic occurrence of salient color distractor’ depth location (i.e., distance from viewer) and tested two search modes. Experiment 1 employed the addition-singleton paradigm, where participants searched for a shape singleton and ignore the salient distractor (singleton-detection mode). Similarly, participants searched for specific shape among four different shapes (feature-search mode) in Experiment 2. Experiment 3 utilized ordinal depth information (i.e., relative depth order) and replicated Experiment 2 to generalize our finding.
RESULTS: For the depth location (or order) where salient distractor was presented with higher probability, we found significant slower response for target selection, indicating learned attention suppression for depth. This finding was replicated in depth order display. However, a larger attention suppression effect under feature-search mode demonstrated different processing of depth from 2D location.
CONCLUSIONS: Statistical learning led to shaped attention selection away from depth location or order of distractor, suggesting that depth information could be learned implicitly. Nevertheless, depth location may be processed as a ‘feature,’ resulting in a larger suppression effect under feature-search mode.

Key words: selection history, attention suppression, depth, statistical learning