Abstract
Under incidental learning conditions, a spatial layout of visually-presented objects can be acquired implicitly and can be used to facilitate visual search (contextual cueing effect; Chun & Jiang, 1998). We studied this type of implicit learning between successive trials using the “contextual cueing” paradigm. Participants received 540 trials of visual search for a rotated ”T” among rotated “L”s. The experiment consisted of training, testing, and recognition phases. In the training phase (450 trials), the pattern formed by the target and distractors on the N-1th trial was reliably associated (correlated) with a specific but different target location on the Nth trial (the distractors on the Nth trials were randomized). That is, the layout of all items in the N-1th trial predicted the target location in the Nth trial. In the test phase (90 trials), the correlation between successive trials was removed so that the pattern in the N-1th trial no longer predicted the target location in the Nth trial. The results revealed clear inter-trial contextual cueing effects: reaction times in the Nth trial increased significantly when the correlation was removed, even though the participants were unaware of the correlation during the training phase. Four additional experiments investigated the critical factors for the inter-trial contextual cueing effect. Experiment 2 suggested that the effect does not occur when only the target location in the N-1th trial was correlated with the target location in the next trial. Experiments 3, 4, and 5 indicated that distractor layout, not target location in the N-1th trial was the critical factor for the cueing effect. We concluded that the correlation between the distractor layout in the N-1th trial and the target location in the Nth trial was learned implicitly, and that this memory guided spatial attention to the target location in the Nth trial.