Abstract
Under incidental learning conditions, the 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 investigated the nature of the implicitly learned spatial layouts in stereo (3-D) displays. Experiment 1 examined whether the learned layouts are limited to 2-D configurations or can encompass three dimensions. Subjects received 720 trials of a visual search task with stereo shutter goggles so that half the items were in the frontal plane and the others were in the back. They searched for a rotated “T” among rotated “L”s. In the OLD condition, 12 different layouts were presented repeatedly for 360 trials. In the NEW condition, a new layout was generated on each of 360 trials. In the last 60 trials of the OLD condition, the disparity of the items was reversed. Reaction times in the OLD condition were faster than in the NEW condition even though the subjects were not able to recognize the OLD layouts (contextual cueing effect). More important, reaction times in the OLD condition increased significantly when the disparity was reversed, suggesting that the 3-D layouts had been implicitly learned during the initial 300 trials. Experiment 2 investigated whether 3-D layouts are encoded automatically or requires selective processing. Subjects searched for a target presented only in a specific depth plane (e.g., front, never in the back). A contextual cueing effect was obtained only when the location of the items in the attended plane was invariant and was consistently paired with a target location. In contrast, repeating and pairing the layout of the ignored items with the target location did not produce contextual cueing effect. These results indicate that visual attention can be guided by implicit knowledge in 3-D layouts, provided that the layout is relevant and selectively attended.
Support: Japan Society for the Promotion of Science