Abstract
The current study used a new flash preview/moving window paradigm to investigate the nature of the information initially acquired from a real-world scene. On each trial, participants were first shown a preview photograph for 250 ms. A word identifying a target object was then presented for 2 s. Finally, participants searched for a target while viewing the scene through a two-degree window that was centered at fixation and moved with the eyes. The scene was not visible outside the moving window. In the first experiment, the preview scene was either identical to the search scene, a different scene, or a meaningless mask. A flash preview benefit, indexed by more efficient eye movements to the target, was found for the identical preview condition over the different scene or mask conditions. The second experiment explored the contribution of scene category to the flash preview benefit. In this experiment, the preview scene was either identical to the search scene, a different scene from the same basic-level category as the search scene (i.e., same gist), or a scene from a different category. The same gist scene did not contain the target object or share the same spatial layout as the search scene. Results indicated that basic-level scene category alone was not the source of the benefit. In the third experiment, the contribution of target recognition to the flash preview benefit was explored by including or excluding the target object from the identical scene preview. Results from this and further experiments exploring other possible sources for this facilitation (e.g., extraction of spatial layout) will be reported. We conclude that the scene information acquired during an initial glimpse can be preserved in sufficient detail and over sufficient time to facilitate subsequent eye movements through the scene, and that the flash preview/moving window paradigm provides a useful tool for exploring the nature of this initially acquired scene information.