Abstract: Preview search is one of the new focuses in attention selection research in the last few years. In preview search task, a subset of distractors appears at least 400 ms ahead of the second subset, which may include the target. The time to search the target can be almost foreign to the number of the former distractors, only depending on the latter. This phenomenon is termed Preview Effect. The most acceptant interpretation on Preview Effect is the Visual Marking theory (Watson & Humphreys, 1997). According to this theory, the former distractors can be ignored or “marked” by top-down inhibition, which is based on the locations of the items. However, Braithwaite et al (2004)’ claimed that the inhibition was not only location-based, but also feature-based. They found some evidences by manipulating color proportions in both former and latter subset. This study designed a new method to certify the role of color in Preview Effect, called colored probe-dot technique. Using this method, a direct color process effect could be observed, without altering the color proportions in both subsets, which may alter the original pattern of preview search.
This study included two experiments, combining preview search task and colored probe-dot task. In both experiments, first a group of preview items were displayed followed by a probe-dot or a group of search items, depending on whether a sound clue appeared in the advance. In the first experiment, the color of probe-dot was that of the preview items or the irrelevant one, while in the second experiment, the color of the search items and the irrelevant color are the alternatives. 14 subjects participated the first experiment, and 16 subjects participated the second, all had normal or corrected-to-normal visual acuity and normal color vision.
The results showed that subjects detected the irrelevant-color probe-dot faster than the preview-item-color probe-dot in the first experiment, and a weak location effect was also observed. In the second experiment, the reaction time to the search-item-color probe-dot is shorter than the irrelevant-color probe-dot, no location effect showed.
We deduce that there are probably two kinds of color-based top-down processes, one is the inhibition set of the preview items’ color, and the other is the anticipatory set of the search items’ color. A location-based inhibition may play an important role in preview search too, depending on dissimilar task situation. The existence of color effect indicates Gestalt organization clues can greatly facilitate preview search. Besides different colors between former and latter subsets, temporal grouping and search array configuration are regarded Gestalt organization clues too

Key words: preview search, top-down process, probe-dot, color