Abstract
When subjects begin a visual search trial there is an initial delay in response, while the items are being initially processed and a hypothesis is generated. But on subsequent exposures to the same display, search can be rapidly resumed by testing a previously generated hypothesis against the display (Lleras, Rensink, & Enns, 2005). Here we examine the nature of these hypotheses, looking in particular at their spatial extent. To do so, we looked at the effect of changing the location of 25% of the distractor items between search exposures in comparison to a no-change baseline. On each trial, a display was shown for 500ms and then disappeared for 1600ms. Then the display was shown again, and this cycle was repeated until the target was found. When items near the target moved during the blank period between exposures to the display, it interfered with hypothesis testing and significantly impaired subjects' ability to rapidly resume search. However, when items distant from the targets were moved between exposures, there was no detriment to rapid resumption. Thus, the hypotheses that guide rapid resumption of search are restricted to a local subset of items.
Supported in part by NIH EY014193