Abstract
In visual search tasks, participants rarely revisit items they have already examined. Therefore, participants are using memory to guide visual search (Peterson et al, 2001). However, the nature of this memory is still unclear. Here, we examined whether memory in visual search is identity based, location based, or both. In the search displays, each of the 12 items appeared inside one of 16 unique color-shape combinations (e.g. a red triangle). The target was a rotated T and distractors were rotated L's. In the first experiment, on about 2/3 of the trials, the surrounding shape of one of the items changed into another unique color-shape combination during a saccade. The changed item was either an item that had not been looked at (new-change) or an item that had been looked at (old-change). After the change, old-change items and old-static items were reexamined equally often, suggesting that when an old item is changed, it is not considered a new item. Likewise, the new-change items were examined at the same rate as new-static items, indicating that the change did not lead to a capture of attention. In a second experiment, an old or new item was removed from the screen for two to five fixations and then reappeared. Old-change items were reexamined at the same rate as old-static items, and new-change items were examined slightly more than new-static items. Therefore, temporarily removing an old item from the display either did not disrupt memory for the item or indicates that the memory is space-based. Furthermore, in both experiments, participants rarely revisited old items. These data support previous research demonstrating the role of memory in visual search, and they demonstrate that memory in visual search is not dependent on the identity of the items or their continued presence on the screen. This suggests that memory in visual search may be based on item locations rather than object representations.
NIH: 1R01MH64505-01