Abstract
There is extensive empirical evidence that Working Memory (WM) representations can bias selection in Visual Search (VS) tasks (reviewed in Soto & Humphreys, 2014). Less has been said about the guidance of attention in VS by the contents of Long Term Memory (LTM), but some studies have shown effects from LTM representations on VS (e.g. Olivers, 2011). As a general rule, when an item to be held in WM or LTM appears as a 'lure' in the visual search display, search is slowed. Here we report conditions under which search is actually speeded by the presence of lures. Fourteen observers searched for a different target on every trial. Targets and distractors were photographs of objects. Observers held either 10 or 32 items in LTM and another item was presented before each trial to be held in WM. A true target was present on every trial. The VS display of 16 items could also contain a WM lure, a LTM lure, both, or none. Observers localized the target with a mouse-click. Under these conditions, RTs were generally faster when lures were present. This was particularly evident when both LTM and WM lures were present. We speculate that this represents a type of familiarity effect. The target is new on each trial and, thus, not very familiar. The lures are more familiar. A rule that rapidly rejects familiar items could speed search if lures can be quickly rejected and might speed search more when two lures can be rejected. This effect is not seen when targets are not present on every trial. Lures tend to slow search in those conditions, perhaps because any familiarity becomes a sign of possible target presence when target presence is uncertain. Both the speeding and slowing of search show the influence of memory on search.
Meeting abstract presented at VSS 2016