Contextual Cueing of Visual Search Tasks: The Elicitation of Implicit Memories in Observer Performance Studies

Abstract

Researchers who design observer performance studies often take steps to negate the effects of memory for radiographs shown in separate viewing sessions. For example, it is common to build in a time gap between viewing sessions and present images in different, random orders. However, little is known about the degree to which radiologists recognize specific radiographs. In the present study, I conducted 1) a contextual cueing task that assessed whether previously seen spatial configurations facilitate implicit memory formation and 2) a delayed recognition test that assessed whether radiologists recognize chest radiographs shown in separate viewing sessions. The results of the contextual cueing task mirrored those of previous studies, in that reaction times for target localization significantly decreased for old (viewed once per block) arrays in comparison to new (never repeated) arrays, and memory for old arrays was implicit. Furthermore, vertically reflecting old arrays eliminated the contextual cueing effect; this locally-invariant remapping significantly increased reaction times for target localization. The results from the observer performance study show that radiologists spend less time assessing old radiographs in comparison to new radiographs (i.e. contextual cueing). Furthermore, the results show that this memory is implicit, as radiologists were not able to distinguish between new and old radiographs. Together, these findings have implications for alternative procedures and design guidelines for observer performance studies in which radiographs are reused in separate viewing sessions.