Distraction from long-term memory in visual search: acquired procedural and template-based memory interfere with contextual cueing after target re-location.

Abstract

Participants can learn to detect search-for target items embedded in repeatedly encountered spatial arrangements of distractor items – the ‘contextual-cueing’ (CC) effect. However, cueing is severely compromised following the re-positioning of the target in an otherwise unchanged distractor arrangement. Previous research demonstrated that this target re-location cost is due to persistent contextual (mis-)guidance of search towards the original, but (after the re-positioning) no longer relevant location in the display array. Given that CC reflects top-down guidance for contextual long-term (LT) memory, this misguidance is an instance of a ‘distraction’ effect that operates from acquired memory, rather than being driven by salient but irrelevant stimuli in the display. While traditional accounts of CC emphasize the acquisition of search-guiding LT-memory ‘templates’ that are specific to particular target-distractor contexts, contextual learning also tunes attentional (oculomotor) scanning routines to the prevailing statistical target-distractor regularities in the display arrangement encountered, yielding a context-unspecific LT ‘proceduralization’ of search. Using reaction-time (RTs) and oculomotor-scanning measures, we confirmed both mechanisms to contribute to initial contextual learning as well as the ‘distraction’ effect produced by re-location of the target to the opposite side of repeated-context displays. We suspect that guidance and misguidance of search by repeated contexts involve two complementary LT mechanisms: procedural optimization of broad, i.e., display-generic, oculomotor scanning routines, and learning of where to expect the target to be located in specific repeated-context displays.