Positional Priming of Visual Pop-Out Search Is Supported by Multiple Spatial Reference Frames

Abstract

The present study investigates the representations(s) underlying positional priming of visual pop-out' search (Maljkodc and Nakayama 1996). Three search items (one target and two distractors) were presented at different locations in invariant (Experiment 1) or random (Experiment 2) cross-trial sequences. By these manipulations it was possible to disentangle retinotopic spatiotopic and object-centered priming representations. Two forms of priming were tested: target location facilitation (i.e. faster reaction times - RTswhen the trial n target is presented at a trial n-1 target relative to n-1 blank location) and distractor location inhibition (i.e. slower RTs for n targets presented at n-1 distractor compared to n-1 blank locations). It was found that target locations were coded in positional short-term memory with reference to both spatiotopic and object-centered representations (Experiment 1 vs. 2). In contrast distractor locations were maintained in an object-centered reference frame (Experiments 1 and 2). We put forward the idea that the uncertainty induced by the experiment manipulation (predictable versus random cross-trial item displacements) modulates the transition from object- to space-based representations in cross-trial memory for target positions.