Memory for salient landmarks: Empirical findings and a cognitive model

Abstract

We test the effect of a landmark’s visual and structural salience on memory retrieval of turning directions at choice points in a VR environment. We find a higher probability for a correct turning decision at intersections where the location of the visually salient landmark converges with the turning direction as compared to intersections where the location of the visually salient landmark diverges from the turning direction. Although altered versions of the intersections were mostly recognized as being novel, we found systematic error patterns depending on the placement in the original intersection. A cognitive model in the ACT-R architecture grounds these findings in an established framework of human memory. Our findings have implications, for example, for the selection of suitable landmarks for navigation assistance systems.