The effects of goal–landmark distance on overshadowing: a replication in humans (Homo sapiens) of Goodyear and Kamil (2004)

Goodyear & Kamil (2004) assessed the ability of Clark’s nutcrackers to find buried food based on a cross-shaped array of landmarks at different distances from the goal. Their findings suggested that close landmarks overshadowed learning about distal landmarks, and this was attenuated when assessing the effect of distal landmarks on learning about close landmarks. In other words, the extent of overshadowing was moderated by landmark distance. In this study, we aimed to replicate their findings in human spatial navigation by using a virtual environment. Three groups of participants were trained in an open environment featuring orientation cues and they had to find a hidden goal with reference to 4 landmarks that were arranged in the shape of a cross and placed at different distances from the goal. Two of the four landmark distances were common across all three groups to allow a comparison of the extent of overshadowing under comparable conditions. Following training, all participants were tested with each of the 4 landmarks individually. Of particular interest was how well participants performed when tested with the common landmarks. Consistent with the results in birds, we observed better performance in the groups with more distal landmarks, suggesting that overshadowing was greater in the groups with closer landmarks and thus dependent on the spatial contiguity between the landmarks and the goal. Landmarks near the goal overshadowed landmarks far from the goal, but the opposite was not true. A second experiment, in which landmarks and orientation cues were misaligned in order to prevent the use of a straightforward solution to the task, replicated the results. The results are discussed in terms of a modification of Pearce’s configural model.