Location learning in virtual environments: The effect of saliency of landmarks and boundaries

In the present paper, we investigated the learning of object location in a virtual reality environment. Previous studies suggested that learning is essentially different when participants rely on landmarks (i.e., single stationary objects in the space) or boundaries (i.e., terrain features), the former requiring associative reinforcement learning, while the latter relying on incidental learning. However, in previous studies the saliency of the spatial cues and the position of objects relative to spatial cues were not carefully controlled, which could account for the specific learning effect found. In our study, we controlled these effects. We presented a virtual arena on a computer screen, featuring a landmark and a boundary cue, and four everyday objects, the location of which had to be learned and recalled. In the learning phase, either one or both cues were present, while in the test phase only one cue was present. Contrary to previous results, we did not obtain the overshadowing effect, as there was no difference in the location errors between the four conditions. At the same time, we found that location errors were influenced by the relative spatial position of the object and cue. This suggests that it is the distance from the cue and not the cue type, which is the main factor behind localization errors.