Unreliable landmarks disrupt egocentric navigation in desktop virtual environments

Navigation relies on the integration of both allocentric (world-centered) and egocentric (navigator-centered) spatial cues. An influential real-world navigational task dissociated allocentric place learning from egocentric path integration, highlighting the decline in allocentric cue use associated with cognitive aging. This report describes a novel computer-based replication of this task, designed as a remote test of navigational ability, that also addresses potential confounds in the original task, including tighter control of spatial cue competition. In Experiment 1a, participants learned a location signaled either by landmark cues in the absence of a stable starting position (allocentric), distance and direction from the starting position in the absence of landmarks (egocentric), or a combination of the two (control). In the allocentric condition, participants exhibited lower accuracy, greater reorientation, and more complex paths compared to the control and egocentric conditions. Experiment 1b employed the same design but featured an egocentric condition with unstable landmarks that did not predict the target location. In Experiment 1b, egocentric condition accuracy was equivalent to the allocentric condition and lower compared to Experiment 1a's egocentric condition. Additionally, in Experiment 1b's egocentric condition, participants took more complex paths and reoriented more to the unstable landmarks. Experiments 2a and 2b replicated findings from Experiments 1a and 1b, indicating their reliability. These data suggest that path complexity is associated with navigational error and underscore the benefits of integrating different sources of information during navigation. Cross-experiment comparisons also show how the presence of unstable visual cues can lead to the adoption of unsuitable strategic approaches. (PsycInfo Database Record (c) 2026 APA, all rights reserved).