A Two-Dimensional Framework for Prior Knowledge in Goal-Directed Navigation

Spatial navigation involves integrating spatial cues with prior spatial knowledge to remember locations and maintain orientation. This study examined the influence of spatial priors (pre-existing knowledge about spatial parameters) using a spatial memory task in desktop virtual reality. We proposed a novel two-dimensional framework for categorizing spatial priors based on spatial reference frame (allocentric vs. egocentric) and temporal scale (global vs. local). Participants memorized and later reproduced target locations. The spatial compactness of the target distribution was manipulated to vary the precision of the allocentric global prior. A central tendency effect, where responses were biased toward the target distribution centroid, indicated reliance on the allocentric global prior. Bias was stronger with more precise distributions, indicating greater reliance on the more precise allocentric global prior. Bidimensional regression analysis revealed additional influences from the egocentric global and egocentric local priors, but not from the allocentric local prior. These findings support the proposed framework and demonstrate that multiple forms of prior knowledge simultaneously guide navigation behavior. This work provides a foundation for future studies linking prior knowledge to spatial navigation and broader cognitive processes.