Locomotion-dependent use of geometric and body cues in humans mapping 3D space.

The ability to represent locations across multiple dimensions of space is a core function of cognitive maps. While the influence of boundary-dependent environmental geometry on spatial representations has been extensively studied in 2D spaces, less is known about the role of boundaries for volumetric spatial memory. Research in humans and other animals has demonstrated distinct processing of the vertical and horizontal spatial dimensions, likely related to species-specific modes of locomotion. Here, we investigate whether different locomotion modes, flying and walking, affect the use of vertical boundaries, leading to possibly distinct volumetric representations. In a Virtual Reality experiment, human participants memorized objects within a symmetric 3D enclosure, and then were asked to replace them in either the familiar or geometrically deformed environments. We found that the flying group exhibited lower vertical than horizontal spatial memory precision, whereas the walking group showed the opposite pattern, an effect related to using their body axis as a vertical "ruler". Within deformed environments, object replacements in the flying group followed the predictions from a 3D-extended boundary-vector-cell-like computational model of spatial mapping that treated all boundaries equally, whereas those in the walking condition favored a modified model that prioritized the ground boundary. Our findings suggest that gravity-related movement constraints promote different utilization of geometric and body-related cues, resulting in flexible representations of volumetric space.