Cassandra Engstrom¹, William H. Warren¹; ¹Brown University

Previously, we found that humans walking through mazes reduce their angular deviation from the beeline to their goal, although they also reduce turns away from their current walking direction (heading) or body orientation. Simulations demonstrated that compared to other uninformed strategies, reducing beeline deviations generates the shortest paths. But does this strategy persist when choosing between routes of equal length? Participants (N=28) viewed displays of rectangular courtyards in a VR headset. Starting at one corner, they walked around the perimeter to a goal post at the opposite corner. If participants reduce deviations from the beeline, they should walk the long side first. To test the influence of heading, half the participants walked to the start corner on an ‘on-ramp’ placed at different angles. To test the role of body orientation, the other half began trials standing at the corner facing different directions. We found that neither group preferred the long side first (ns). However, both groups chose routes that minimized turns from their start orientation (p<.001). We considered two possible explanations for this unexpected result. (1) Because participants can see the paths ahead, they depend on the spatial layout and discount other visual variables that normally correlate with the shorter path, like beeline deviations. (2) Because the courtyard is rectangular, they learn that route lengths are equal and similarly discount visual variables. To disassociate these alternatives, we repeated the experiment (n=30) except that participants began each trial behind a wall with two doors. Only the goal post, but not the courtyard, was visible when they selected routes. Both groups now preferred the long side (p<.001), even if it meant turning away from their start orientation. These findings suggest that visual information and knowledge of the environment play distinct roles in wayfinding behavior, and that the former dominates route selection.

Acknowledgements: NSF 1405001060; T32MH115895 (NIMH)