Effects of topological instability on spatial problem solving in a virtual desktop maze task

Real-world environments are unpredictable and dynamic. However, most spatial learning and problem-solving research uses stable environments. Our study examined how humans solved a desktop virtual maze task under conditions of topological instability, where the structure of the environment changed on every trial. We developed two virtual desktop environments, one with a stable structure and one with a reconfigurable structure. In the stable maze, the environmental structure remained the same across trials. In the reconfigurable maze, the internal organization of the maze changed after each trial based on the participants’ route selection. Participants ( n = 132) were instructed to find and collect coins hidden in two of four chests. Both groups showed significant improvement across trials. A trend toward slower acquisition of Optimal behavior—collecting both coins without approaching unnecessary compartments—in the reconfigurable group did not survive multiple comparison correction. The reconfigurable group showed slower reductions in both path length (IRR = 1.10) and trial duration (ratio = 1.26) across trials. After controlling for path length, the difference in duration was reduced but remained significant (ratio = 1.17). Topological instability imposed a cost on task-solving efficiency, measured by path length and trial duration. It could not be determined whether it also slowed task-rule acquisition in the present sample, and replication with a larger sample is needed. This study provides a novel tool for investigating problem-solving in changing environments. With proper adaptation, this tool could be used in developmental, aging, and clinical research.