A Study of Integrating Spatial Relationships Across Separately Learned Routes: Navigating Segmented Paths in Copetown

Whether humans form cognitive maps is controversial. One view is that the ability to generate detours and shortcuts demonstrates retention of direction and distance information integrated within a common frame of reference. Another view is that spatial representations are not Euclidean, given findings of biases, distortions, and lack of recognition of impossible spaces in VR. A compromise comes from an individual-differences perspective, suggesting that some people in some environments may be able to use a common reference frame to integrate across routes. We created Virtual Copetown to examine within-route knowledge, integration between routes with experienced connections, and integration between routes requiring inference. We also examined cognitive correlates of the ability to make these judgments. Our results indicated that some people were more accurate across all kinds of pointing judgments, crucially including the ability to judge inferred relations. Their accuracy was not related to the number of node-path-segments separating locations. Other people were less accurate overall, and less accurate for between-route relations than within-route relations, had worse mapping scores, and did decline in accuracy as the number of nodes separating locations became high. Variability was related to self-reports of navigation strategy use.