Abstract
Like honeybees (Dyer, 1991), humans do not seem to build a metric “cognitive map” from path integration, but rely on visual landmarks to take novel short-cuts in a known environment (Foo et al., 2001). In the present experiment we investigate dependence on path integration and local landmarks when learning the layout of a new environment. Do people, like ants (Collett, et al., 1999), first depend on path integration to learn the locations of landmarks, and then shift to rely on landmarks? Or do they rely on landmarks from the beginning? Participants walked in a 40 × 40 ft virtual environment while head position and orientation were measured with a sonic/inertial tracker. Training: participants learned two legs of a triangle with feedback: the paths from Home to A and Home to B. A configuration of colored posts surrounded the A location, with another cluster placed on the path between H and B. Test: participants walked the learned legs or the novel short-cut between A and B without feedback. On catch trials, one cluster of posts were translated by 2m, so as to probe reliance on landmarks denoting a place (H-A), a known route (H-B), a route home (B-H), or a short-cut (B-A). Catch trials were initiated immediately after the onset of learning to track changes in the reliance on landmarks vs. path integration. Preliminary results suggest that participants' early use of path integration decreases during learning so they come to rely on landmarks that are associated with targets and routes. Thus, like ants, active navigation becomes dependent on local landmarks with increased familiarity with the environment.
Supported by NSF LIS IRI-9720327