Human place learning in a virtual Morris water task: some important constraints on the flexibility of place navigation.

We investigated human place learning in a computerized version of the Morris water task (VMWT) under comparable conditions to those employed by Sutherland et al. (Sutherland et al. Psychobiology, 1987;15:48-57) with rats. Participants viewed a computer-generated environment and were trained to locate a hidden goal in one half of a circular pool (region 1). The opportunity to navigate in and view cues from region 2 was systematically varied during training. Participants were then started from region 2 to assess transfer. Accurate transfer performance was dependent upon prior experience viewing distal cues from region 2 while on a trajectory to the goal, a finding we interpret as inconsistent with the automatic formation and modification of a cognitive map (O'Keefe J, Nadel L. The Hippocampus as a cognitive map. Oxford, UK: Clarendon Press, 1978). Additionally, the transfer data reported here closely match the data obtained by Sutherland et al. with rats suggesting some generality in the principles involved in place learning.