Evidence for flexible navigation strategies during spatial learning involving path choices

In a classic study in 1946, Tolman et al. found that rodents chose the optimal path to a hidden goal location when given the option to take a shortcut. Subsequent studies, though, found mixed results, including a previous study in humans showing non-optimal response strategies. Here, we tested human participants in a virtual sunburst maze identical to that employed by Tolman. Across three different experiments, we found that participants consistently employed non-optimal response strategies involving either following the encoded route or using the light as a beacon cue. Adding distal boundary cues (Experiment 2) or enhancing idiothetic cues (Experiment 3) led to the same overall results. Our findings suggest participants’ search strategies are often suboptimal when given the opportunity to spontaneously take a short cut.