Persistent and stable biases in spatial learning mechanisms predict navigational style

A wealth of evidence in rodents and humans supports the central roles of two learning systems--hippocampal place learning and striatal response learning--in the formation of spatial representations to support navigation. Individual differences in the ways that these mechanisms are engaged during initial encoding and subsequent navigation may provide a powerful framework for explaining the wide range of variability found in the strategies and solutions that make up human navigational styles. Previous work has revealed that activation in the hippocampal and striatal networks during learning could predict navigational style. Here, we used functional magnetic resonance imaging to investigate the relative activations in these systems during both initial encoding and the act of dynamic navigation in a learned environment. Participants learned a virtual environment and were tested on subsequent navigation to targets within the environment. We observed that a given individual had a consistent balance of memory system engagement across both initial encoding and subsequent navigation, a balance that successfully predicted the participants' tendencies to use novel shortcuts versus familiar paths during dynamic navigation. This was further supported by the observation that the activation during subsequent retrieval was not dependent on the type of solution used on a given trial. Taken together, our results suggest a model in which the place- and response-learning systems are present in parallel to support a variety of navigational behaviors, but stable biases in the engagement of these systems influence what solutions might be available for any given individual.