A MODEL FOR PLACE NAVIGATION IN MAMMALS

Mammals are able to process the location-based (site-dependent) inform ation provided by the spatial layout of landmarks to return to an inco nspicuous place (the goal) whose location has been memorized exocentri cally by reference to the landmarks. We have developed an algorithm ac counting for this place navigation ability. This algorithm is based on the angular information provided by the bearings and angular sizes of landmarks, rather than on metric distances. It is implemented by a sy stem that combines elementary computing procedures and two types of si mple associative memory network. One type of network enables the anima l to learn how the environmental perspective changes when it makes a o ne-step translation in any direction, whereas the other type enables t he animal to learn how the current panorama is oriented with respect t o an overall reference direction. This system highlights the fundament al role of exploration of the environment prior to place navigation, a nd assigns specific functions to hippocampal place cells and postsubic ular head-direction cells.