Identifying location by dead reckoning and external cues

Golden hamsters can orient towards specific points in their environment using location- based visual cues and/or dead reckoning based on vestibular and proprioceptive signals. The relative weight of these different kinds of information was investigated in an apparatus consisting of three identical, square compartments joined by tunnels, with the subject's own nest box at one end. Each compartment contained a feeding site and a weak light spot, the relation between the feeding site and the light spot being different in the three compartments. The animals were trained to hoard food in succession from the three feeding locations, in darkness. During test trials, the light spots were either suppressed or moved to new locations, thus being set in conflict with other kinds of spatial information. In the majority of trials, the subjects proceeded fairly directly to the feeding places, independently of the presence and position of the light spots. This performance may be explained through rote motor learning, which may be initiated with respect to tactile cues previously associated to the goal. However, more flexible, indirect trajectories towards the goal suggest that the subjects kept track of their location within the test space and therefore depended simultaneously on dead reckoning and a map. A control experiment in which the goals were moved along with the visual cues excluded the use of olfactory cues from the food source and confirmed the role of dead reckoning.