Spatial learning in a three-dimensional maze

Abstract The ability of rats, Rattus norvegicus to orient spatially was tested in three-dimensional structures using a place learning task procedure. Mazes were built from a network of bars allowing the rats to move easily horizontally as well as vertically. Three groups of hooded rats were trained in plane mazes oriented horizontally, vertically or tilted at 45°. Rats trained in the vertical and the tilted 45° mazes were able to focus on the goal location earlier than the horizontally trained group suggesting that orientation is easier in a maze with a vertical component. Naive rats were then tested for a place learning task in a three-dimensional cubic maze. Two transfer tests without food reinforcement were conducted in the middle and the late part of the training to see if the rats would search around the previous spatial position of the food. In the first transfer, the rats were able to locate precisely only the height of their goal. Only during the second transfer were they also able to locate its horizontal coordinates. This indicates that the animals learned the position of the goal in two steps and it provides evidence for a priority given to the vertical dimension of space, when relevant for orientation. This could be related to the vertical dimension of space being parallel to the force of gravity. On the one hand, vertical movements could produce stronger kinaesthetic cues than horizontal movements; on the other hand, rapidly acquiring a high precision along the vertical dimension allows the animals to reduce the number of vertical movements and therefore the energetic cost of the search.