Anisotropic usage and encoding of three-dimensional space

Decades of work in the field of spatial cognition have revealed fundamental aspects of how navigation is structured in two-dimensional, horizontal spaces and have exposed remarkable insights into how the brain processes spatial information. However, animals and humans daily encounter navigation tasks in complex environments that include many vertical features. Until now, it has been only poorly understood how navigation behaviours are organised in three-dimensional spaces and how the additional vertical dimension is encoded. The aim of the present thesis was threefold: first, to review behavioural and neurophysiological literature involving three-dimensional navigation; second, to characterise spatial behaviours in environments with a vertical dimension; and third, to study neural substrates of three-dimensional navigation. Rats’ navigation abilities were tested in two types of environments: the pegboard, which is a vertical climbing wall, and the lattice maze, which is a three-dimensional climbing cube. Foraging and detour experiments were conducted in both environments and revealed that the use of three-dimensional space is horizontally biased. Such biases were manifested in three ways. Rats exhibited more horizontal than vertical movements, structured their foraging behaviours in a horizontal fashion and showed preference for initially-horizontal paths when offered two routes requiring equivalent effort to reach a goal. In order to understand how the vertical dimension is represented at a neural level, hippocampal place cells and entorhinal grid cells were recorded on the pegboard during foraging. In horizontal environments, place cells exhibit location-specific firing, whereas grid cells map space with multiple hexagonally arranged firing fields. The experiments with the pegboard revealed that both cell types represented the horizontal dimension better than the vertical dimension. Furthermore, grid cells formed vertically aligned stripes on the pegboard, suggesting that encoding of the vertical dimension was more strongly impaired in grid cells than by place cells. It is suggested that in rats, three-dimensional spaces are – both at a neural and a behavioural level – anisotropically represented.