Spatial representation by a combination of grid cells and place cells.

Animals are capable of navigating through an environment. This requires them to recognise, remember and relate positions. Spatial representation is one of the main tasks during navigation. The brain seems to have a world centric positioning system such that we remember positions of objects in relation to a reference frame. Such spatial representation is believed to have been constructed in the hippocampus and related brain areas. Place cells, head direction cells, grid cells and border cells seem to transform vestibular information to spatial information in the brain. However, when psychological studies reveal how these areas are connected together, the process of transforming vestibular information to the kind of representation seen in place cells is still in question. In this thesis, we develop a hypothetical model, consisting of head direction cells, grid cells, sensory cells, border cells, place cells and connections between each. Our aim here is to create a biologically feasible spatial representation of the environment with activity patterns of neurons in the model similar to firing patterns of place cells in the hippocampus. With this model, we simulate a subset of the components in order to develop a spatial representation. This subset consists of a single dimension head direction representation, periodic two dimension grid cells representation, and finally a monotonic two dimension place cell representation. We offer two approaches for connections between grid cells and place cells, hardwired and learnt connections. The hardwiring approach provides us an understanding of how grid cell firing can be transformed into place cell representation. The learning approach illustrates a biologically plausible method in the development of spatial representation. Our results show that it is possible to achieve spatial representation by use of Bienenstock, Cooper and Munro (BCM) learning method which has a dynamic threshold. To further evaluate our model and determine its practical application we embed it in a mobile robot. A robotic implementation provides an opportunity for us to evaluate the model under the presence of noise both in the internal information and from the environment. Although our robotic experiments are limited, they demonstrate that our model provides a fundamental biologically plausible infrastructure for robotic navigation