1. The role of grid cells in spatial localisation and computation
- Author
-
Manson, D., O'Keefe, J., and Burgess, N.
- Subjects
570 - Abstract
From recording single neurons in rodents and primates, and observing the effects of lesions in humans, it is well recognised that the mammalian entorhinal cortex, and the broader hippocampal formation, perform cognitive tasks relating to both space and memory. In this work I propose a number of specific roles for grid cells and experimentally test aspects of their function. I also give an overview of a piece of software developed for streamlining the process of researching grid cells. The main theoretical chapter discusses how grid cells might be used to perform two-dimensional vector calculations. Simulations show that computing the vector to a goal can be performed accurately, and in a short time window, using a network topology that might conceivably be initialized during development. It is then suggested that such a network might also support vector addition. A short experimental chapter follows, in which some preliminary data is presented to test this hypothesis. The main experimental chapter examines how grid cells respond to variations in the level of spatial information present in the environment. In the data presented, grid cells appear to be more stably anchored in a cue rich arena, as compared to an otherwise identical cue poor arena. Differences in grid scale are examined in relation to the inaccuracy of self-localisation, as inferred from spatial tuning. The final chapter discusses the above findings further and makes suggestions as to their broader significance and the significance of other related recent findings.
- Published
- 2017