The representation of space in mammals

Mathis, Alexander (2012): The representation of space in mammals: resolution of stochastic place and grid codes. Dissertation, LMU München: Graduate School of Systemic Neurosciences (GSN)

Abstract

Animals require cognitive maps for efficiently navigating in their natural habitat. Cognitive maps are a neuronal representation of their outside world. In mammals, place cells and grid cells have been implicated to form the basis of these neuronal representations. Place cells are active at one particular location in an environment and grid cells at multiple locations of the external world that are arranged in a hexagonal lattice.

Abstract

As such, these cell types encode space in qualitatively different ways. Whereas the firing of one place cell is indicative of the animal's current location, the firing of one grid cell suggests that the animal is at any of the lattice's nodes. Thus, a population of place cells with varying parameters (place code) is required to exhaustively and uniquely represent an environment. Similarly, for grid cells a population with diverse encoding parameters (grid code) is needed. Place cells indeed have varying parameters: different cells are active at different locations, and the active locations have different sizes. Also, the hexagonal lattices of grid cells differ: they are spatially shifted, have different distances between the nodes and the sizes of the nodes vary in their magnitude. Hence, grid codes and place codes depend on multiple parameters, but what is the effect of these on the representation of space that they provide?

Abstract

In this thesis, we study, which parameters are key for an accurate representation of space by place and grid codes, respectively. Furthermore, we investigate whether place and grid codes provide a qualitatively different spatial resolution.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Keywords: Grid Cell, Entorhinal Cortex, Place Cell, Hippocampus, Population Coding, Fisher Information, Maximum Likelihood Estimator, Navigation, Optimal Coding Hypothesis, Noise Correlations
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 570 Biowissenschaften, Biologie
• Fakultäten: Graduate School of Systemic Neurosciences (GSN)
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 30. Juli 2012
• 1. Bericht­erstatter:in: Herz, Andreas