Recent advances in our understanding of spatial coding in the entorhinal cortex and hippocampus have made this network a powerful model system for neural computation in high-end cortices. With our discovery of grid cells as the brain´s metric for space in 2005, spatial navigation became one of the first non-sensory cognitive functions of the brain to be understood in reasonable mechanistic detail. Grid cells are cells that have spatially localized firing fields tiling environments with a periodic hexagonal firing pattern in a manner that enables accurate self-localization. Much has been learnt during the past decade about properties of grid cells and how they are organized in anatomical space. We have shown, for example, that grid cells cluster into a small number of discrete layer-spanning modules with semi-independent firing properties. However, the fundamental principles underlying the clustered anatomical arrangement of grid cells remains poorly understood. Limited anatomical sampling has obscured whether the grid system operates as a unified system or a conglomerate of independent modules, and it has remained unclear whether modular organization is unique to grid cells or also applies to other cell types of the spatial representation system.
The present project aimed to identify some of the principles determining how grid cells self-organize into functional clusters. By adapting state-of-the-art multisite recording technology, we have increased the number of cells recorded in individual animals by an order of magnitude compared to data obtained only a few years ago. The project has been organized around three tasks. In the first, which was completed by the end of the first project year, as planned, we showed, based on data collected before the startup of the project, that modular organization is present only in grid cells and not in head direction cells. Head direction cells - cells that fire in relation to the animal´s direction of movement - showed topographical organization along the dorsoventral axis of the medial entorhinal cortex, with sharper directional tuning at the dorsal pole than the ventral pole. However, the decrease in directional tuning was gradual and very much unlike the step-like scale change observed in simultaneously recorded grid cells. In the second task, we have tested whether the increase in grid scale over modues follows a similar organization, with a similar scale factor, in different species - rats and mice. Our observations suggest that despite the slight difference in scale across rodent species, the increase in grid scale between successive modules is constant, which would be consistent with a genetic and experience-independent organization of the network. The third and final project aims to determine, with recordings of gamma-frequency oscillations in firing patterns of cells from different modules, whether grid modules undergo patterns of synchronization over time and in relation to behavior of the animal. We have recorded synchronization of network activity across several weeks of postnatal life in mice, seeing considerable coherence in network activity during the first weeks of life. These coherent patterns may be instrumental in wiring networks for grid cells. Taken together, the results from these three tasks will help us understand how a high-end brain function - the mapping of self-position - is computed in interconnected parallel networks of the cerebral cortex.
Recent advances in our understanding of spatial coding in the entorhinal cortex and hippocampus have made this network a powerful model system for neural computation in high-end cortices. With our discovery of grid cells as the brain´s metric for space in 2005, spatial navigation became one of the first non-sensory ?cognitive? functions of the brain to be understood in reasonable mechanistic detail. Grid cells have spatially localized firing fields that tile environments with a periodic hexagonal firing p attern in a manner that enables accurate self-localization. However, the organization of the grid network remains poorly understood. Limited anatomical sampling has obscured whether the grid system operates as a unified system or a conglomerate of indepen dent modules. By adapting state-of-the-art multisite recording technology, we are now able to increase the number of cells recorded in individual animals by an order of magnitude. Recent data from more than 150 grid cells in individual rats show that the cells cluster into a small number of layer-spanning anatomically-overlapping modules with distinct geometric features. The discreteness of the grid modules, and their apparent autonomy, differ from the graded topography of many sensory maps, suggesting th at a hitherto unknown set of algorithms underlie the formation of grid modularity. The aim of the present project is to identify these algorithms. We shall identify the factors controlling the scale relationship between modules, establish whether interact ions between grid modules are dynamic and what mechanisms control the dynamics, and determine whether modularity is a unique property of grid cells or involves all functional cell types of MEC. The discovery of modularity in the grid-cell system has enabl ed analysis of a system of interconnected parallel-computing networks with a realistic potential for revealing fundamental principles of neuronal operation on abstract data.