In this way the connectivity between place cells, normally identified with the CA3 recurrent connections, is updated to reflect the relative position of their fields in space and can be used to test or infer potential routes [41]. A weakness of this approach though is that the animal must thoroughly explore an unfamiliar environment before it can navigate effectively; specifically

the network cannot identify routes that traverse unvisited sections of space. Thus, the system cannot exploit potential shortcuts when changes to the environment occur. Conversely, Pifithrin-�� manufacturer it does mean that the network learns about the relative accessibility of points in known space, allowing the shortest route to be selected and Epigenetics inhibitor dead-ends avoided. Muller

et al.’s [41] model of the CA3 place cell network as a resistive grid took advantage of this effect to determine the shortest viable route to a goal. An alternative proposal is that navigation could be affected by moving to maximise the similarity between the place cell representation of the goal and current location. However, such an approach is only successful when travelling between points separated by less than the diameter of the largest place field. Beyond this distance the overlap between representations will be flat affording no gradient to follow. Although the size of the largest place fields is unclear, recordings made from the ventral hippocampus of rats suggests that fields

might exceed 10 m in diameter [43]; though larger than a typical experimental room this is much smaller than the range of wild rats which can be hundreds of metres [44]. By contrast to place cells, the spatial Non-specific serine/threonine protein kinase activity of grid cells is inherently regular, spanning the available space with repetitive firing patterns [19] that may provide a spatial metric (though see [45]). In the medial entorhinal cortex medial entorhinal cortex (mEC) grid cells are known to exist in functional modules, the cells in each module having grid-like firing patterns that are effectively translations of one another; sharing the same orientation and scale but having different offsets relative to the environment 19, 46, 47 and 48] (Figure 1b). Modules are distributed along the dorso-ventral axis of the mEC with those at more ventral locations tending to be of larger scale such that the size of the peaks in the grid firing pattern and the distance between them is increased 19, 23 and 47]. Analysis of the grid code suggests that it provides an extremely efficient representation of self-location; modules of different scales behaving similarly to the registers in a residue number system such that capacity of the network greatly exceeds the scale of the largest grid 49 and 50].