Know your neuron: Grid cells

Know your neuron is a mini-series here at the Fan Club where we look at some of the curious response characteristics of neurons discovered through electrophysiology. We will work our way through the simple properties of cells that respond to spatial locations or head direction and on to increasingly abstract cells, which encode the identity of individuals or perhaps even the relation between actions of others and those of the self.

Given how straight-forward their response properties are, it is curious that grid cells were only discovered relatively recently (Hafting et al, 2005). According to one of my professors, this is simply because researchers didn’t put their rats in sufficiently large boxes to see the pattern previously, but who knows… The response properties of grid cells are perhaps best illustrated in a figure (this is from the excellent Scholarpedia article on Grid cells, written by two of their discoverers).

In this figure, the x and y coordinates correspond to the spatial location of a rat, which is running around freely inside a large box. The black lines in the left figure shows how this particular rat explored the box in a fairly haphazard manner. However, an electrode inserted in the rat’s subcortex picks up a signal that is anything but chaotic: the responses of said neuron are given as red dots in the left figure, while the right figure gives the firing rate distribution (ranging from blue for silent and red for the peak rate of responding). Although the rat is running about randomly, this neuron is responding in a grid, seemingly coming on an off in response to the animal’s spatial location.

This response pattern is similar regardless of the exact box, so the pattern does not reflect learning of the spatial layout of this box as such. Indeed, when the light is turned out, the grid cell continues to respond as before while the rat runs around in the dark (Hafting et al, 2005 – this is not terribly impressive when you consider how little rats use vision, however). Yet, at least some grid cells showed a response pattern that rotated 90 degrees with a corresponding shift in an external visual cue, so although the grid cells do not depend on the external environment, they do seem to take cues from it when available.

Grid cells are found in Entorhinal cortex of the rat, a region which provides the main input to the Hippocampus, long known as a centre for memory consolidation and spatial navigation. Thus, these cells are a prime candidate as building blocks for more complex spatial cells that are observed in the Hippocampus proper. Different grid cells in Entorhinal cortex appear to have partially overlapping response fields, so that a given location might produce a response in some grid cells but not others. There is also a topographical structure, where the size of the grid increases as one moves from the dorsomedial to the ventrolateral side of the area. Incidentally, this pattern matches a spatial scale increase of more complex cells in the Hippocampus proper.

The obvious implication is that by combining the output of a number of partially-overlapping grid cells, perhaps a cell with more complex spatial properties could be built. The next issue of Know your neuron is about one category of such neurons, which appear to respond only to a specific spatial location.

References
Hafting, T., Fyhn, M., Molden, S., Moser, M., & Moser, E.I. (2005). Microstructure of a spatial map in the entorhinal cortex. Nature, 436, 801-806. DOI: 10.1038/nature03721

Moser, E., & Moser, M.-B. (2007). Grid Cells. Scholarpedia, 2, 3394.