I don't completely understand the advantage of having grid cells as a navigation aid. I know they are called the 'GPS' of the brain. I also read that they can encode locations in a large space, and I saw that a paper came out recently showing how to do 'path integration' with grid cells. But usually in the brain, similar patterns code for similar things. I don't think this is true of grid cells. Is it true that if - lets say - 3 grid cells encode a particular location, that a similar pattern (maybe sharing 2 of the 3 cells), will encode a nearby location?

  • 1
    $\begingroup$ Could you add an APA citation to the paper you mention? $\endgroup$
    – Steven Jeuris
    Commented Oct 11, 2023 at 11:20
  • $\begingroup$ Grid cells are not found in the hippocampus but in the entorhinal cortex - could you update the title accordingly? $\endgroup$
    – leonos
    Commented Oct 19, 2023 at 23:35

1 Answer 1


First, a comment: Place cells, not so much grid cells, are thought to reflect a positional signal in the brain. Each place cell specifically activates in 1 or a few locations in a given environment. For example, place cells A, B, C will be active in location 1 and place cells C, D, E will be active in location 2. Note that a place cell can be active in multiple locations (called 'place fields'), like C in our example, which is still fine as the ensemble of place cells firing in that location will still be unique.

similar patterns code for similar things

Indeed, two similar subsets of active place cells, say ABG vs ABE, are likely to reflect nearby locations, as the regions of activity of place cells are gaussian-like and nearby place fields overlap.

Then, Grid cells, unlike place cells, have many, repeating active regions of interest ('grid fields') and as such it is more difficult to retrieve precise location information from them. The combination of active grid cells XYZ, unlike for place cells, will occur many times in a given environment, due to the repeating nature of their firing. However, with enough co-recorded grid cells, it is probably possible to extract accurate position information from their population firing, especially since each grid field might have a different firing rate, but I'm not aware of any experimental work doing that (the most related seems to be Gardner et al., 2022 but they do not decode 2D position). With that in mind, it should still be the case that similar patterns of firing will correspond to similar locations, but you will need to observe the activity of a lot more cells for this to be true, compared to place cells.

Additional material:

PS: Grid cells are not found in the hippocampus but in the (monosynaptically connected) Entorhinal Cortex; place cells are the ones found in the hippocampus.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.