El Duvelle NeuroDec 10, 2023

I love #GridCells but I have to say their influence for any neural process is probably highly over-inflated.

When you actually record from #MEC you see there are so many non-grid cells… it’s really a huge contrast to #Hippocampus where most of the cells are going to be #PlaceCells either in the current environment or another.

Show threadmatt nolanDec 10, 2023
@elduvelle_neuro You're right, non-grid cells within MEC are massively overlooked and could be as important - definitely worth spending time on. Our most recent work (preprint here: https://www.biorxiv.org/content/10.1101/2023.05.12.540491v1, updates soon) suggests grid cell roles are restricted to path integration (which follows Bruce McNaughton's early suggestions, and nice previous work from Hannah Monyer's lab).
Show threadmatt nolanDec 10, 2023
@elduvelle_neuro Also relevant, during a location memory task we many non-grid cells show ramping activity (https://doi.org/10.1016/j.cub.2022.08.050), also suggesting (to us at least) that there's lots more going on in addition to grid activity.
Show threadEl Duvelle NeuroDec 10, 2023

@mattnolan thanks for the input! So I’ve always wondered why people think that grid cells “perform” path integration instead of path integration information (distance and direction to a start point) being used (as an input) in grid cell firing?

Also what do you make of the results on grid cells in 3D where the grid completely breaks down? Are rats (or bats) incapable of doing 3D path integration? Or is the grid not actually that important for grid cells, more of a side-effect when looking at them in a 2D flat surface with very regular geometric features?

Show threadmatt nolanDec 10, 2023
@elduvelle_neuro Re. 3D. I guess whether grids are important needs behavioural experiments in which z-xis position is relevant to solving a task. It could also be interesting to take existing 2D CAN models and ask what else they need to account for the 3D data. E.g. Is it just a case of changing the sensory inputs that provide an anchor, or is additional information about velocity on the z-axis required? Maybe someone has done this already?
Show threadEl Duvelle Neuro

@mattnolan Good points… still, when rats do foraging in 2D, they can have very nice grid cells, but as soon as they move on to 3D, still foraging, the grids break down… 🤔

Roddy (from Grieves et al 2021 says that he’s not aware of any continuous attractor model of grid cells that has been extended to 3D but there should be 2 options: 1) either the grids can integrate movements in the z-axis and that should give rise to columns in 3D, or 2) they can’t and the grid should completely break down in 3D. However, the data showed mostly randomly-located spatial fields that were more spatially-stable than chance, so neither of those 🤔 (similar findings by Ginosar et al., 2021 but for MEC cells that were not necessarily grid cells). But maybe there are now more recent CAN models that we are not aware of!

This reminds me that Gily has a new review paper on the role of #GridCells for #Navigation that I #Need2Read: Ginosar et al., 2023

Irregular distribution of grid cell firing fields in rats exploring a 3D volumetric space - Nature Neuroscience

Grieves et al. show that when rats explore a 3D space, grid cells in the entorhinal cortex exchange their usual spatially regular firing patterns for more irregular ones, suggesting that 3D space is mapped differently than previously thought.

Nature
Dec 10, 2023 at 8:56pmWeb
Show threadmatt nolanDec 11, 2023
@elduvelle_neuro I wonder if knowing more about the 'anchoring' inputs would be useful here? When we built CAN models (e.g. Pastoll et al. 2013) we assumed place cell inputs as the anchor, but any spatially stable signal will do, e.g. Lisa Giocomo's group suggested border cells, but some kind of a 'view cell' could also be sufficient. Depending on the z-dependence of the anchor (place, border, view, or whatever) possibly many different 3D grid structures are feasible?
Show threadEl Duvelle NeuroDec 11, 2023
@mattnolan yes, definitely!