The grid cells are found in the entorhinal cortex, which is basically a part of the neocortex. The place cells found in the hippocampus receive input from the EC. The neocortex/hippocampus are not structures found in non-mammals, but I wouldn't be surprised if some other kind of spatial metric mechanism that was used for navigating (and whatnot) was found in reptiles/amphibians/birds too.
Your grid cells map to a 2D space like the game Asteroids. Instead of trying to represent an "infinite" space on an endless plane, they map to a finite space that has no edges. Like an ant walking on a donut, the modules of grid cells treadmill around closed loops (x and y axes).
@@robertmihalinac your cells are not literally moving around the torus. You can think of it being similar to how retinal cells map your vision; they do not accurately represent reality but are how the brain "thinks" about reality. Grids cells map to the torus because it is an efficient way to think about "infinite" space. Almost all of the cells in your body don't need to know where they are in space; they only need to be told what to do by a system that does know where you are and where you want to go (the brain). There is the possibility that your grid cells do not only map space but are involved in other forms of memories. Using a torus to represent time for example is less intuitive than say a circle/ring. A 1D track would also make more sense to map to a circle since you don't use the second dimension. As he mentioned in his talk, head direction cells map to a circle. All of those are single dimensions (time, position, or rotation) so they only need a 1D circle instead of a 2D circle of circles (torus). If other cells use loops like that, I'm not sure.
@@SirFake314 That's what I understood from his talk, but what is not clear to me (both from his talk and actually from reading the paper itself) is, what is the significance of this transformation other than just a mathematical game of representation. To me this torus model of a patch does not, for example, helps the animal or the researcher to interpret the environment. Maybe in the future, the research will show significance of this mathematical transformation, but for now it does not help, as I see it. No?
So in fact what is the significance of this discovery? It looks to me that it has no physical meaning rather than just a mathematical exercise. @4:00 you explain that a single field (patch) can be imagined as folded into torus but in reality the neurons forming the field are not folded. Yes, you can see the signal of neurons firing stronger somewhat that way, maybe, what is that what you created this model for?
Are we hoping to see (in the future research, i.e. predicting that) the neurons that a part of the grid cells system to form a physical torus in layer 2 of entorhinal cortex? So at each scale/module, the cells would from a physical torus in layer 2, and you could see, I don't know, say 10 of those of those toruses for 10 different scales.