Comments on: Space perception is hard-wired http://charbonniers.org/2010/08/10/space-perception-is-hard-wired/ A blog on consciousness by Janet Kwasniak Wed, 23 May 2012 06:20:09 +0000 http://wordpress.org/?v=2.2.1 By: Lucas Wilkins http://charbonniers.org/2010/08/10/space-perception-is-hard-wired/#comment-37669 Lucas Wilkins Wed, 03 Aug 2011 06:14:07 +0000 http://charbonniers.org/2010/08/10/space-perception-is-hard-wired/#comment-37669 <p>I don't know how this means 3D perception is hard-wired. Even if they mapped a 3D space - which they don't - grid cells definitely don't work like actual space. To see why this must be the case (experiments show this too), think just how many you would need to represent everywhere a rat could possibly travel. A rat that travels a sufficient distance would have to "go off its own grid" if the grid worked exactly like euclidean. Rats do solve this problem, but it means that its not quite like euclidean space - still leaving the question of how a coherent euclidean space comes to be.</p> JK: Thank you very much for your comment. Three points - first, the quote did not mention border cells. These are cells that fire when activity moves to grid cells near the edge of the grid. The idea (not proven but generally accepted, I think) is that there is a store of 'maps' in the hippocampus. They are various scales and so on. A single map is selected and "superimposed on the grid". When movement comes to the edge of the map then another is placed on the grid. Second point - I am not sure that the grid gives 3 dimensions. I would guess that it gives 2 plus 'landmarks' that give some depth to the perception. I presume that object perception has more to do with 3D perception then the navigation grid. Third point - I think the hard-wired support comes from the navigation system being ready to actually do some navigating before the rat can have created a navigation system by learning. So maybe I over-stated the case in calling it 3D space rather then ability to navigate - thanks for the correction. On the other hand, my belief that something as essential as our concept of space would be hard-wired has not been shaken by anything I have read to date. If we learn 3D space, we must learn it very, very quickly because we need it in place before we can learn most other things. I don’t know how this means 3D perception is hard-wired. Even if they mapped a 3D space - which they don’t - grid cells definitely don’t work like actual space. To see why this must be the case (experiments show this too), think just how many you would need to represent everywhere a rat could possibly travel. A rat that travels a sufficient distance would have to “go off its own grid” if the grid worked exactly like euclidean. Rats do solve this problem, but it means that its not quite like euclidean space - still leaving the question of how a coherent euclidean space comes to be.

JK: Thank you very much for your comment. Three points - first, the quote did not mention border cells. These are cells that fire when activity moves to grid cells near the edge of the grid. The idea (not proven but generally accepted, I think) is that there is a store of ‘maps’ in the hippocampus. They are various scales and so on. A single map is selected and “superimposed on the grid”. When movement comes to the edge of the map then another is placed on the grid. Second point - I am not sure that the grid gives 3 dimensions. I would guess that it gives 2 plus ‘landmarks’ that give some depth to the perception. I presume that object perception has more to do with 3D perception then the navigation grid. Third point - I think the hard-wired support comes from the navigation system being ready to actually do some navigating before the rat can have created a navigation system by learning. So maybe I over-stated the case in calling it 3D space rather then ability to navigate - thanks for the correction. On the other hand, my belief that something as essential as our concept of space would be hard-wired has not been shaken by anything I have read to date. If we learn 3D space, we must learn it very, very quickly because we need it in place before we can learn most other things.

]]>