Numenta Research Meeting - June 1, 2020

Jeff Hawkins brainstorms how sensorimotor models might be built up from purely sensory data, how this might fit into a cortical column, and the importance of magnocellular and parvocellular cells.


Here is the Link to the paper mentioned at the end:


Hmm, Moire patterns?
Where have I seen that proposed before?

Or maybe here?

1 Like

Maybe Retinal origin of orientation maps in visual cortex?

That actually would give a consistent model for both direction sensitive and grid cells.


I find the Moire inteference of two slightly different gride cells is very interesting. But about Theta grid, this paper suggests a very complex idea.
I believe we can maintain the theta frequency constant simply by using rotational scaling rule. It likes the phase changing, and does not need to change grid spacing.

Grid spacing and phasing and orientation are all variable.
There is much experimental evidence to support this.

@Bitking yes I see! I want to understand a little more about how ganglion cells (GC) connect to visual cortex.
In my current experiment, GC cells are designed as hexagonal grid. Parvocellular and Magnocellular provide On/Off cells, where on/off grid cells will have parameters for making moire inference effects in parvo and magno!
Normally, Moire inference is happened by any motion of objects or Animals motion!
In retina, motion is detected by magno!
That is why I do not understand how information from magno cellular will be used for creating moire effects?
Any idea?

@jhawkins could you please explain me how big is The ratio between small And large Field of view For magno connected to L4 and L2?
Is it biological plausible that magno has two different field of view? Thanks

The mentioned observations remember me of sitting in a standing train while seeing another train outside the window starting to move. At first, I think the train I’m sitting in starts to move (as all of the visible world outside moves) until I recognize that I don’t feel any acceleration.