Implications of Reference Frames on memory and aging

Science
1

Taking in mind reference frames and cortical columns, even with 150,000 cortical columns remembering hundreds of reference frames, it still represents much less than what we actually remember (which is something Hawkins has clearly stated).

Are there any theories as to why that happens? if it indeed is the biological limit, then does that mean a lot of things (reference frames) we learn are actually transferable to other domains of our life?

Also, if indeed 150K cortical columns are enough to perform and remember everything we currently know, then why is it that as we age, we able to learn a lesser amount of new things in comparison to as we are young? wouldn’t all the reference frames we gained in the youth be transferable to new things any old person learns?

Lastly, if a cortical column retains its reference frames forever, then wouldn’t it mean that there would be no space for newer reference frames learnt by our brain?

Could it be somehow related to what we call “muscle memory”? like if someone hasn’t played the piano in 10 years, but come near it - they might be able to recall certain parts of songs, some notes, their memories of piano, etc. Does that indicate that reference frames are not lost but just become less active unless something triggers it? but that would initiate the same argument that there aren’t enough columns to store every reference frame we learn throughout life.

I am pretty confused. can anyone shed some light on this?

2

I don’t recall reading a limit of hundreds of reference frames, but a limit of hundreds of objects. From TM, I do know that the numbers of cells involved in a CC have an astronomical capacity WRT numbers of transitions between representations that can be learned, so my sense is there is a much greater capacity to learn reference frames and behaviors that are related to a given object.

In any case, hundreds per unit is not a very constraining limit when you consider how semantics work. Analogous to how a tiny 1024x32 array of cells in a typical TM layer can represent an astronomical number of distinct SDRs, an array of 150,000 cortical columns could potentially represent an astronomical number of distinct concepts.

2 Likes