http://www.scientificamerican.com/article.cfm?id=marijuana-reveals-memory-mechanism

I wonder what the implications are for brain preservation and whole brain emulation? If glial cells are important, then saving and emulating the neurons alone probably won't be enough.

New Comment
11 comments, sorted by Click to highlight new comments since:

I would be surprised if the selection pressure that led to intelligence hadn't taken advantage of absolutely everything it possibly could.

[-][anonymous]120

If it's just important for working memory, then it would seem like they could get "you" back without them - assuming that working memory implementation across humans is mostly similar. You'd need analogs of glial cells, but not necessarily the exact

[-]gwern120

Yes. Just because something is necessary for the brain to operate doesn't mean specific versions need to be preserved rather than some generic version. It is necessary for you to have a skull to think and live, but I don't think many people expect the individual bumps and indentations to need to be preserved for a faithful upload...

On the other hand, people probably have slightly different versions of glial cells, and which model of glial cells you have might have some effect on your personality and intelligence.

Actually, it's probably not just one model per person. I've been told that there's more than one sort of ATP (copying errors and variations), even inside one person, and if ATP isn't universally identical, then you can't expect anything else to be.

I know next to nothing of biology, but I would naïvely expect the structure of the ATP, ADP, AMP, etc. to be fixed across all organisms with mitochondria. Shouldn't copying errors or variations that produce something other than ATP in place of ATP kill any eukaryote, let alone a human? Perhaps you mean variations to ATP synthase?

I know little biology, too. I can check back with the person who told me that-- the way I understood it was that there are slightly different versions of ATP, some of which are more efficient, but all of which work.

The WP article doesn't mention anything of the kind, and when I studied freshman biochemistry nothing like that was mentioned, either. OTOH there's a lot of variations in the (vastly bigger, more complex and more diverse and numerous) organic molecules that work with ATP.

If possible, better ones. Working memory seems to be a human weak point.

When I took neurobiology, we learned that glial cells aided in, well, pretty much everything. And I thought we specifically addressed memory too. But I could be misremembering.

This should theoretically raise the amount of computing power required to create AI through brain emulation

I've been of the opinion for a while that we should be planning on modelling glia just to be on the safe side.