What Can a Cell Remember? | Hacker News Reader

What Can a Cell Remember?

quantamagazine.org

83 pointschapulin13 comments

I think the memory of slime molds (which are referenced in the article) can go further than merely relying on paths of slime they leave behind.

Recently, I acquired a sample of Physarium polycephalum and have been keeping it as a sort of "pet", if one can call it that. For those who don't know, slime molds like Physarium are actually considered a single-celled organism, at least when it's in its plasmodial phase. People typically feed them oat flakes because that's what they seem to love most, though I started trying some other foods to see what my little slime mold would be willing to eat. Carby things like pieces of bread, etc. The funny thing is that it seemed to really like those other foods, even multiple feedings in a row, but would then spontaneously refuse to respond to those same foods again. I've heard some anecdotes suggesting I'm not the only one to witness this. It really does seem like the slime mold is "remembering" at a level that may go beyond slime trails.

I heard that if you eat the same thing for a long time, you can get sick because you aren't getting certain micronutrients.

Could it be that one carby food isn't the same nutrients as another carby food?

Unfortunately, seeing this item right next to https://news.ycombinator.com/item?id=44755394 primed me in the wrong direction. ("What phones were connected to it?")

Not if one subscribes to Assembly Theory https://en.wikipedia.org/wiki/Assembly_theory

not what?

John Tyler Bonner writes on this theme this in The Evolution of Culture in Animals. While we usually associate learning and culture with higher animals, where do we draw the boundary? Can a single cell learn and have memory? Can it develop a culture? Can it have a memory?

A great book which expanded how I think about our own learning and cultural heritage

"Remember?" this sounds more like we have so little observation of cells that when they respond to stimuli they have receptors for, we just never know that receptor existed.

Anthropomorphizing cells as anything beyond little machines seems silly.

If a cell has an injury that impacts its mobility but not its ability to survive because it can adapt -- is that a memory? Are nutrients within a cell? How about its current position in relationship to a food source? Or a buildup of some molecule in response to environmental factors? If not to any of these, why not? Because it's not neuronal by-nature?

The past dictates the future and there are many, many, many, many, MANY ways to encode the past for future self-preservation.

You're doing yourself an intellectual disservice by sticking to one form of "memory". I'd even argue that you're the one anthropomorphizing the hardest by limiting "memory" to an idealized human form of memory.

"Memory" can be used to refer to any form of statefulness, which is mostly what's being discussed here. I agree the article engages in some unwarranted anthropomorphism in some places, but there's still plenty of interesting material without that.

Cellular memory is well-established in epigenetics, where DNA methylation and histone modifications persist through cell divisions without changing the underlying DNA sequence. These mechanisms allow cells to "remember" previous states and environmental exposures, which is fundamentally different from simple receptor-based responses.

You might want to look into the work of Michael Levin’s lab to give you a broader perspective on the intelligence of individual cells and tissues. Your machine metaphor is arguably as misleading as anthropomorphization.

Edit: did you read the article? The examples described go far beyond your straw man about undiscovered receptors.

'member the first visual neural nets? just a few black spots in a grid, then colored spots, ... those grids were memory.

Given how much more capable these little machines are than anything we can program, I think that "remembering" is hardly a stretch at all.

Reply on news.ycombinator.com