Hacker News new | past | comments | ask | show | jobs | submit login
Brain DNA Refolds Itself to Aid Memory Recall (quantamagazine.org)
62 points by scottlocklin on Nov 14, 2020 | hide | past | favorite | 12 comments



So (assuming a similar mechanism in humans) when you are taught, it primes memory formation, but it's only when you are tested that the memory forms.

This is, "tests", far from being merely evaluative or diagnostic, are educative.

This effect is well-known, but I feel it's easier to act on with a physical model.

If you work hard to learn something but don't then test (or use/recall) it, then all that priming was wasted.

The mechanism sounds like it's struggling to form memories, but it may also play a role in only remembering what's needed (not filling your head with useless learning, hence the need for the contrived "use" of tests) It could even be a separately evolved filter.

(pre-emptive qualification: there's some learning before the test else you couldn't recall it.)

PS I've just finished Spivak's Calculus chapter 1 problems (about one a day), and he introduces more new material there than in the body. There is tremendous reuse of previous problems (at times to a ridiculous extent, see 6(d) in the answer book) - I noticed that this consolidated my memory of them, but the physical model additionally tells me I'd best "test" myself on all the problems - especially the later ones which haven't yet been reused.


Yes, the act of testing i.e. "recall" consolidates memory. Recall may take place many times - obviously, later when the mouse returns to the scene. At that point, the memory might be strongly burnt in if the environmental cues are highly salient e.g. food is found, or a predator strikes.

Recall consolidation also takes place more mildly in the normal course of resting and sleeping. It is a default system idle process in animals. So, if the mouse visits a new maze, and then later on takes a break to just chill and snooze a bit, or do nothing in the corner of her cage, her system idle process will replay the sense data from the maze experience, consolidating the memories mildly.

The article is reporting that the ultra small scale technology required to image where the memories are physically being consolidated in the chromosomes has advanced to a promising point.

What is the physical data structure of memory? In computers, it is the silicone, organised into physical bytes. In mammals?

The last para of the article is indicative of what the researchers are exploring:

“We are currently ignoring 95% of the genome,” Marco said. He was taught to call it junk DNA. But like the enhancers that drive this aspect of memory encoding, the rest of these genes may take on crucial roles as well. “Although we mapped the genome, we still don’t understand most of it,” he said.

He is speculating that the data structure for memory storage is the junk DNA which comprises 97% of the human genome. Perhaps proteins surrounding the DNA in the chromosome packaging act as pointers. Perhaps the non-coding DNA can represent data structures such as a body map, or an environmental location map, or a timeline, or a dictionary. The technology will help scientists to identify what the pointers are pointing at. Over time, they hope to make sense of this, in the same way that cryptographers developed Reverse Encryption Algorithms through Shannon's principle of confusion and diffusion.


There's a difference between testing your knowledge (that is, simply trying if what you learned actually works), and being tested on your knowledge (that is, being judged on what was taught, which is what unfortunately happens in schools.)


I wonder if this is the mechanism of how animals can be born with instincts for complex behaviors.


It makes a lot of sense since it's unclear how the neuronal connectome could be passed on to offspring.


This is why modern Identity politics which is based on the notion that human behaviour is entirely a social construction is so silly. Proof that even your memory is guided by your DNA.


Biological mechanisms like this are not directly related to human behaviour patterns.

This is a pretty big non sequitur. And a conflation of "identity politics" with sociology. And a misrepresentation of sociology!


But thats like saying legs aren't related to human behaviour patterns. Without the DNA for legs, I proposition, humans would act very differently.


Correct! Likewise, without the DNA for a brain, humans would act very differently.

Please give an example of a human without the DNA for a brain.


It’s not black or white. How we behave is obviously heavily influenced by our initial DNA but there’s clearly a heavy influence on “mind” from external factors. It has been shown (proven?) that we can grow beyond things like “identity politics.”


To be fair, chromatin architecture is highly dynamic. The epigenetic marks are established throughout your life as a response to environment.


There is a big difference in definition depending on what part of the term "epigenetics" you emphasize.

Emphasizing "epi" means you are looking at architecture/mechanistic of chromatin and how they affect function. Emphasizing "genetic" means you care about whether the mechanism of interest is actually inherited.

There are some mechanisms that could perhaps be inherited through maternal contribution of enzymes involved in creating the marks. But trying to imagine how somatic epigenetic markers can be transferred to germ cells and then somehow transmitted to offspring past the epigenetic erasure that happens during fertilization is tough. It is definitely an area of research that I think is actually more interesting than studying the mechanisms themselves.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: