Hacker News new | comments | ask | show | jobs | submit login
To Remember, the Brain Must Actively Forget (quantamagazine.org)
353 points by digital55 6 months ago | hide | past | web | favorite | 119 comments



Relevant convo between Sherlock Holmes and John Watson (My choice of Italics highlights).

Below excerpt from A Study in Scarlet, Sir Arthur Conan Doyle, 1887.

——

My surprise reached a climax, however, when I found incidentally that he was ignorant of the Copernican Theory and of the composition of the Solar System. That any civilized human being in this nineteenth century should not be aware that the earth travelled round the sun appeared to be to me such an extraordinary fact that I could hardly realize it.

“You appear to be astonished,” he said, smiling at my expression of surprise. “Now that I do know it I shall do my best to forget it.”

“To forget it!”

“You see,” he explained, “I consider that a man’s brain originally is like a little empty attic, and you have to stock it with such furniture as you choose. A fool takes in all the lumber of every sort that he comes across, so that the knowledge which might be useful to him gets crowded out, or at best is jumbled up with a lot of other things so that he has a difficulty in laying his hands upon it. Now the skilful workman is very careful indeed as to what he takes into his brain-attic. He will have nothing but the tools which may help him in doing his work, but of these he has a large assortment, and all in the most perfect order. It is a mistake to think that that little room has elastic walls and can distend to any extent. Depend upon it there comes a time when for every addition of knowledge you forget something that you knew before. It is of the highest importance, therefore, not to have useless facts elbowing out the useful ones.

“But the Solar System!” I protested.

“What the deuce is it to me?” he interrupted impatiently; “you say that we go round the sun. If we went round the moon it would not make a pennyworth of difference to me or to my work.”


The flaw in the attitude of Sherlock Holmes is that he does not consider the, for the lack of a better term, "compressibility" of knowledge. Then again, Shannon wouldn't do his thing for another half century, so it would not be fair towards Sir Doyle to expect him to have taken this into consideration.

The brain actually does a lot of lossy compression tricks to remember things together, plus I doubt that knowledge about the Solar system takes up any significant amount of space in the brain, since physics works with models. Very "compressible" information, really. But if we look at how much lore people remember about modern pop-culture trivia, like comics or TV series or games, it's a different story. The amount of mental effort and "mind space" that fans dedicate to remembering, say, World of Warcraft lore is mindblowing to me at times.


I think there's a related problem that would have been understandable at the time: if you don't understand how the world works, you are potentially missing out on patterned information which allows one to form more accurate models of the world across domains. The assumption that's made seems to be that different problem domains are all independent, but that's not completely true. That's probably the main reason that mathematics and language are so useful.


> The assumption that's made seems to be that different problem domains are all independent, but that's not completely true. That's probably the main reason that mathematics and language are so useful.

To be fair, the effectiveness of multi-disciplinary research seems to only have been acknowledged in recent decades, even though by and large the biggest breakthroughs have almost always been thanks to people looking outside of the silos of their field. The sciences are really "tribal", and it used to be much worse.

For example, when I started studying in the early 2000s I remember the ML bachelor (then: AI) having recently split off from CompSci and being super-territorial.

I also heard stories of CompSci having split off from mathematics a few decades before that, and during my Interaction Design (IxD) master I learned that it separated from HCI in the nineties (short version: HCI was idolising the methods of the quantitative sciences too much for its own design-oriented good, IxD decided to bring in methods and insights from the qualitative fields).

I wonder if this was better or worse in Doyle's time.


I think a significant portion of my brain (Similar in size to the knowledge from my Bachelor's degree) was dedicated to remembering every quest and map up to WotLK. It's surprising how much I still remember from 7 years ago.


Funny enough, that might also reveal something about how human brains store information: locations on a map sounds suspiciously compatible with the method of loci (aka "memory palace", thanks to a more recent incarnation of Sherlock Holmes)

[0] https://en.wikipedia.org/wiki/Method_of_loci


Music is also very easy to remember.


IIRC there is recent research confirming it is at least partially stored in the same are responsible for muscle memory, which supposedly explains quite a few things.


If that is true, the obvious question is why is it stored there?


Well, there are strong ties to movement in music. There is a reason that you don't dance to paintings, literature or poetry.

So one option could be that music arises partly from our motor functions to begin with.


>There is a reason that you don't dance to paintings, literature or poetry.

Ha ha, well put. Although poetry is sort of like music, a bit, at least rhyming poems are. Just looked it up, this article is interesting:

https://en.wikipedia.org/wiki/Rhyme


Came here to say this, which is the first thing I thought of after reading the title of this article.

Incidentally, I first read the Chinese version of "A Study in Scarlet". And I remember vividly (because I read it many times) what Sherlock said was translated as "Even if I had known it I would have tried to forget it". It wasn't until I read your comment did I realize it's actually "Now that I do know it I shall do my best to forget it." Funny.

EDIT: formatting


That makes a lot of sense. I need to reconsider my podcast and reddit habits.


I actually wonder what the neurological research is on this subject. Does too much information crowd other information out?


The brain will keep around information that is used automatically. In my experience its a very bad idea to forgo learning something out of the expectation that it won't be used. I did this often, and it was catastrophic for me -- it turns out we're not very good assessors of what we will need.

Doyle was a crank.


There's a related corollary to this question though, which is whether systems that require less human memory enable you to remember other things better.

For example, your phone can dial dozens or hundreds of your friend's phones without you remembering their numbers. This means you never need to use their phone numbers and hence don't remember them. A reasonable question you might ask is whether you are now free to spend that brainpower memorizing other, more useful things? Or are you actually less capable of memorizing other things because you are exercising your memory less?

There are other similar examples, for example password managers and even just general Google-able trivia. Is it a good thing or a bad thing for general intelligence that humans need spend less time memorizing these things?


I felt this way for a long time about geospatial directions, but when your phone / gps dies and you have no map, it's kind of a terrifying feeling of helplessness


I skirt around this by instead of depending on turn by turn I study the main route and then only go to turn by turn at the endpoint. That way I gain the knowledge of the macro-infrastructure and main roads - I tend to give directions to out of towners even in cities I’ve been in for a week.


So from what I understand, there is a tradeoff to whether or not you memorize something.

When you have a piece of knowledge memorized, it has the capability to be activated in parallel with other known knowledge. These parallel activations result in a new / novel idea to your neural network. If these activations haven't been had before, they may be a breakthrough in science, technology, or otherwise.

So in one aspect, if you do not memorize, you can not associate and create new things.

Luckily, with phone numbers and locations it might not be important or valuable to have them memorized. But that's hard to say. Many scientific breakthroughs and discoveries were made by people who had cross-discipline knowledge and associated two things that no one thought were valuable together. Potentially, even googleable factoids could potentially contribute to a scientific breakthrough.


It reminds me of a story about Seymour Cray (founder or Cray super computer company). I've read it multiple times from different sources, copying it from first search result now.

He also invented the now-famous algorithm for buying cars: you go to the dealer closest to your house, point to the car closest to the door, and say: "I'll take that one."

The algorithm wastes the least time on unimportant things (like buying cars) to leave you the maximum time for doing important things (like designing supercomputers).


Minor and probably not that interesting anecdote: I always struggled to remember sequences of numbers. Even if I had looked at the number just a fraction of a moment ago, they were gone. I would have to do things like repeat them over and over or remember two numbers at a time. I was always buzzing in my head anout remembering the numbers and that produced a noise that I think made me forget.

Then I started cleaning up my attention span a bit, less social media, less mindless scrolling, more thoughtful long term tasks and planning. It helped a lot, but it wasn't perfect still.

Then one day I tried reading a number sequence once, not making a single noise in my brain, and I jotted it down again with ease. It seemed the secret for me was to trust my memory had the number in working memory somewhere, rather than trying to persist it my concious stream of thought which inevitably garbled it.

The brain is weird.


A similar thing: for the longest time I could remember my Dutch phone number only if I was "thinking" in Dutch. So if I was having a conversation in English when someone needed it, I basically had to mentally "switch" languages. Which proved so hard that the only way to do so was to say my number out loud in Dutch, listen to myself and write down the numbers, then read the numbers again in English.


hmm, I am a bit opposite - no issues with remembering even longer numbers (learned long serial numbers of all posters on the wall as a child just out of boredom), but introduce to me 2-3 people I haven't seen, and I simply can't remember their names. Feeling like a incompetent idiot guaranteed.

Names stick like 2 seconds around in my memory and are gone. Heck sometimes even a single person is an issue. My memory ain't greatest but for names specifically it is almost not writable. Tried few mind tricks but not really successful. I am sure there must be some way to get this to more acceptable level, anybody has any idea?


Me too, i can never quite seem to remember names, it can be very embarrassing, sometimes i even forgot the names of my cousins i haven't seen in a long time.

My theory is that, we don't internalise names and link it to the respective person for it to be anything memorable. but what do i know...


> It seemed the secret for me was to trust my memory had the number in working memory somewhere, rather than trying to persist it my concious stream of thought which inevitably garbled it.

That happens to me too! PIN keys, or bank passwords I know I know, if I try to think to remember them, I'm unable to remember them correctly. Instead, just seeing the input interface, blank my brain and I start pressing keys... And I write them down perfectly!

Indeed, the brain is very weird.


Now that's a good way to remember my btc private keys ;)


Have you tried remembering the number as a picture rather than a list of digits? Works for me.


I'm not the grandparent, but after doing "Learning How To Learn" course on coursera[1] (very good), I finally realised I suffer from Aphantasia, where I don't see images in my imagination.

Even though I'd tried loads of memory techniques over the years as I have a very poor memory, I'd never really understood people actually see images. I thought they meant that they were, like me, simply listing the properties the image would have in your head.

So all these techniques like "remembering the number as a picture" never worked well from me, and finally explained why all the common techniques people recommend are pretty useless. It turns out people have a huge variation on how good they are at visualising things in their minds eye, and I'm right near the bottom in terms of skill.

On the plus side it turns out that I can make images appear, having done some exercises to help it, although I didn't keep at it and it still doesn't come naturally. Most people can improve and there are very few cases of people not actually being able to do it. I also read of one man who'd had a stroke and lost his ability to picture things in his brain.

Oddly, I am very quick at generally orienteering myself in new places and picking up map layouts in games, even though I don't visualize anything in my head.

[1]https://www.coursera.org/learn/learning-how-to-learn


I’ve never heard of Aphantasia before.

Funnily enough it never occurred to me to think of things as pictures for my whole childhood. Even though I am almost a completely visual thinker.

It’s worth thinking about how you think. Eg if I imagine time passing as images of numbers, I can keep track of time and talk at the same time. If I count them out in my head, I always loose track within seconds.


Oh wow, I just tried counting to ten with pictures while thinking a conversation. Neat trick!

I do something similar for separating my hands on the piano, the left hand becomes a pattern rather than literal chords and I can picture the hand placement and still think of the melody. Forever a work in progress that is though.


I wonder if that would work with the drums, I can never keep the foot pedal independent of the hi-hat/ride.


I worked with a graphic designer with aphantasia! If you will excuse the wordplay, I could never imagine what it was like inside her head, as her process would surely be very different to mine. But she was very skilled and had been doing it for years. She also didn't realise until she was in her 30s that it was something she had.

She was capable of fluently talking about changes and the effects they would have on the design and so the fact that this could all happen without ever picturing the changes was super interesting.


I remember sitting at a bar in a distant city in some other country one night, and the person next to me, who I had just met that day, suggested that we do a shot together. Now these shots were big, and it had already been quite the night, and I remember thinking, "If I do this shot, it's probably the last thing I'll remember tonight."

It was, but that's not the point of this anecdote. This experience gave me the idea for the premise of a sci-fi story: a drug which, once taken, has no affect on you whatsoever outside of one crucial impact, which is that whatever happens while you're on it, you won't remember once it wears off a few hours later.

The scientific findings related in this article seem to indicate that developing such a drug would be possible. Would people take it? I suspect they would, although I also suspect that it would lead to regrettable behaviour. (That is probably, in fact, its chief advantage.) What about you - would you take it?

Are experiences less meaningful if you don't remember them?


There might be a market for this with high end consultants or psychiatrists as a more secure alternative to signing an NDA. It'll be something similar to the movie Paycheck

You give the consultant the drug, then proceed to share your secrets with them and get their expert advice on your issue. Then by the end of the day/week/whatever they'll forget everything you told them and your secrets will be secure.


Philip K. Dick explored this idea in his short story "Paycheck"[0], which was made into a movie[1].

Edit: Did you edit your comment? I'm pretty sure you hadn't mentioned Paycheck. No problem, just making sure I'm not going crazy. :)

[0] https://en.wikipedia.org/wiki/Paycheck_(short_story)

[1] https://en.wikipedia.org/wiki/Paycheck_(film)


You're not going crazy, it's normal for the brain to actively forget, didn't you read? :P


Lol, no, no edits :)


I feel that fundamentally won't work. Experts like that become experts because they are exposed to years and years of clients. One can build useful abstractions without giving away secrets.


I'm sure those experts would only accept jobs like this a very small percentage of the time and only for very large amounts of $$$. Otherwise they're not just loosing out on their expertise, they're loosing out on life


End of the day/week? I assume they'll be detained and watched like a hawk the whole time? There's all sorts of ways to record information that persists even when our memory has been wiped clean. Just take Memento, for example, where the guy tattooed his body with information he kept forgetting, but they don't even need to do that. Pen and paper, talking into a phone, telling a friend the instant you leave, wearing a wire, all sorts of ways to bypass a forgetful pill.


I recall a cyberpunk RPG had an implant that did something similar. It blocked the transfer between short and long term memory.


[Spoiler] Another usage: Larry Niven used this for blackmail information, Neutron Star.


I may be mistaken but believe William Gibson had a similar concept in one of his novels where prostitutes would work while in such a state. They could do disturbing work and be free of the consequences of those memories.

I also believe that if such a drug were available it would find first use in the military by certain specialist units.


Neuromancer was the novel, “meat puppet” was the term.


Such a drug exists. Namely, anything of the benzodiazepine class.


Can it or has it been used for UX testing? Having a way to "reset" people for interface testing would be useful. Preferably without side effects, of course.


Drugging people with hypnotics to test your shitty UI? This is the most HN thing I’ve seen in a long time ;)


bingo^

short/longterm memories will be shot full of holes when on benzos for a while. even one night of benzo usage and you will forget most of the night, mix in a beer by accident and the whole night will be erased.


Benzos have anxiolytic effects. I think OPs premise is a drug that has no psychoactive effects (apart from the no long term memory thing)


Most people recognize that as Ambien

https://popula.com/2018/07/22/the-ambien-diaries/


Ambien is not a benzo. The bizarre behaviour it causes seems to be related to sleepwalking, which is not a common side effect of benzodiazepines.


I met someone who was on a clinical trial for Ambien. He'd pass out almost immediately, but first he'd see crazy wild hallucinations. But only for like 15 seconds to a minute.

When it was finally released on the market, the dosage was a fraction of what he had take in his clinical trial.


I think you have to feel really crazy adventurous or desperate for money, to become a guinea pig for new drugs, that mess with your brain ...


kind of, but not exactly.

benzos wipe out most of your ability to inhibit behavior along with your ability to form memories. it practically guarantees that you're a mess.

your concentration goes to shit, along with most of your higher functions... at least at the quantity that it takes to make you black out.

there are a lot of stories about what happens when the "xan man" (the xanax overdose state) comes out to play. most of them are bad.


Take enough “xan’s” and you’re describing the same effects


Personnaly I would never use such a drug. I conceptualise myself as a tought-based kernel with a bunch of meat around to make it run. This kernel tries hardly to think about itself as 'one thing' (you don't achieve much with schizophrenia), and that relies heavily on this kernel being able to trace a continuous stream of tought back to its birth (at least as far as it can remember). Alter this stream and you loose yourself (the bigger the altering the bigger the loosing).

Two more things: (i) I'm aware that not everybody thinks the same, mainly because fellow humans acts as a constant reminder of who you were before, and (ii) yes, sleeping is disturbing to me, I try not to think about it too much.


What's your opinion on the common teleportation dilemma? If there were a machine, that would instantly destroy my body and assemble it somewhere else (even from completely different atoms) so that from my point of view, I instantly vanish and appear elsewhere, does this count as interruption to my line of consciousness? Doesn't really seem like it to me. After all, my memories and thoughts are stored in the patterns of matter and those are preserved.


Teleportation isn't very concerning to me, because it's basically against known physic rules for now on. Theoretically I would be very wary of teleportation, on a pragmatic level: if I don't own the teleportation process on an extreme level then there's too much risk to be altered in some (maybe dramatic) way.

However, there's a variant that sucks my thoughts sometimes: what about transferring a full brain's backup into a valid recipient?

Classical application would be life extension: going into a "blank" body/brain (or big enough computer) when mine is too old to carry on. It's still very inconceivable nowadays, but much more attainable than teleportation I think.

You can't "switch off" the old brain, copy it, and "switch on" the new one, because that would pose the same questions as teleportation. Neither I wouldn't start a copy process, wait for it to complete, commit suicide and set alive the new body/brain in the same time.

The obvious solution would be to find a way to connect two brains (be it an organically or computerized brain), extend the thought in such a way that there would be full redundancy of memories between the two devices, and the switch off the old one.

I guess I read too much hard sci-fi :)


No I would not want to take it. "buyers remorse" will be so huge that I will regret for the rest of my life. For example if something bad happened then its fine but if something good happen and I can't remember, that would suck !


Alternatively, what happens if you are accused for a crime that you didn't commit during your "blackout period"?


While not a crime, i had some friends be very pissed off when i awoke from being blacked out drunk. To this day i do not know what faux pas i did...


Have you asked them? What sort of friends cannot broach the subject with you?

There's no reason to continue living in that situation if you can make amends or discover that they judged you in a faulty manner.


Seconded. They probably can’t remember. It’s a powerful thing to hold over somebody. If they ever fancy a chick who’s fancying you at a party for example, with that type of person something is sure to come out to put yhem above you. I’ve had some shitty drinking buddies and cut them all out of my life. Good riddance.

Life is for fucking up and growing, which requires feedback.


Not remembering committing a crime has never been a valid defence thus far and we have drugs that cause far more extreme behavioural changes besides just memory loss.


It would become a "rape drug" slipped into drinks. The new Roofies.


Passing time without any knowledge or memory is probably the main reason to be taking drugs in prison. Yours sounds like the perfect, side-effect free solution!


Interestingly, I wouldn't use it for that (if at all), because it doesn't actually pass the time at all. You still live the time, experiencing it in full. Ie, a boring day would still be a boring day.

On the flip side, a drug that let you retroactively forget would be heavily used I bet. Traumatic experiences are often wished to be forgotten. Though, I wonder what a human would be like if they didn't have negative experiences. One who was able to remove a big portion of mind shaping experiences. Further yet, I wonder if your mind ever truly forgets. If a dog bites you as a kid, and then you take the forget pill, are you still afraid of dogs?


I ran into an article recently that hinted that memory erasure may well be possible in the near future...


This sounds pretty close to my experience with Xanax.


That is an exact description of "Soy Sauce" from the "John Dies at the End" series.


anterograde amnesia is a property of some anesthesics used in emergency: the basic idea is that it may hurt, but the patient won't remember.

(actually, it is more of a dissociative effect - ketamine does that)


Ketamine absolutely stops you from feeling pain. It's one of the things that's dangerous about using it recreationally, you can injure yourself and not even be aware of it. Your leg could be bleeding out, but you're too ketty to notice and just keep partying on.

I haven't seen anybody seriously injured form this, but I've heard second hand stories.


Haldol often can cause complete memory loss in a portion of the population. I have actually worked with drug. Has a bunch of other side effects long term so I don't reccomend it.


They made a movie about that- it's called Men In Black.


They don't often willingly get their memorizes blanked though.


Sounds like scopolamine:

https://en.wikipedia.org/wiki/Hyoscine


I'd like to take a pill that helps you forget movies so I can rewatch them.


But then you also wouldn't know which ones to not bother rewatching...


IMDB top 100


commonly accomplished by smoking a largish joint beforehand.


Mike Tyson, a Tiger..


If we remembered everything, he said, we would be completely inefficient because our brains would always be swamped with superfluous memories.

Yet there are a few people who, it has elsewhere been claimed, can remember details from every day of their life, on demand. What's 'superfluous' to one of us may be essential to others... a strictly subjective judgement. Some of us are specialists, others generalists.

The trouble with such reporting is that there's little science involved, and a whole lot of anecdotal reportage. Of course, our brains are enormously complex, having to operate in a wide diversity of human and chemical environments. Given the number of variables, solid memory science has yet to emerge ... in articles like this one, at least.



I've heard of these people, but if I remember correctly they were of the type to obessesively recall the events/details of their day - like a form of OCD.


For the curious one, I think this is called Hyperthymesia.


Looks like Married With Children was on to something back in '94:

"When the Bundy TV set blows its condenser, Al decides to try out for a place on a new sports trivia game show, hoping to win $10,000 for a new TV set as the first prize. But when he is denied because of his lack of personality, Al tries to transfer his knowledge of sports to Kelly to win the show for him. But for each fact that she takes in, another falls out."

https://www.imdb.com/title/tt0642312/plotsummary?ref_=tt_ov_...


> Eleven of those genes, they discovered, were still active on one side of the animals’ brains but not on the other, even after the animals had apparently forgotten about the shock.

What exactly does active genes mean here and how can they do that in a living slug and be different on both sides? Afaik, genes are mostly constant throughout one individuals' lifetime, and usually the same in all cells

With almost a GB of data in the DNA of a single cell and 100 billion human brain cells, the DNA sounds like a great storage medium for memories (725 exabytes)! But that's not how it works, right?


The entire genome in each cell contains all of the genes for every type of cell in your body. In different types of cells and in different states certain genes are upregulated and downregulated. Essentially the process for turning a specific gene into the protein it codes for can be turned on or off. That is what is meant by an active vs inactive gene.

Note: the activity of a neuron and the activity of a specific gene (eg rac1) are two completely different kinds of activity. The first is electrochemical activity and the second is production activity. Where changes in production can change the physical number of receptors available for the electrochemcial process.


Use it, or lose it!


Makes sense. It is best to forget useless things and fill the brain with useful ones. I was also going to say something else, but I forgot.


Well, whether or not it is "useless" depends on the person. Most human have emotional attachment to memories. And hence those that are more emotional ( whether it is inside or outside ) tends to remember things better and in tiny details.

And I actually doubt we "forget" about it. It is not like a computer sudden decide to do GC and you have more memory space. I think our brains tends to archive those "useless" things into somewhere deep, that takes lot of energy and take it back out.

I don't believe information is lost in our brain, it is most encrypted, compressed ( lossy ) and stored differently.


Apparently each time a memory is accessed, it gets written down back again. Imperfectly.


As someone who has moved to Japan for the last ~12 years and speaks the language daily, I struggle to remember terms I am absolutely postive I know but can't recall while in the middle of conversations with my English speaking employees. Or even worse, deep in personal thought I have to start googling around to recall a term I know I know.

There are linguistic studies which show that this is the norm (i.e. as a second language learner using said language daily, you begin to forget your mother tongue), but I have to admit it is quite disconcerting when I have to take a multi-second break to recall an uncommon word before I can carry the conversation further.

I guess the moral of the story is that, yeah, you forget things you don't actively use.


I recently started taking noopept for memory health and I'm getting a side effect when I feel like I'm a college freshman again and don't know much about whatever I'm doing. It's helpful because I try new things that I wouldn't usually. I've learned a lot about bash and vim in the past two weeks as a result and I've tried new playing styles on guitar, while in the past I'd just play tabs from bands I like. I'd highly recommend it to anyone looking to learn quickly and break through mental blocks.


The book Behave by Robert Sapolsky describes this topic (among others) nicely.

https://www.penguinrandomhouse.com/books/311787/behave-by-ro...


This book was fantastic, and elegantly laid out in how it described behaviors that human beings take.



I know nothing of the neuroscience but I have a theory we look at memory wrong. We see it like storage. You take a cup, put it in a cabinet, and bring it out later.

But what about memories as your brain, not pulling it from storage, but entirely regenerating the experience. So its remaking the cup every time, not pulling it from the cabinet.

False/faulty memory is common. We’re regenerating the experience inaccurately. Like if you drew a picture then had to draw it again. It would look the same but slightly different. The cup after all if stored would always be retrieved exactly as it was.


I only skimmed the piece, but it doesn't look like it touched on the fact that the brain apparently "resets" during sleep (edit: specifically to foster new memories the next day, from what I gather). I have seen several articles in the last couple of years about the brain shrinking, etc, during sleep.

A quick search pulled up, for example, this piece:

https://www.newsweek.com/brain-synapses-shrink-during-sleep-...


One of my favorite papers on the subject, from a colleague's lab:

"Unstable neurons underlie a stable learned behavior"

https://static1.squarespace.com/static/5859e85dd2b8571a0a859...

>Motor skills can be maintained for decades, but the biological basis of this memory persistence remains largely unknown. The zebra finch, for example, sings a highly stereotyped song that is stable for years, but it is not known whether the precise neural patterns underlying song are stable or shift from day to day. Here we demonstrate that the population of projection neurons coding for song in the premotor nucleus, HVC, change from day to day. The most dramatic shifts occur over intervals of sleep. In contrast to the transient participation of excitatory neurons, ensemble measurements dominated by inhibition persist unchanged even after damage to downstream motor nerves.

Brain circuitry truly is amazing. The connectome is vital to the function of the brain, and slight perturbations can severely damage that function; however, the connectome can change over night even in the context of precise motor skills. It's a sort of Catch-22.


Tangential but interesting: There's a great Borges short story called Funes the Memorious, which imagines a man who could never forget anything. To remember the events of a day, he must spend an entire second day, going through it second by second.

https://marom.net.technion.ac.il/files/2016/07/Funes-the-Mem...


Over the years i have come to think that the brain can be better likened to a de-duplicating storage system, but with some nasty flaws that all too often result in mixed data on reading.

Meaning that experiences that are similar enough gets overlaid each other, resulting in time flying past as we get older and to the same stuff day after day. This because we seem to experience time retroactively.


Current neuroscience suggests that brain operates in broadly two regimes - learning during wakefulness and model selection during sleep. Both these activities require "forgetting" of specific kinds. Specifically model selection involves trimming synapses. The goal is to learn a model of the world at the right level of detail, what AI folks call avoiding over fitting.


Memory seems rather like a multi-tier cache hierarchy to me. You can only really keep something like seven objects in your mind at once, and everything else seems to be pushed to slower sections of your memory.

Within this analogy cache eviction strategies start to have massive effects on our ability to process information.


There's also the fundamental theory of this which ultimately rests on thermodynamics, e.g. [1] [2].

Basically, any finite organism only has a finite number of configurations available to it; but to 'remember' something is to a set up a long-term correlation between the past and the present, which uses up some of the 'available degrees of freedom' to e.g. respond to shorter-term changes in the environment.

If you never forget anything, then eventually you 'run out of bits', whereas a system that allows some forgetting has greater flexibility in e.g. forgetting the distant past in order to remember to tie your shoes right now.

There are some neuroscientists that are starting to bring thermodynamic views into looking at actual biological systems as well [3]

[1] https://en.wikipedia.org/wiki/Landauer%27s_principle

[2] https://physics.aps.org/articles/v11/49

[3] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5108784/


Micheal Pollan in Botany of Desire has a lot of good stuff about the importance of forgetting: https://books.google.com/books?id=Woywyw8LlcgC&pg=PA162&lpg=...

> For it is only by forgetting that we ever really drop the thread of time and approach the experience of living in the present moment.


Don't forget that you are not a fruit fly.


I have probably reach the top of my Remembering capabilities.

Often I feel actively the urge to make free space in my mind to store new information. Technology helps me keeping offloading useless/unimportant facts and store them somewhere else, but the overload of information is being an issue


Is that why smoking weed makes things seem profound, from instantly forgetting the less salient stuff?


Well I guess the saying "if you don't use it, you lose it" must be true


>> Hardt said. “It filters out the stuff that the brain deems unimportant.”

Most of the things I learned at school I forgot; all the math and so that I never need in my life. More than half of the curriculum is just rubbish for our brains IMAO.

Maybe this helps to start making compulsive education more pragmatic. At this moment I see kids coming from school knowing how to do advance math they'll never ever need. But basic knowledge about what is healthy food or not, or knowing how the industry is poisoning almost anything you can buy in the supermarket, they have no clue.. And I believe knowing those things do matter for our brain, we would not easily forget.


After I ran my head against math for some time and actually noticed changes to my mind structure, I am sceptical. Studying math is not (just) about the knowledge you gain, it is about the reasoning structure behind it, like a workout for the brain.

Learning things is not (just) about the facts you know afterwards. I read so many books I forgot at least half of what was written there, but my worldview and thinking was altered from that.

Not that the other stuff isn't very important too of course..


Just because you wouldn't remember how to tackle an advanced math problem doesn't mean it has been wasted.

You've gained experience and you can identify what you don't remember. Even if you can't remember it it is no longer magic, which is invaluable by itself.

Also the process of learning these abstract concepts is arguably the most important skill you learn in higher education. You might recognize what is required and you do know how to pick it up when you are faced with it - or know when you need external help.


This article seems to operate under the premise that photographic memory is not a thing. Is it considered an urban myth now?


So, simply put...

- short term memory: transaction journal

- long term memory: persistent table

- sleep: commit and garbage collection

As I often said, Computer Science can explain everything.


I should hope that it can at least explain your brain, which is a computer.


This is a huge misconception.


  For Robert Calin-Jageman, it’s exciting that forgetting
  seems to be a biological process like digestion or excretion
  because that means it can, at least in theory, be ramped up
  or down.
Huh? Of course it's biological. Why would one believe otherwise unless they made naive deductions from the metaphors du jure--computers and neural networks. Perhaps the journalist is misinterpreting the researcher, and was, indeed, working backward from modern metaphors to conclude that it's somehow novel to think the mechanisms intrinsically corporeal.


I think the keyword here is "process" not "biological"; i.e. that it's an action taken rather than random decay.




Applications are open for YC Summer 2019

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

Search: