Can your brain really be full? | Hacker News Reader

Can your brain really be full?

theconversation.com

60 pointsarnie00158 comments

No more than a neural network can be full. To anybody who has worked with neural networks, the following comes as no surprise:

    old information is sometimes pushed out of the brain 
    for new memories to form

Weights in a NN like the brain would just gradually shift. Older, seldom-accessed information will slowly fade.

That said, a NN can suffer from the trying-to-be-jack-of-all-trades, ending-up-an-idiot problem.

Another possibility is that two activation patterns are so similar that they interfere and thus they merge, one replaces the other or they both become useless noise. See also: https://en.wikipedia.org/wiki/Forgetting#Interference_theori...

this sounds like whats happening during an approach that is sometimes used in therapeutic settings: where a patient or client is induced into recalling a memory with negative emotional content, and then guided into transforming that memory by introducing competing neutral or if possible positive interpretations. for this purpose the result isnt useless noise, but instead its useful because the patient or client then experiences less negative pattern matching in daily life. at least thats the idea.

I have that problem trying to recall some books that I've read. I used to read a lot, and so many of the books I read as a teen seem to have merged into bigger story threads.

You could alternatively say that the brain is constantly full.

More so as you age. As a youth, you have very little total information.

My brain is definitely full. Every new thing I learn comes at the expense of something else I know (I guess this should be past tense).

The funny thing is I still remember when my brain had spare capacity - I used to fill it with arbitrary facts like what page I was up to when I was reading dozens of books at once. When I was young I never read one book at a time or used bookmarks. I would just remember the page I was up to and whenever I came back to a book I just turned to the correct page. Of course now I have to use bookmarks and every book I read is at the expense of some part of my past :(

Don't know why you're downvoted.

Technically, our brains do get full, even this article basically says so.

It's just that we don't experience error messages or shut down when that happens - instead, old memories are replaced with new ones.

The total capacity is limited, but the ability to create new memories is unhindered.

I wish we could choose what we forget, but it seems it's possible to choose what you remember - just relearn that stuff or recall it more often, then it will be at the top of the search results, so to say :-)

Yes you can choose to some extent what to remember, but if you have a good memory it is impossible to recall all your old memories so you don't forget them. My memory is still really good (I am hardly suffering from dementia), but it is not able to add new memories without losing the old. I hate this, but I can't really do anything about it :(

What about your ability to understand and process new information?

I'm finding almost everything to be much easier to understand now, compared to how it was when I was younger. Back then, I could focus, read at much faster speeds and my memory seemed way better, if measured by the amount of information I could remember. Entire texts were memorized with one or two passes. But then I'd have to spend time figuring out what to do with the information.

Now, my feeling is that wherever data I absorb is slotted right next to familiar concepts, so even though I read slower, the amount of 'post-processing' is much reduced.

It feels like a compression algorithm. "This new concept is like A and B, except for X, and Y, which are similar to Z". Sometimes, I caught myself saying that aloud. I've had to develop coping mechanisms however, to avoid jumping to conclusions, as my brain loves to do.

Entirely unfamiliar concepts are now way harder. From foreign languages (with different roots) to... Haskell.

So you're saying it's a hard number and, one-for-one, if something new goes in, something old goes out?

Pretty much. Since around my late 20s everything new going in has been at the expense of something old. The only interesting thing is the loss is not a first in, first out system nor is it based on importance. I have retained pointless memories from the past and lost things that were quite important. All I can say is that my capacity is finite and I have reached the limit.

I'll concede that this might be true - but you most certainly can't validate it based on these descriptions. In fact, by definition, you are unable to forget "something" for every other "something" you learn. In fact - if you know you forgot something - is that not a new piece of information? Now that I'm writing it out I think it's definitely nonsense actually - sorry. As a parallel consider music (a "single" domain for simplicity) - you are implying you can hear a new song and forget an old one, AND KNOW that you forgot an old one(?). You're trolling us for sure.

No I am not trolling. The reason I know (more precise I think I know) that every new piece of information I learn is at the expense of something old, is that the old is not always lost without the metadata about the old being retained. I am often left with the memory of having once known something, but now lacking the actual knowledge. To give you an example I often come across plant species where I know I once knew the name (I can even often remember where I first learned its name), but I can now no longer tell you what it is. Basically I have memories of what my brain used to be like without it being like this anymore.

Kind of like how I can't remember a thing about a rafting trip I was supposedly on when I was 19, but I clearly remember running into someone I was on that trip with, and discussing/reminiscing about it, when I was 22. So at 22 I clearly had the memory (enough to talk in detail about it), and I now (more than 20 years later) have a recollection of that conversation, but other than that the memory of that trip at 19 is totally purged. (except some vague recollection about glowing moss in the woods).

I experience a very similar thing. You aren't alone in being able to look at meta-states of your mind and realize it's different, or things are slipping.

For me, I noticed I no longer had instant recall of various remote control buttons. Was a small thing, but until that time, I stored maps of buttons and modes for everything. Now, I sometimes do, but I find I have the general rules, and need to parse devices and apply them now, not just thought as action as it was before.

So you forget things. Everyone does. Now demonstrate that forgetting is caused by learning. It seems you just assumed that part.

This is definately hard to prove particular since we are talking about my own perception of my brain. Due to the nature of my career I have never had to stop learning new concepts and skill and the pace of this learning has been pretty constant over the last 25 years.

What I can say is that up until my late 20s this learning did not seem to come at the expense of past learning or memories. Since then I have noticed that while I can still learn new things at around the same pace as in the past, that new learning seems to kick out the old.

Since my ability to learn has not changed (I have had to do courses and other external tests that confirm this personal observation), then the most parsimonious explanation is that new learning is coming at the expense of old learning which is just another way of saying my brain is full :)

Of course it is possible that my memory storage capacity is just getting smaller with age, but if this was the case I would expect it to impact on my ability to learn new things now. My perception is my brain can hold around 30 years worth of learning and as you get old that capacity gets stretched to the point that gaps start to appear. While better than the alternative, getting old is not good :(

One way of demonstrating this is if the rate of forgetting old information increases when you are doing some intense learning (either studying in school, or taking additional vacations, etc).

Has that been your perception? How do you measure that?

You can take a hash, delete the file, and record the deletion. You'll use less space, know something is missing, and even be able to identify the deleted file if you encounter it again.

Or maybe you can jam an SSD in your ear and use it as swap space?

I thought this was why people carry around phones - why remember anything when you can just ask siri?

I think it's possible to know that you forgot something. You can remember the title of a song, but the lyrics/details are deleted and replaced with a much shorter "memory was deleted to free space" message :-)

This is true - the metadata about a memory and the actual memory are often stored seperately. In regard songs memory loss seems to occur with the song title before the cord progression is lost. This results in the classic old person lament that all new music sounds the same.

So then the brain must have some kind of weighting algorithm to decide what memory of equal value to jettison. Surely learning a few new URLs poses no risk to remembering which kind of car you drive?

And on the subject of memories, are all memories information? If you learn something new in mathematics, do you risk losing the memory of the cake at your fourth birthday party, or only something technical you learned about archery?

Also, if what you are saying is true, you are vulnerable to a DoS attack on your brain. I could flood you with new information right now and, if you read it, you might not remember your mother's name by the end of it.

I've found that my weighting is a function of recall. The more often I recall a memory, the more likely I am to continue to remember it, conversely unrecalled memories tend to fade more and in a sense become compressed or move to cold storage which requires more concentration to recall.

I've always assumed this was an evolutionary result of weighting memories based on their usefulness, wherein the most useful memories are correlated with their recall frequency. I.e. they are the most useful for surviving my day to day life since I seem to need them a great number of times.

Your mention of the idea of DoSing someone's brain reminds me of Ewen Cameron's experiments attempting to "depattern" people. (see http://en.wikipedia.org/wiki/Unethical_human_experimentation... )

Rather horrifying.

Cameron's "experiments" (torture) seemed to revolve more around sensory depravation and electric shock therapy (one of the major side effects of electro-shock is serious memory loss).

This does bring up an idea I had from years ago where instead of making prisons as boring as possible we make them amazingly interesting places where the prisoners have to use their brains constantly all day. The idea being that all this information will slowly push out the ideas and thoughts of the old person to be replaced with a new ideas and memories that make the person less likely to commit crime once released. The person should come out of prison a better person than they went in.

I am pretty sure that passing-out and altered states of consciousness are the brain's anti-DoS measures. You would have to counteract all those measures to successfully DoS a brain, and that would probably take a lot of drugs.

I doubt that. It is more likely that you are impatient. Recall of older experiences can be frustratingly slow.

To recover old info, first concentrate on it and recall as much as possible (what you call "metadata"), and then simply cease focusing on it. Your mind will "re-hydrate" the archived experience in the background. Sometimes this may take a day or two, but usually it requires less than 10 minutes.

Reconsider the memory periodically: each time you should feel an indicator of whether progress is being made (even though the memory isn't yet fully recovered).

One of the problems with doing this is the "recovered" memories are not exactly reliable.

One of the good things about my memory is if I remember something it accurate. I used to win lots of bets (for drinks mostly) by exploiting this - people would tell me X happened and I would say no it was Y and would they like to bet on it. I was always amazed at how many people would accept these bets even after losing to me many times - it seemed they forgot that I had an accurate memory :)

I feel the same way, especially after thinking about it while reading this article. It's the same thing even with small and specific things as well. For example, we did a home winemaking course and forgot how to drive afterwards. The brain is truly a marvel.

Has anyone who knew what they were talking about every seriously suggested that your brain could "fill up", like a hard disk running out of free space? It seems like the closest you could get to that is your neurons becoming completely static (i.e. not-plastic), unable to form new connections or sever existing ones. That would kill you pretty quickly, or maybe another way to look at it is that it happens to you when you die, so the answer I guess could be "sort of, when you die".

But while you're alive? Assuming you don't just "add more space", to continue the analogy, it seems like what your brain does is probably closer to lossy compression, over and over again. And the important bits remain more-or-less readable.

To be technical ... from a information theoretical point of view your brain definitely can't hold infinite information so it definitely fills up in the sense you can't possibly be adding information without evicting/corrupting something else. I personally don't think we can reach that point within a human lifetime though.

Something more interesting to think about is the limits of http://en.wikipedia.org/wiki/Working_memory which you probably fill up on a day to day basis.

The brain is constantly evicting, shifting, and changing memories as new stimuli come in and change the connectivity of our neurons. In that sense the brain, from inception, is full in the same sense that in an artificial neural network every neuron has a weight. The only thing that changes are the weights. (Note that I realize ANNs are a dramatically overly simplified representation of the brain, but the point remains).

The brain may also choose to re-encode a memory as part of a recognized pattern, in a sense, finding a way of compressing memories as recognizing them as part of a kind.

Lossy compression is still a loss of information.

when we don't reach it in our lifetime, why do we have that over-capacity than?

The brain already consumes something like 20% of our energy.

Why would you assume that unused spare capacity is a significant fraction of that expense?

Maybe that excess capacity helps with efficiency? I know that some computer filesystems and flash devices get extremely slow when they start hitting 99% capacity.

It's like a piece of fabric one could keep weaving on - all the threads that get weaved on will always remain but only visible if it's near the top layer - That's my impression.

Having no scientific background and inclined to do computer analogies, I tend to think that our brain works like a RRD database [1], that is, we don't completely discard old information; instead, we store an aggregation of the said information as new information comes in. Then, these aggregations become our knowledge about a given subject; we don't need to store every detail, as this aggregated view is enough to help us understand new information.

Of course, if we ever need to reconstruct every detail of old info, we simply cheat, in the confabulation sense [2], possible with disastrous consequences [3]

[1] http://oss.oetiker.ch/rrdtool/

[2] http://www.spring.org.uk/2013/02/reconstructing-the-past-how...

[3] http://www.scientificamerican.com/article/do-the-eyes-have-i...

Without reading the article: of course. Evolution would not allow for much overcapacity.

However this will be done through lossy compression (you forget irrelevant shorter term stuff, you forget details, things get abstracted and generalized etc.)

Not that I disagree with your premise, but you can only use evolutionary arguments for populations in equilibrium. The human population has undergone so much recent evolution (I actually know of no non-domestic species that has changed more in the last 100,000 year) that you can't use evolutionary arguments like this. There may well be people alive that will never use all the capacity of their brain given their lifespan.

I suspect this is wrong, although I recognise your apparent authority in this field.

The human population has gone through more well-studied evolution than other populations; but, for example, novel species such as the London Underground mosquito are unlikely to be rare. You may claim that this is effectively a domestic species, but I claim that's just what we can easily study.

Human lifespans are not very much greater: chimpanzees live to be 60 years old, and they are 8,000,000-12,000,000 years of evolution away. The difference in average lifespan is primarily due to progress with infant mortality rate.

Speciation and evolution are not directly connected, but this is a minor point. For a population to be in equilibrium the population size has to be approximately constant. The human population (and our domestic species) have expanded so much in recent times that we are not in equilibrium. This means that there are lots of genes of great selective value that have yet to spread through the human population. My favourite is the ApoA-1 Milano gene which basic stops you from getting heart disease [1]. This only arose a couple of hundred years ago and unfortunately most of us don't have it.

1. http://en.m.wikipedia.org/wiki/ApoA-1_Milano

Yes, human evolution (and heart disease) is overtly very well-studied, to the extent that we can identify the likely patient zero for a gene mutation occurring in recent generations.

Our study of any other species comes nowhere close to that. To think that we can achieve that level of detail in another species is almost laughable. There are only a few species where we can guess what their equivalent to heart disease would be.

This doesn't provide evidence that humans are evolving more quickly: not at all.

We are touching on a lot of different concepts here, but evolution speed is proportion to population size not how much it has been studied - on the basis of genetics fruit fly are much better understood than humans.

My original point is that you can't apply evolutionary principles to humans since our massive population expansion and recent change from hunter-gather lifestyles to being farmers make all such assumptions void.

I always thought the cricket thing was unreasonably, almost impossibly, fast http://evolution.berkeley.edu/evolibrary/news/061201_quietcr... But maybe chirping isn't all that important to cricket reproduction.

"...old information is sometimes pushed out of the brain for new memories to form."

Reminds me of this Married with Children episode I've seen ;) http://www.imdb.com/title/tt0642312/

That's my greatest fear.

I'm terrified of the possibility that there could be a day when I hit my maximum capacity and have to forget things in order to be able to learn new things.

You forget things every single day.

Perhaps but I can't remember what.

Yes, temporarily at least. Flood someone with information and they will stop to make sense of the information they already have.

I'm sure the beer helps "tidy" it up. =)

Could it be that we consume too much useless information on social media and the internet?

Url changed from http://www.scientificamerican.com/article/can-your-brain-rea..., which points to this.

Reply on news.ycombinator.com