Comment by HarHarVeryFunny
14 hours ago
I wonder how much of this is directly related to age (biological aging), and how much is just someone's brain becoming "full" due to more memories getting added every year?
It seems that memories must be stored as embeddings with single multi-neuron assemblies (cortical columns?) storing multiple embeddings as a kind of contents-addressable memory that is able to keep memories distinct due to the very high dimensional space (# neurons per assembly) being used. However, you'd expect that at some point if you store too many memories in a single assembly the recall accuracy is going to go down.
You'd expect that with big brains being so costly, evolution has only equipped us with brains big enough to store a lifetime of memories, so it would be odd if memory didn't suffer as we get old.
To make a computer analogy, it's a bit like a hash table getting too full. Say you had a hash table without any overflow mechanism... up to a point recall may still be pretty good, but as the table gets closer to full there will be more hash collisions and likelyhood of "false recall". Obviously the brain is not a computer, but the analogy may hold up reasonably well if you consider the hash table keys and values as embeddings and the store operation being an embedding merge rather than overwrite.
The brain does not have the Von Neumann bottleneck. Unlike most current digital systems, the brain doesn’t have a separate memory registry it needs to pull from.
Engrams, that is, the physical trace of a memory, are not stable through life. They start out in the hippocampus, but as the stimulus recedes in time without reinforcement, it moves away.
No evidence exists though that the memory is encoded in one set of cells. This spatial segregation of memory is the worst hangover from the “brain is a computer” analogy. Even if it is, why in the world would it be like our digital devices which specifically have the Von Neumann bottleneck? In biology, memory and processing are not segregated.
There’s growing evidence the memory is much more distributed over the network, and is recomposed based on salience overlap with a new stimulus.
Another factor to keep in mind is circadian rhythms. There’s growing evidence for how much the memory system and timekeeping system overlap, at a molecular level. Every neuron (and other cell) has an intrinsic clock that ticks at roughly 24 hours, and continues to do so even in total darkness.
When you encode the memory has a lot to say, based on your chronotype, on how and how well you will remember it. Same with learning: there’s a time of day based variation.
Sleep, and dreaming, is when these memories seem to get replayed and critical features and connections are incorporated into the system and its regime, awaiting the right triggers to access a state similar to when the memory formed.
I’m stitching across a lot of different research, and I want to be clear many aspects of this system are not yet fully worked out.
But what we do know points to a system that works with different physical and algorithmic priors, and the dynamics are sharply distinct from current digital computers.
Hijacking your post with a dubious segue because I’m itching to bounce these thoughts off somebody:
I’ve been consuming a lot of talks / writing recently about “enactive” pictures of how our brains function. From what I gather, recent studies have called into question the entire idea of real world concepts being “represented” by an area of the brain at all. While it’s true that atandard fMRI-style snapshots of brain activity are semi-stable over the course of a short experiment, it’s not true over longer timeframes. The response to the same stimulus will change over time. They refer to this as “representational drift” in the literature, and some people are using this to bolster theories of mind that they consider non-representational. They instead emphasize the brain as a kind of dynamical system that learns to “resonate” with the world to pull itself back into homeostasis. The focus shifts away from facts and memories as data, and sees neuronal plasticity more as a mechanism for tuning the brain’s resonant frequencies. This obviously places high importance on the spiking, recurrent nature of actual neurons, as opposed to the neurons-as-functions / back-propagation / ML approach.
My mind’s not made up on how interesting and revolutionary this approach is / isn’t. The distinction seems to be about whether learning is more like “writing to disk” or “tuning a PID controller” - but in either case, the world is leaving a stateful imprint on your brain that will impact how it processes future data. Is that important to understanding how brains work, or is it just semantics?
Some years ago, I heard a retired Lutheran priest ponder, in radio, about the concept of a prayer, and whether, as an edge case, an unborn child could be able to pray, with the child of that age having no understanding on the required concepts. This is apparently one of those questions that people ponder under the umbrella of philosophy of religion. His conclusion was that prayer was about harmonizing one's self with the universe. It's a beautiful thought, (and I think religions, of all kinds, are really good at providing the fertile ground for thoughts like these).
It's akin to how the saying goes, that when we argue, we need to reach the same wavelength where the other one is to reach an understanding.
In communication studies or sociology or linguistics or one of those fields, there's the idea that communication is about making pacts about meanings, and finding the common ground to understand and delimit the message.
In a sense, from that basing, one could argue that understanding can be seen as an act of harmonizing. I think there's something universal in it.
Ooh thanks for the hijack! It’s so much easier to talk to someone who isn’t stuck in a picture of the brain from the 1980s.
Yes, I fall more towards the camp that a lot of our cognitive models, built from times when we didn’t have the resolution of understanding we have of the capacity of even a single neuron, and before we knew how astrocytes played a role, suffer from being an abstraction describing an abstraction. They are not tethered in the dynamics of the molecules and cells that give rise to the behavior, but rather from an interpretation of observed behavior.
This paper from the field of chronobiogy is one I’d recommend that helpfully contrasts this:
https://www.sciencedirect.com/science/article/abs/pii/S00393...
>In circadian research, the models are not proposals regarding the basic architecture of circadian mechanisms; rather, they are used to better understand the functioning of a mechanism whose parts, operations, and organization already have been independently determined. In particular, circadian modelers probe how the mechanism’s organized parts and operations are orchestrated in real time to produce dynamic phenomena—what we have called dynamic mechanistic explanation.
And what you’re describing, the enactivist description of cognition, (and 4E cognition more broadly as a framework), is one way the neuroscience community is trying to move past these issues.
Few things that give me confidence these are the right track:
1. Circadian rhythms are evolutionarily ancient. Bacteria have em. Plants have em. But different molecular tools shape very different clocks, though the same 24 hour cycle is being tracked. 2. The way these rhythms are generated is not through some central system that broadcasts the information to other regions. Instead, it’s instantiated in every cell in the body, and the behavioral rhythm is due to the synchrony between cells. Resonance absolutely plays a role, and has been well documented. The brains role, via the suprachiasmatic nucleus or SCN, is to orchestrate this synchrony, but it is not the source of the rhythms. 3. This slow rhythm definitely regulates cognition (time of day effects in learning, memory formation, recall etc are well documented), but turns out, the molecular mechanisms by which the clock responds to light hugely overlap with the molecular mechanisms of learning in the synapse, and even more recent work has shown clock proteins are actually in the synapses and synaptic activity affects the clock.
All this points to nested oscillators with cross frequency coupling, and even better, because this is all grounded in actual molecular dynamics, there’s plenty of falsifiability. The phase amplitude links are best established for the faster rhythms, the famous “brain waves”. Highly recommend György Buzsáki‘s work on this:
https://www.jneurosci.org/content/32/2/423.short
What’s missing is going down into lower frequency rhythms, and testing how exactly they all couple. We have a lot of the pieces, but no single experimental paradigm that has looked at the full sweep over different times in the same organism. It’s not easy to do, but we’ll get there.
Clock disruption, depending on how you do it, has huge impacts on time perception, cognition, memory, aging AND consciousness. As that data and evidence gets more and more saturated, I hope we see more studies account for chronotype and the internal dynamical state of their test subjects when assessing outcomes.
Obviously I’m biased (also did chronobiology in school), but hopefully I’ve left you curious. Happy to answer more questions all this may have set off.
2 replies →
> The brain does not have the Von Neumann bottleneck
Obviously not, which is why I didn't say it did!
However, if you want to identify where long-term memories are stored, then that is in the cortex, but it should go without saying that this doesn't make the cortex the storage component of a von-Neumann architecture!
> No evidence exists though that the memory is encoded in one set of cells
I'm not sure what you are trying to say.
Memories are presumably stored as embeddings - a distributed representation, and an episodic memory may well be stored as "chained together" episodic "scenes/chunks" where each chunk recalls the next.
However, a distributed representation isn't the same as a holographic one, and any redundancy may well still be localized within given cortical columns, so I think you may be wrong if you are saying that individual memories/chunks are not confined to one set of cells (some localized neural assembly such as a cortical column).
> Obviously not, which is why I didn't say it did!
But you sneak it into your assumptions on what a memory must be.
> However, if you want to identify where long-term memories are stored
And why do you assume there is a specific “where” for the memory?
> then that is in the cortex
There’s good deal of evidence disproving this in the way you’re stating this. The cortex is involved in sensing, and yes, the sensory information associated with a memory will recruit appropriate cortical cells. This doesn’t mean the memory resides in the cortex. And detailed episodic recall keeps recruiting the hippocampus even for old memories, which is odd if the memory were somehow only in the cortex.
> I'm not sure what you are trying to say.
Let me restate what I’m saying then:
There are four claims bundled together in the way you were describing biological memory: that a specific ensemble is activated when a given memory forms; that it’s the same ensemble that gets activated over time when that memory is retrieved; that it’s spatially compact, like a column in region of the brain; and that this ensemble it’s dedicated to that memory, or similar memories .
The first is well supported. An engram, a network of neurons, is indeed activated when a memory first forms, and gets stabilized due to repeated stimulus. Re-activating these neurons in a different context can make the subject (a mouse) behave as it would if it had contextual signals to evoke said memory.
However: 1. This engram is not in one particular part of the brain. There’s a cortical part that overlaps the sensory regions that were involved. But plenty of other regions are part of the engram 2. There’s turnover, over the course of weeks, when the specific cells involved in the engram drift, while the behavior remains stable. 3. The same synapses participate in many memories.
> Memories are presumably stored as embeddings
No. Let’s consider songbirds, which are an excellent worked out example (in an animal without the complex columnar cortical architecture mammals show, by the way).
What’s learned is a temporal sequence, with neurons in a nucleus in their brains each firing one brief burst at a fixed point in the motif, so the content of the memory is its dynamics rather than any value. The circuit that evaluates the match against the tutor template is the same circuit generating the output being evaluated. Song degrades overnight during sleep replay and recovers the next day, and in seasonal species the song nuclei change size across the year with neurons added and lost while the song persists. There’s no read that leaves the item untouched, no persistent address, and no substrate holding still. “Stored as” imports all three.
And it goes below the neuron or synapse. Hearing a tutor song drives immediate early gene expression that habituates with familiarity, and singing drives large transcriptional changes in the song nuclei that differ by social context for the same motor output. Since transcription runs on minutes to hours and the proteins turn over, any persistent state has to be actively regenerated rather than deposited.
TLDR: the memory isn’t a static store. There’s no persistent “location” for it, distributed or otherwise, though specific locations can be in the chain that’s activated for retrieval/production. Instead, memory, over time, is driven by a dynamical regime that adjusts its dynamics to account for the temporal pattern in the salient stimulus.
Nothing, down to the epigenetic changes in the chromatin of these neurons, can be seen as “the” location of “a” memory, especially over time.
> if you are saying that individual memories/chunks are not confined to one set of cells (some localized neural assembly such as a cortical column).
That is indeed the case.
2 replies →
None of that changes whether there is a physical capacity, which I think was the larger point? There is no reason to believe distributed memory doesn't suffer from the capacity component of the bottleneck. In fact iirc there was some late 80s/early 90s papers on the memory capacity of NN. Btw I would advise against the absolute statement that there's absolute segregation of memory and processing.
I guess my point is the brain not being "von Neumann" in architecture or digital is not proof that isn't a "computer" of some sort.
> None of that changes whether there is a physical capacity, which I think was the larger point?
Capacity in what sense? Are we saying it’s X MB of data the brain can store? That claim is steeped in assumptions.
On the other hand, no one is claiming the brain has infinite memory or anything. And it’s certainly not a very accurate memory system. I’m arguing against “capacity” being understood as “these specific physical components located here and here we can ID store memories, and can get crowded with too many memories” sense.
This most particularly fails because not all memory is even identical in the brain, whether we mean the physical changes associated, the topology of the information, or how it’s activated.
> There is no reason to believe distributed memory doesn't suffer from the capacity component of the bottleneck.
I didn’t know there was a capacity component to the bottleneck, only a bandwidth one.
All I’m trying to say is that analogy to current typical memory storage systems to explain the brains memory processes is not helpful.
> I guess my point is the brain not being "von Neumann" in architecture or digital is not proof that isn't a "computer" of some sort.
Indeed, since the word computer was first used for humans. But what kind of computer matters enormously. Ising machine? Quantum+classical stack? Reservoir computer? All those frameworks have processes in the brain they can point to as homology.
Which points to a possibility: maybe the brain is multiple types of computers interacting. And the physical realization of these computing architectures aren’t spatially separated but thread through each other in the biochemistry and physical dynamics of cells.
2 replies →
Does this mean we have unlimited memory storage?
This is really fascinating. The actual mechanisms of the human mind are distinct from computer systems, yet there are some parallels.
There are some hints that increased memory access times scale with the amount of information the brain has stored vs the more typical narrative that aging decreases the capabilities of the brain.
> Our results indicate that older adults'; performance on cognitive tests reflects the predictable consequences of learning on information-processing, and not cognitive decline. We consider the implications of this for our scientific and cultural understanding of aging.
[0] https://pubmed.ncbi.nlm.nih.gov/24421073/
A fascinating story of "Patient Sh." [1], the man who did not forget.
[1] https://en.wikipedia.org/wiki/Solomon_Shereshevsky
"His memory was so powerful that he could still recall decades-old events and experiences in the smallest details. After he discovered his own abilities, he performed as a mnemonist; but this created confusion in his mind. He went as far as writing things down on paper and burning it, so that he could see the words in cinders, in a desperate attempt to forget them. Some later mnemonists have speculated that this was a mentalist's technique for writing things down to later commit to long-term memory. Reportedly, in his late years, he realized that he could forget facts with just a conscious desire to remove them from his memory, although Luria did not test this directly."
I believe that one need to have superhuman memorization abilities to have definite confusion due to too much remembered. More trivial explanation of this effect in normal ageing persons is age-related brain shrinkage.
Our brain consists of approximately 86 billions quantum computers [2] controlling tens-of-thousands chemical neural networks with at least 10 coefficients, communicating [4] using lasers [5] (coherent light is laser light).
Live with that. ;)
> Live with that. ;)
Well, sure, it's a computer of sorts, although in the abstract sense of being able to perform computations so is our liver, so perhaps not a very useful concept.
What I meant (as I assume you realize) was "not a von Neumann architecture computer", but I'd also fairly confidently assert that it's not a quantum computer either.
Our ANN model of a neuron is obviously too simple (especially being a synchronous model, not a real-time asynchronous one), but it's hard to imagine that all of the classical chemistry, let alone quantum, details are important. It's necessarily built out of chemistry, but selection is happening at the level of behavior - presumably depending only on a much higher level set of abstract capabilities (ability to learn, etc), not the exact details of chemistry.
The success of LLMs, a crude prediction mechanism built atop a crude ANN, does tend to support the idea that low level details don't matter. Timing will matter if we want to go beyond LLMs to AI that can learn time-based things and not just sequence order, but how much else will matter remains to be seen!
I suspect the anxiety of being unable to forget things has an inherent selection bias. You will only stress over the things you can't forget, the things you have forgotten you won't stress over...
Generally claims of eidetic memories are overstated, doubly for older claims, but Nigel Richards memorized a French dictionary in 9 weeks, over 6k words a day, x2 including the alphagram.
I consider myself to have a decent memory, but that is 200x what I'd think myself capable of, assuming I want to retain it all at the end of the 9 weeks.
Our brain consists of approximately 86 billions quantum computers [2] controlling tens-of-thousands chemical neural networks with at least 10 coefficients
By that logic, a Blackwell GPU contains 208 billion quantum computers.
Each transistor is a quantum mechanical system and takes hundreds of model parameters to describe. So apparently a GPU is a 208-billion-node quantum supercomputer.
"The brain is not a computer" != "Parts / aspects of the brain do computation".
"A computer" in the first statement is IMO obviously intended in the common usage sense of the term, i.e. the brain is not a desktop computer or smartphone, or Turing machine, or etc, and thinking of it like these things will cause more error than insights. Also, billions of interlinked mini bio quantum computers arguably produce something with emergent properties and behaviour much, much more complex than "a computer". I am with GP, the analogy to "a computer" is not super helpful here unless you highly restrict the meaning.
But yeah, the Shereshevsky case is a super interesting one, thanks for linking!
> "A computer" in the first statement is IMO obviously intended in the common usage sense of the term, i.e. the brain is not a desktop computer
No one has ever thought that a brain is a desktop computer, and that is not the common usage sense of the term in this context.
1 reply →
The book Moonwalking with Einstein covers Patient S and memory in general. It's an enjoyable and quick read on the topic, would definitely recommend.
I see it differently: as a teen, I could remember the exact word for lots of things, I could recall numbers and data with high precision. Nowadays, I'm pretty bad at all those things, but I can very well understand metaphors and connects concepts from wildly different topics — which I couldn't back then.
I see this kind of "interconnecting ideas" of my current years as somewhat connected with "fuzzy memory". It's like, when I retrieve data, it's not as precise, and sometimes I get neighbouring data, but that can actually of use.
I feel like I feel this increasingly as I age. I used to be able to remember nuanced details of API and computer systems. Now, so many of those memories simply get stored as “FLAG: double check the docs” (because I’ll forget).
I’m a lot better at assessing things from a high level, knowing where to find the information I need, and knowing how to assess it for accuracy - but I can’t remember details anymore. Books become summaries. Trips become a few key snapshots representing the trip.
I am not a biologist so I cannot say this authoritatively, but I’m fairly certain that’s not how neuroplasticity works.
I only mentioned neural assemblies storing multiple embeddings - what about that contradicts your understanding of neuroplasticity?
How so?
I don't think there is a "full" state, but watching the changes as the years go by I have thought of collisions. Not really like a hash table, though--most collisions are with things that would be filed together. I think it's more like the brain returns a random item matching the category. Brains are expensive, memory that gets fuzzy with age is probably the better choice than trying to feed a brain that does better.
Yes, evolution is not going to penalize everyone with a "100-year brain" when the rest of our body seems optimized for something more like a 30-40 year lifespan. Evolution is all about propagating your genes, and not too many 100-year olds are going to be doing that!
In any case, fuzzy associative recall is really the "design spec" for a brain, so semi-graceful degradation and fuzzy recall comes for free.
I'm not sure if old-age memory unreliability is totally random - I would not expect it to be. I'd expect these associative recall failures to happen when there are too many memories partially matching the same "key" (activation pattern), which presumably happens mostly when they do have something in common.