I'm so sorry you went through that -- concussions can be really scary.
You raise a really good point about physical trauma being different from protein synthesis inhibition. Memory loss can happen in lots of different ways.
I can only imagine what having conversations floating around without their proper context must be like -- I'm sure it is a very strange experience.
And thank you for the kind words about consistency!
I guess it depends on the trauma. During a concussion, the biggest concern is that the brain's axons can stretch or shear due to the rapid movement of the brain inside the skull. This damage disrupts communication between brain cells and triggers an energy crisis as neurons struggle to regain balance. Memory loss of the event often happens because the hippocampus doesn't get the inputs it needs during the disruption. As a result, the brain doesn't encode the memories of what happened.
So human memory has the destructive read problem! The earliest digital computers used ferrite core memory (basically rings of magnetic ferrite suspended on a 2D grid of wires) which equally had to be written again after every read.
[The most amazing thing about the movie is that it WASN'T based on a novel by Philip K. Dick!]
That's so interesting - I had no idea. I have so many questions! Was this primarily due to the limitations of the technology at the time? And did the rewriting after every read introduce latency? what about storage? and ram?
It was due to the technology being used. The ferrite is more or less the same as in modern magnetic media so it's not intrinsic to the technology, but the system of X and Y wires used to access the core apparently destroyed the stored value as it read it.
It was all batch processing (punched cards, punched tape, line printers) and it was all pretty slow so latency wasn't an issue. You submitted your cards then some hours later you got your printout. We're talking late 1960's.
Ferrite rings were the storage or RAM, usually just called core. You also had hard disks and mag tape. Printed circuits came next.
I was an apprentice engineer in the computer department of a warship about a million years ago. We had about 1k of ferrite core, exactly as you describe. If you weren't actively using the memory, you had to save it to disk because it would only last for seconds.
The ship had two computers: one for fire control and the other for target tracking and they used to crash all the time. Part of my duty as a computer watchkeeper was to reboot the computer when it crashed. There were three steps:
1. There were 24 bit switches on the control panel and you had to remember 5x 24-bit commands that would tell the computer to read 5-hole paper tape.
2. You load a reel of 5-hole paper that would teach the computer to read 8-hole paper tape.
3. You load a reel of 8-hole tape that would teach the computer to read the disks and magnetic tape.
Only then could the computer get back to firing guns and missiles and I’d have to wind all those paper tapes back onto their reels because the reader would just dump the tape into a big basket.
About that time I was working at Institute of Oceanography in Surrey. We had ships with similar kit (IBM 1800's, CAMAC), tho I managed to avoid ever going on a research cruise.
A few years previously it had been the Admiralty's National Institute of Oceanography and a lot of the wartime radar work was done there. Allegedly it was still classified and still sitting in boxes on the top floor 30 years later.
Conversely we were regarded as low security and a convenient dumping ground for people they didn't know what to do with. You've heard of Philby, Burgess and Maclean, well we had the fourth one; nice chap, but he always had too much time to chat.
These were Ferranti 1600Bs that used an assembly language called FIDS-ABC. A, B, C were 5 bit registers that could add. The first 9 bits made up commands (I forget what they stood for). I could program in in FIDS-ABC back in the day.
I believe Mr Blunt knew a lot about art too.
I later became a sonar engineer and the documents for our main sonar, 2001, were stolen in another famous spy scandal with Lt Bingham. We called the Russian copy, 2001ski.
ICL - International Computers Ltd was the UK's national computer company, created by govt dictat by the merger of International Computers and Tabulators (ICT), English Electric Computers (EEC) and Elliott Automation in 1968. Govt Depts were constrained to buy ICL kit unless they could find a good reason to buy (better) kit from other suppliers. I worked there for a while: the attitude was that govt depts should be forced to buy our kit.
It's now Fujitsu, recently famous for the Horizon Scandal, still headquartered in the same building in Bracknell where I worked.
A story from our ship: they had trouble booting the computer one day - the paper tape kept running thru but it didn't boot. The captain said he'd take over while they got a coffee. When they got back they noticed the blank lead-in and the start of the tape was missing. They asked the captain about this and he said, yes, it got damaged so I tore it off, but there's plenty left...
Simona Ginsburg and Eva Jablonka in their book, The Sensitive Soul, describe studies that manage to cause synapses to retract, leading to loss of long term memory, but apparently only temporarily, indicating the memories are stored intracellularly. And they discuss theories that it might be stored epigenetically, which resonates with the efforts you discuss around RNA. But my impression is this last proposition remains shakier than the more general idea that it's stored somewhere inside the cell.
Thanks for the book recommendation, Mike! This sounds very much up my alley.
I'm curious - in the studies you mention from The Sensitive Soul, do you remember if they suspected these memories were temporarily hidden? In other words, I'm intrigued by why it was temporary -- if it was driven by a molecular change, a change in protein synthesis, or a change in neuronal structure. The fact that it was temporary is really interesting.
I'm fascinated by the memory research, too. I think we might be reading some really interesting science on memory in the next few years.
You're welcome Suzi! I should mention that they're principally focused on minimal consciousness in animals rather than the full human package. But I found their broad coverage of the literature fascinating. It's a long academic and technical book, which I know won't be an issue for you, but for me at times they delved into organic chemistry, going beyond my ability to follow entirely.
Here's the snippet from their book where they begin the discussion on intracellular memory.
"Epigenetic cell memory seems to be crucial for long-term behavioral memory, and there are now serious challenges to the idea that long-term memory (which depends on transcriptional activation and protein synthesis) is based on the persistence of synaptic structures. Studies of Aplysia neurons connected into networks in a petri dish, as well as intact living Aplysia, suggest that long-term memory does not reside only in synaptic connections. Treatments that disrupted long-term memory led to the loss (retraction) of the synapses that were formed following sensitization, as well as to the random loss of other synapses, yet long-term memory later recovered. This implies that synapses do not store memories but rather express memories that are stored intracellularly, in epigenetic marks.36 Other studies have shown that a single cerebellar Purkinje cell can encode the memory of a specific time delay in ferrets, again suggesting that intercellular connections cannot be the only locus of memory.37 These studies reinforce the conclusion that intracellular, epigenetic mechanisms are crucial for long-term learning and memory."
Ginsburg, Simona; Jablonka, Eva. The Evolution of the Sensitive Soul: Learning and the Origins of Consciousness (The MIT Press) (pp. 315-316). The MIT Press. Kindle Edition.
The studies they cite here are:
Chen, S., Cai, D., Pearce, K., Sun, P. Y. W., Roberts, A. C., and Glanzman, D. L. (2014). Reinstatement of long-term memory following erasure of its behavioral and synaptic expression in Aplysia. eLife 3, e03896. doi:10.7554/eLife.03896.
Jirenhed, D.-A., Rasmussen, A., Johansson, F., and Hesslow, G. (2017). Learned response sequences in cerebellar Purkinje cells. Proceedings of the National Academy of Sciences USA 114, 6127–6132. doi:10.1073/pnas.1621132114.
They follow this with a longer discussion on epigenetic mechanisms supporting memory.
Their papers on Unlimited Associative Learning actually cover the central thesis of the book. (Although the book obviously goes into a lot more detail.)
A story from the earliest days of neuroscience suggests two things. First, that emotions exist in memories physically apart from declarative or episodic memory. And second, that emotional memories influence behavior unconsciously. More recent research supports this.
In 1911, a French physician named Edouard Claparede encountered a female patient with damage to both sides of her hippocampus. This kind of damage makes the consolidation of short-term memory into long-term memory impossible. His patient was incapable of creating any lasting episodic or declarative memories. Every visit Dr. Claparede made to his patient was like his first: he had to introduce himself to her as if it were the first time. Since her procedural memory was intact, she engaged in social rituals and shook hands with Dr. Claparede. One day, Dr. Claparede hid a sharp pin in the palm of his hand. When they shook hands, the pain of the prick startled her, but it was superficial and healed quickly. Because of her condition, she forgot about the incident. Despite this, when Dr. Claparede visited this patient again, she refused to shake hands with him and could not explain her own behavior. She had never been hesitant to shake his hand before. The act of shaking hands previously solicited a response with neutral or positive valence. Yet, now the doctor’s outstretched hand invoked a strong, negative feeling.
There are two parts to remembering something: remembering it and finding the memory again when you need it. Breaking either of these counts as forgetting. Any theory needs to explain the many familiar properties of memory: can persist for a life-time, refreshing memories keeps them alive, old memories often take a long time (even days) to retrieve, people with eidetic memories remember "everything", memories degrade with time and even change to false memories. This sounds like weak chemistry. Proteins would be too stable. Bacteria use DNA for memory.
yes, good point. Sometimes it is difficult to tell if memory failure is because of a failure to encode or a failure to retrieve. My first published paper looked into this difference. And I still find it fascinating.
The most popular theory is that we don't really 'search' for memories. Instead, memories are created each time we remember them. A memory might begin (or be triggered by) something in our environment (like a smell, sound, or sight), or maybe one memory (or thought) triggers another memory (or thought). What we create (or reconstruct) depends on our current situation -- our current needs, goals, attention, and the setting.
You hear a word then pattern-match that sound against your hearing memory bank and find the right match which links to the meaning node.
Or you read the same word and pattern-match to your written words memory bank, then link to the same meaning node. After that it's just ChatGPT.
But sometimes it takes a couple of days to retrieve a fact so there's another distinct process.
As for building memories each time... why not reassembling? The question boils down to whether a single memory atom contains the whole event or the event is broken down into components which form a group of linked atoms.
The next question is to what extent each memory has its own private atoms and to what extent it reuses pre-existing atoms shared with other memories - I suspect mostly the latter and this works better if you have smaller atoms.
The research seems to support this, too. Memories aren't stored as whole events. It seems to be broken down into parts, which can be reused in many memories.
I agree with Malcolm, excellent choice of movie. Total Recall (based on a PKD story) would serve more to confuse rather than to enlighten. Great film though, imo. Thanks for this.
From the story about interfering with the rat’s amygdala, I'm intrigued by the idea that memories are stored in one place and used in another. I’m thinking of it as preventing the rat from experiencing a memory: although the memory is perfectly fine, the lessons stored in the memory don’t make it to where they can be useful. Is that about right?
That makes me wonder how the memories make it across the brain. I know that a memory is stored as a particular configuration of synapses. What is it about that configuration that gets sent ‘down the wire”? Is it even sent down a single wire (AKA axon)? Perhaps it is a whole new array of neurons of neurons and synapses that carry the message?
In previous lives, I have been an electronics engineer and a software engineer. Is the message more like an analog signal in a single wire? Or is it more like a digital signal? Or something else entirely. In a previous discussion, you told me that the transmission of a message is more like a conversation between the sender and the receiver. Does the anisomycin in the amygdala prevent the signal from being received? Or prevent the amygdala talking back?
You are making me wish I’d been a neuroscientist! I have so many questions!
I’m really interested in what a memory looks like and how it is stored across a bunch of neurons. Where can I read more about this? I have a pretty good understanding of how neurons work at the level of dendrites and action potentials etc (I’ve read Kandel’s Principles of Neuroscience cover to cover) but I am interested in the conceptual layer above that describes the shape of a memory and how it gets used and passed around.
Maybe there’s a book or a website I could start with? Thank you!
I’ve always had an excellent memory but it has been declining in recent months. It has been declining in very specific ways though.
My episodic memory is still excellent but I can’t recall vocabulary or proper nouns. If I see a word, I know what it means, but it won’t come to me. I know it’s all in there but it won't come out. Everyone has tip-of-the-tongue syndrome but this is about once per sentence for me now. Thinking about your rats forgetting their fears, I wonder if this is related to my memories being blocked from where they will be used.
It’s hard to make new memories too, especially of facts and proper nouns. I’ve never had to take notes before but now, even to remember a name from one website to type into another, I have to write it down or copy-paste it or I will forget it in two seconds. It makes studying for my philosophy degree rather difficult!
Mike's book suggestion is a good one. But Mike did suggest it might be a bit of a slog.
There's a new book that is getting a lot of great reviews:
Why We Remember by Charan Ranganath -- published this year (2024). It seems accessible, although I haven't read it.
There are the classics like:
In Search of Memory by Eric Kandel and The Seven Sins of Memory by Daniel Schacter. They are getting a bit old now, but they're still a great read.
How We Learn by Stanislas Dehaene is more recent (but I haven't read this one -- it has great reviews, though)
and if you're up for a challenge:
The Brain from Inside Out by György Buzsáki. -- Although it's not sticky about memory, it does explore how neural networks generate behaviour and cognition, including memory. It’s conceptual but rooted in a lot of detailed neuroscience. Buzsáki's work has changed the way I think about the brain. His book can be a slog -- I wouldn't recommend it to someone with little understanding of neuroscience.
I've read How We Learn. My favourite so far. One reason I started my philosophy degree was to learn more about the Philosophy of Mind but since I read the Dehaene book, I decided that philosophers don't have anything useful to say about it.
I read the papers that The Sensitive Soul is based on. Very good and not too hard but I don't need to read the book too.
I’m intrigued by the idea that you can somehow inject RNA into a neuron. Do you squirt RNA into the CSF and it somehow gets taken up by the neurons? Or do you somehow inject it directly into the neurons themselves? Into the soma?
When neurons fire with the right sort of synchrony, the combined electromagnetic energy reaches a physics based zone that exists as sight, sound, itchiness, pain, thought, and all elements of consciousness. So I think this explains how neurons from around the brain can create a unified consciousness that operates in a serial way (unlike the brain itself which functions in a massively parallel way — we think one step at a time while our brains preform countless separate non-conscious computations per second).
If it wasn’t for memory however, each moment of consciousness should essentially be lost to a given organism. Thus apparently we not only evolved to have present consciousness in term of our senses and thought, but synchronous neuron firing adds a highly crippled and condensed assessment of past consciousness to the present consciousness EM field. And why are some things remembered while others are not? Repetition should explain some of this though I consider value (or feeling good/bad) to be the essential fuel which drives the conscious mode of function. We tend to think more about horrible and wonderful experiences to thus entrench them deeper into our memories. This should help explain why further iterations can also become extra fictional. Recalling a memory should depend upon thinking the sorts of thoughts that synaptically spark chains of firing that exist as the desired EMF memory, and whether a name, experience, or whatever.
On your words not coming to you, while quite irritating and obviously not a good sign, it seems to me that your episodic memories and intellect itself should be a far more basic element of who you are.
Injecting RNA into neurons can be done in several ways, and the method used depends on the research goals and the type of experiment being conducted.
Yes, some directly inject RNA into the soma (cell body) of a neuron. This method allows precise control of the RNA delivered to individual neurons, but it’s technically challenging.
Other times (like work with sea slugs), I think it's injected into their blood (if I remember correctly).
Electroporation is sometimes used. This technique uses an electric field to temporarily disrupt the neuronal membrane, allowing RNA to enter directly into neurons, including their soma and processes.
As you mentioned, RNA can be introduced into the CSF, too.
The latest theory on how memory works isn't that memories are stored in one place and accessed in another. There isn't really one place where a memory is located. Rather, they're reconstructed -- or created -- each time. Memory is the process of coordinating the activity of multiple brain regions working together. So memories don't travel down neurons like electrical signals through wires or data packets through a network. Instead, remembering involves many brain regions simultaneously recreating patterns of activity that together form our experience of remembering.
That's a great question about whether the anisomycin in the amygdala prevents the signal from being received or prevents the amygdala from talking back.
When memories were recalled, they became fragile and susceptible to modification. The researchers found that giving anisomycin during this vulnerable period destroyed the emotional component of that memory pattern. This suggests that each time we remember something, we have to rebuild it - and during that rebuilding process, memories can be altered or lost.
So, the researchers don't think the anisomycin blocked a stored memory from being accessed or signals from being received -- they think that the memory pattern related to the emotional component of the memory was destroyed.
For me, the answer would be no unless it is something so terrible that people around me and I would be better off if my specific memories were erased.
All my positive and negative experiences and memories make me who I am. Also, you learn from your experiences/memories what to do and what not to do, preparing you for the future.
There is an interesting episode called "Passing Through Gethsemane" of the mostly excellent science fiction TV series "Babylon 5" that explored an ethical side to the erasure of memory. A notorious serial killer is caught and sentenced to "mindwipe", which leaves him with a new and much more agreeable personality, and no memory of his former life of crime. Without going into further details of the plot, the main question it raises is: to what extent is it ethical to punish someone for crimes he/she can't remember committing? And, if not, does it become ethical if the criminal can somehow be forced to remember those crimes?
The ethical questions quickly become more fundamental than a simple question of whether we should/ought to punish someone who doesn't remember their crimes.
The question we need to ask is: What are we trying to achieve with punishment -- retribution, rehabilitation, or prevention of future crimes? And how does memory play into those goals?
Agreed. Here in the USA there are a large number of people in prison. Some of them are quite dangerous, and society needs to protect itself from them. But I believe that many others are incarcerated for minor offenses that don't justify imprisonment. And, as is true in various other countries, the prisons are not well run and some have become hotbeds of corruption, violence, sexual assault, disease, etc. The whole system is badly in need of reform.
People also need to think more about the issues you mentioned: is the goal of imprisonment (or other forms of punishment) to deter future crime, rehabilitate the offenders, take revenge on them for their crimes, or some combination of those? I don't think our current system does a very good job achieving any of those goals , except maybe a perverse sort of revenge. There must be a better way, although I don't think it involves "mindwipe".
Actually, there is a dearth of thinking about lots of important issues. Unfortunately, thinking takes time and energy and makes people's brains hurt, so most prefer to avoid it if they can.
There's some interesting research that suggests that it is possible to have long term memories of events without short-term memory of them. Long-term memories may form on a separate, parallel path without needing consolidation in short-term memory. They blocked short-term memory formation in mice. They couldn't remember an hour later but could remember a month later. Described here:
I'm so sorry you went through that -- concussions can be really scary.
You raise a really good point about physical trauma being different from protein synthesis inhibition. Memory loss can happen in lots of different ways.
I can only imagine what having conversations floating around without their proper context must be like -- I'm sure it is a very strange experience.
And thank you for the kind words about consistency!
I guess it depends on the trauma. During a concussion, the biggest concern is that the brain's axons can stretch or shear due to the rapid movement of the brain inside the skull. This damage disrupts communication between brain cells and triggers an energy crisis as neurons struggle to regain balance. Memory loss of the event often happens because the hippocampus doesn't get the inputs it needs during the disruption. As a result, the brain doesn't encode the memories of what happened.
Hi Jack! The short answer is -- yes; it is possible to stimulate other brain regions besides the hippocampus to evoke memory-like experiences.
So human memory has the destructive read problem! The earliest digital computers used ferrite core memory (basically rings of magnetic ferrite suspended on a 2D grid of wires) which equally had to be written again after every read.
[The most amazing thing about the movie is that it WASN'T based on a novel by Philip K. Dick!]
That's so interesting - I had no idea. I have so many questions! Was this primarily due to the limitations of the technology at the time? And did the rewriting after every read introduce latency? what about storage? and ram?
Hahaha, he's written a few, hasn't he?
It was due to the technology being used. The ferrite is more or less the same as in modern magnetic media so it's not intrinsic to the technology, but the system of X and Y wires used to access the core apparently destroyed the stored value as it read it.
It was all batch processing (punched cards, punched tape, line printers) and it was all pretty slow so latency wasn't an issue. You submitted your cards then some hours later you got your printout. We're talking late 1960's.
Ferrite rings were the storage or RAM, usually just called core. You also had hard disks and mag tape. Printed circuits came next.
This is great, thanks!
the paleocomputer age!
I was an apprentice engineer in the computer department of a warship about a million years ago. We had about 1k of ferrite core, exactly as you describe. If you weren't actively using the memory, you had to save it to disk because it would only last for seconds.
The ship had two computers: one for fire control and the other for target tracking and they used to crash all the time. Part of my duty as a computer watchkeeper was to reboot the computer when it crashed. There were three steps:
1. There were 24 bit switches on the control panel and you had to remember 5x 24-bit commands that would tell the computer to read 5-hole paper tape.
2. You load a reel of 5-hole paper that would teach the computer to read 8-hole paper tape.
3. You load a reel of 8-hole tape that would teach the computer to read the disks and magnetic tape.
Only then could the computer get back to firing guns and missiles and I’d have to wind all those paper tapes back onto their reels because the reader would just dump the tape into a big basket.
24 switches means 24 bit words so presumably ICL?
About that time I was working at Institute of Oceanography in Surrey. We had ships with similar kit (IBM 1800's, CAMAC), tho I managed to avoid ever going on a research cruise.
A few years previously it had been the Admiralty's National Institute of Oceanography and a lot of the wartime radar work was done there. Allegedly it was still classified and still sitting in boxes on the top floor 30 years later.
Conversely we were regarded as low security and a convenient dumping ground for people they didn't know what to do with. You've heard of Philby, Burgess and Maclean, well we had the fourth one; nice chap, but he always had too much time to chat.
I don’t recognise ICL.
These were Ferranti 1600Bs that used an assembly language called FIDS-ABC. A, B, C were 5 bit registers that could add. The first 9 bits made up commands (I forget what they stood for). I could program in in FIDS-ABC back in the day.
I believe Mr Blunt knew a lot about art too.
I later became a sonar engineer and the documents for our main sonar, 2001, were stolen in another famous spy scandal with Lt Bingham. We called the Russian copy, 2001ski.
ICL - International Computers Ltd was the UK's national computer company, created by govt dictat by the merger of International Computers and Tabulators (ICT), English Electric Computers (EEC) and Elliott Automation in 1968. Govt Depts were constrained to buy ICL kit unless they could find a good reason to buy (better) kit from other suppliers. I worked there for a while: the attitude was that govt depts should be forced to buy our kit.
It's now Fujitsu, recently famous for the Horizon Scandal, still headquartered in the same building in Bracknell where I worked.
A story from our ship: they had trouble booting the computer one day - the paper tape kept running thru but it didn't boot. The captain said he'd take over while they got a coffee. When they got back they noticed the blank lead-in and the start of the tape was missing. They asked the captain about this and he said, yes, it got damaged so I tore it off, but there's plenty left...
Oh dear!
Fascinating as always Suzi!
Simona Ginsburg and Eva Jablonka in their book, The Sensitive Soul, describe studies that manage to cause synapses to retract, leading to loss of long term memory, but apparently only temporarily, indicating the memories are stored intracellularly. And they discuss theories that it might be stored epigenetically, which resonates with the efforts you discuss around RNA. But my impression is this last proposition remains shakier than the more general idea that it's stored somewhere inside the cell.
Thanks for the book recommendation, Mike! This sounds very much up my alley.
I'm curious - in the studies you mention from The Sensitive Soul, do you remember if they suspected these memories were temporarily hidden? In other words, I'm intrigued by why it was temporary -- if it was driven by a molecular change, a change in protein synthesis, or a change in neuronal structure. The fact that it was temporary is really interesting.
I'm fascinated by the memory research, too. I think we might be reading some really interesting science on memory in the next few years.
You're welcome Suzi! I should mention that they're principally focused on minimal consciousness in animals rather than the full human package. But I found their broad coverage of the literature fascinating. It's a long academic and technical book, which I know won't be an issue for you, but for me at times they delved into organic chemistry, going beyond my ability to follow entirely.
Here's the snippet from their book where they begin the discussion on intracellular memory.
"Epigenetic cell memory seems to be crucial for long-term behavioral memory, and there are now serious challenges to the idea that long-term memory (which depends on transcriptional activation and protein synthesis) is based on the persistence of synaptic structures. Studies of Aplysia neurons connected into networks in a petri dish, as well as intact living Aplysia, suggest that long-term memory does not reside only in synaptic connections. Treatments that disrupted long-term memory led to the loss (retraction) of the synapses that were formed following sensitization, as well as to the random loss of other synapses, yet long-term memory later recovered. This implies that synapses do not store memories but rather express memories that are stored intracellularly, in epigenetic marks.36 Other studies have shown that a single cerebellar Purkinje cell can encode the memory of a specific time delay in ferrets, again suggesting that intercellular connections cannot be the only locus of memory.37 These studies reinforce the conclusion that intracellular, epigenetic mechanisms are crucial for long-term learning and memory."
Ginsburg, Simona; Jablonka, Eva. The Evolution of the Sensitive Soul: Learning and the Origins of Consciousness (The MIT Press) (pp. 315-316). The MIT Press. Kindle Edition.
The studies they cite here are:
Chen, S., Cai, D., Pearce, K., Sun, P. Y. W., Roberts, A. C., and Glanzman, D. L. (2014). Reinstatement of long-term memory following erasure of its behavioral and synaptic expression in Aplysia. eLife 3, e03896. doi:10.7554/eLife.03896.
Jirenhed, D.-A., Rasmussen, A., Johansson, F., and Hesslow, G. (2017). Learned response sequences in cerebellar Purkinje cells. Proceedings of the National Academy of Sciences USA 114, 6127–6132. doi:10.1073/pnas.1621132114.
They follow this with a longer discussion on epigenetic mechanisms supporting memory.
This is amazing. Thanks, Mike!
I found and read a few papers by them that were kind of precursors to the books. Very good.
Their papers on Unlimited Associative Learning actually cover the central thesis of the book. (Although the book obviously goes into a lot more detail.)
https://link.springer.com/article/10.1007/s10539-020-09772-0
That's the one I read last night.
Great post.
A story from the earliest days of neuroscience suggests two things. First, that emotions exist in memories physically apart from declarative or episodic memory. And second, that emotional memories influence behavior unconsciously. More recent research supports this.
In 1911, a French physician named Edouard Claparede encountered a female patient with damage to both sides of her hippocampus. This kind of damage makes the consolidation of short-term memory into long-term memory impossible. His patient was incapable of creating any lasting episodic or declarative memories. Every visit Dr. Claparede made to his patient was like his first: he had to introduce himself to her as if it were the first time. Since her procedural memory was intact, she engaged in social rituals and shook hands with Dr. Claparede. One day, Dr. Claparede hid a sharp pin in the palm of his hand. When they shook hands, the pain of the prick startled her, but it was superficial and healed quickly. Because of her condition, she forgot about the incident. Despite this, when Dr. Claparede visited this patient again, she refused to shake hands with him and could not explain her own behavior. She had never been hesitant to shake his hand before. The act of shaking hands previously solicited a response with neutral or positive valence. Yet, now the doctor’s outstretched hand invoked a strong, negative feeling.
More on emotion and memory at https://tomrearick.substack.com/p/why-emotion-matters
There are two parts to remembering something: remembering it and finding the memory again when you need it. Breaking either of these counts as forgetting. Any theory needs to explain the many familiar properties of memory: can persist for a life-time, refreshing memories keeps them alive, old memories often take a long time (even days) to retrieve, people with eidetic memories remember "everything", memories degrade with time and even change to false memories. This sounds like weak chemistry. Proteins would be too stable. Bacteria use DNA for memory.
yes, good point. Sometimes it is difficult to tell if memory failure is because of a failure to encode or a failure to retrieve. My first published paper looked into this difference. And I still find it fascinating.
I'm intrigued by how we search for memories. Do we know? Are there theories? (Obviously there are!)
The most popular theory is that we don't really 'search' for memories. Instead, memories are created each time we remember them. A memory might begin (or be triggered by) something in our environment (like a smell, sound, or sight), or maybe one memory (or thought) triggers another memory (or thought). What we create (or reconstruct) depends on our current situation -- our current needs, goals, attention, and the setting.
Yes, sorry I keep thinking linearly.
You hear a word then pattern-match that sound against your hearing memory bank and find the right match which links to the meaning node.
Or you read the same word and pattern-match to your written words memory bank, then link to the same meaning node. After that it's just ChatGPT.
But sometimes it takes a couple of days to retrieve a fact so there's another distinct process.
As for building memories each time... why not reassembling? The question boils down to whether a single memory atom contains the whole event or the event is broken down into components which form a group of linked atoms.
The next question is to what extent each memory has its own private atoms and to what extent it reuses pre-existing atoms shared with other memories - I suspect mostly the latter and this works better if you have smaller atoms.
The research seems to support this, too. Memories aren't stored as whole events. It seems to be broken down into parts, which can be reused in many memories.
I agree with Malcolm, excellent choice of movie. Total Recall (based on a PKD story) would serve more to confuse rather than to enlighten. Great film though, imo. Thanks for this.
I agree about Total Recall. It's a brilliant film, but it might take 2000 words just to explain the plot! Inception has the same problem.
From the story about interfering with the rat’s amygdala, I'm intrigued by the idea that memories are stored in one place and used in another. I’m thinking of it as preventing the rat from experiencing a memory: although the memory is perfectly fine, the lessons stored in the memory don’t make it to where they can be useful. Is that about right?
That makes me wonder how the memories make it across the brain. I know that a memory is stored as a particular configuration of synapses. What is it about that configuration that gets sent ‘down the wire”? Is it even sent down a single wire (AKA axon)? Perhaps it is a whole new array of neurons of neurons and synapses that carry the message?
In previous lives, I have been an electronics engineer and a software engineer. Is the message more like an analog signal in a single wire? Or is it more like a digital signal? Or something else entirely. In a previous discussion, you told me that the transmission of a message is more like a conversation between the sender and the receiver. Does the anisomycin in the amygdala prevent the signal from being received? Or prevent the amygdala talking back?
You are making me wish I’d been a neuroscientist! I have so many questions!
I’m really interested in what a memory looks like and how it is stored across a bunch of neurons. Where can I read more about this? I have a pretty good understanding of how neurons work at the level of dendrites and action potentials etc (I’ve read Kandel’s Principles of Neuroscience cover to cover) but I am interested in the conceptual layer above that describes the shape of a memory and how it gets used and passed around.
Maybe there’s a book or a website I could start with? Thank you!
I’ve always had an excellent memory but it has been declining in recent months. It has been declining in very specific ways though.
My episodic memory is still excellent but I can’t recall vocabulary or proper nouns. If I see a word, I know what it means, but it won’t come to me. I know it’s all in there but it won't come out. Everyone has tip-of-the-tongue syndrome but this is about once per sentence for me now. Thinking about your rats forgetting their fears, I wonder if this is related to my memories being blocked from where they will be used.
It’s hard to make new memories too, especially of facts and proper nouns. I’ve never had to take notes before but now, even to remember a name from one website to type into another, I have to write it down or copy-paste it or I will forget it in two seconds. It makes studying for my philosophy degree rather difficult!
I think Mike’s book suggestion above might be a good place to start reading…
Mike's book suggestion is a good one. But Mike did suggest it might be a bit of a slog.
There's a new book that is getting a lot of great reviews:
Why We Remember by Charan Ranganath -- published this year (2024). It seems accessible, although I haven't read it.
There are the classics like:
In Search of Memory by Eric Kandel and The Seven Sins of Memory by Daniel Schacter. They are getting a bit old now, but they're still a great read.
How We Learn by Stanislas Dehaene is more recent (but I haven't read this one -- it has great reviews, though)
and if you're up for a challenge:
The Brain from Inside Out by György Buzsáki. -- Although it's not sticky about memory, it does explore how neural networks generate behaviour and cognition, including memory. It’s conceptual but rooted in a lot of detailed neuroscience. Buzsáki's work has changed the way I think about the brain. His book can be a slog -- I wouldn't recommend it to someone with little understanding of neuroscience.
Thank you!
I've read How We Learn. My favourite so far. One reason I started my philosophy degree was to learn more about the Philosophy of Mind but since I read the Dehaene book, I decided that philosophers don't have anything useful to say about it.
I read the papers that The Sensitive Soul is based on. Very good and not too hard but I don't need to read the book too.
Kandel sounds interesting.
One more question!
I’m intrigued by the idea that you can somehow inject RNA into a neuron. Do you squirt RNA into the CSF and it somehow gets taken up by the neurons? Or do you somehow inject it directly into the neurons themselves? Into the soma?
Consider some of my own answers Ragged Clown.
When neurons fire with the right sort of synchrony, the combined electromagnetic energy reaches a physics based zone that exists as sight, sound, itchiness, pain, thought, and all elements of consciousness. So I think this explains how neurons from around the brain can create a unified consciousness that operates in a serial way (unlike the brain itself which functions in a massively parallel way — we think one step at a time while our brains preform countless separate non-conscious computations per second).
If it wasn’t for memory however, each moment of consciousness should essentially be lost to a given organism. Thus apparently we not only evolved to have present consciousness in term of our senses and thought, but synchronous neuron firing adds a highly crippled and condensed assessment of past consciousness to the present consciousness EM field. And why are some things remembered while others are not? Repetition should explain some of this though I consider value (or feeling good/bad) to be the essential fuel which drives the conscious mode of function. We tend to think more about horrible and wonderful experiences to thus entrench them deeper into our memories. This should help explain why further iterations can also become extra fictional. Recalling a memory should depend upon thinking the sorts of thoughts that synaptically spark chains of firing that exist as the desired EMF memory, and whether a name, experience, or whatever.
On your words not coming to you, while quite irritating and obviously not a good sign, it seems to me that your episodic memories and intellect itself should be a far more basic element of who you are.
Injecting RNA into neurons can be done in several ways, and the method used depends on the research goals and the type of experiment being conducted.
Yes, some directly inject RNA into the soma (cell body) of a neuron. This method allows precise control of the RNA delivered to individual neurons, but it’s technically challenging.
Other times (like work with sea slugs), I think it's injected into their blood (if I remember correctly).
Electroporation is sometimes used. This technique uses an electric field to temporarily disrupt the neuronal membrane, allowing RNA to enter directly into neurons, including their soma and processes.
As you mentioned, RNA can be introduced into the CSF, too.
There are probably other ways, too.
The latest theory on how memory works isn't that memories are stored in one place and accessed in another. There isn't really one place where a memory is located. Rather, they're reconstructed -- or created -- each time. Memory is the process of coordinating the activity of multiple brain regions working together. So memories don't travel down neurons like electrical signals through wires or data packets through a network. Instead, remembering involves many brain regions simultaneously recreating patterns of activity that together form our experience of remembering.
That's a great question about whether the anisomycin in the amygdala prevents the signal from being received or prevents the amygdala from talking back.
When memories were recalled, they became fragile and susceptible to modification. The researchers found that giving anisomycin during this vulnerable period destroyed the emotional component of that memory pattern. This suggests that each time we remember something, we have to rebuild it - and during that rebuilding process, memories can be altered or lost.
So, the researchers don't think the anisomycin blocked a stored memory from being accessed or signals from being received -- they think that the memory pattern related to the emotional component of the memory was destroyed.
It's fascinating stuff. It's a good project for me to learn some more once I hit retirement.
For me, the answer would be no unless it is something so terrible that people around me and I would be better off if my specific memories were erased.
All my positive and negative experiences and memories make me who I am. Also, you learn from your experiences/memories what to do and what not to do, preparing you for the future.
Those are my thoughts, too.
There is an interesting episode called "Passing Through Gethsemane" of the mostly excellent science fiction TV series "Babylon 5" that explored an ethical side to the erasure of memory. A notorious serial killer is caught and sentenced to "mindwipe", which leaves him with a new and much more agreeable personality, and no memory of his former life of crime. Without going into further details of the plot, the main question it raises is: to what extent is it ethical to punish someone for crimes he/she can't remember committing? And, if not, does it become ethical if the criminal can somehow be forced to remember those crimes?
Fascinating!
The ethical questions quickly become more fundamental than a simple question of whether we should/ought to punish someone who doesn't remember their crimes.
The question we need to ask is: What are we trying to achieve with punishment -- retribution, rehabilitation, or prevention of future crimes? And how does memory play into those goals?
Agreed. Here in the USA there are a large number of people in prison. Some of them are quite dangerous, and society needs to protect itself from them. But I believe that many others are incarcerated for minor offenses that don't justify imprisonment. And, as is true in various other countries, the prisons are not well run and some have become hotbeds of corruption, violence, sexual assault, disease, etc. The whole system is badly in need of reform.
People also need to think more about the issues you mentioned: is the goal of imprisonment (or other forms of punishment) to deter future crime, rehabilitate the offenders, take revenge on them for their crimes, or some combination of those? I don't think our current system does a very good job achieving any of those goals , except maybe a perverse sort of revenge. There must be a better way, although I don't think it involves "mindwipe".
Actually, there is a dearth of thinking about lots of important issues. Unfortunately, thinking takes time and energy and makes people's brains hurt, so most prefer to avoid it if they can.
There's some interesting research that suggests that it is possible to have long term memories of events without short-term memory of them. Long-term memories may form on a separate, parallel path without needing consolidation in short-term memory. They blocked short-term memory formation in mice. They couldn't remember an hour later but could remember a month later. Described here:
https://neurosciencenews.com/parallel-pathway-long-term-memory-28188/
It will be interesting if that is replicated.