Why and how is this kind of long-term potentiation (LTP) possible?
Is LTP even needed for all sorts of synaptic plasticity and long-term memory formation?
------------
Longer version:
Long-term potentiation (LTP which is necessary for synaptic plasticity and long-term memory formation) needs repeats and reinforcement of the engrams to be triggered.
However, apparently everybody automatically "absorbs" a lot of information immediately and also permanently, even without needing any extra effort (at least any conscious effort), which seems to be needed for LTP to happen. Everyone seems to have this ability, although it is even stronger in those with better memories.
People simply "learn" many things once; and many of those learned items remain there for a pretty long duration, and in many cases even for the rest of their lives. This seems to happen without any repeats, at least without any apparent or conscious efforts to remember or re-remember those memories. This is the case for a lot of semantic information (especially the information of interest or importance to the person) as well as a large portion of the contents of episodic memory.
Why and how is this kind of LTP possible?
Perhaps attention plays a major role here, e.g., being interesting and important automatically triggers LTP without a further need for repeats.
But such effortless long-term memorization happens also in the case of a lot of semantic information or autobiographical events that are not inherently interesting or significant to the person.
Is LTP even needed for all sorts of synaptic plasticity and long-term memory formation?
I prepared the attached consideration of your question.
Although this document shows that sharper memory can be produced by a biological variance or mutation governing the density of white matter, the issue of memory creation is that 2/10ths of a second produces a very clear and permanent engram without further repetitions.
re-encountering the triggers to memory merely adds access to the existing memory since, while the new access is experienced the trigger restores the engram neurons to active, and they bind in with the new active neurons - expanding the existing engram, and especially expanding access to it.
Trying to change the memory is impossible, however it is possible to create new memory and integrate it but alternative contexts are needed.
Visual + auditory + tactile + proprioceptive cues have to be configured to help the new memory establish an alternative engram dimension, to counteract the sharp Mega memory instance that is erroneous.
There are different types of memory formation, and not all of them require LTP. For example, some forms of memory can be formed through a single exposure to a stimulus, such as classical conditioning (e.g., associating a bell with food), or through a single episode of learning (e.g., remembering a phone number). These forms of memory formation are thought to rely on different mechanisms than LTP.
However, for more complex forms of memory, such as long-term declarative memory (i.e., memory for facts and events) or spatial memory, LTP is thought to play a crucial role in their formation and maintenance. LTP involves the strengthening of connections between neurons in response to repeated activation, which allows for the creation of stable and long-lasting memories.
While it is true that attention and interest can facilitate memory formation, it is also the case that the brain is constantly processing and encoding information, even when we are not consciously aware of it. This process, known as incidental learning, allows for the acquisition of information without explicit effort or attention. Additionally, factors such as emotional arousal, novelty, and relevance can also enhance memory formation and consolidation, even for information that may not be inherently interesting or significant to the individual.
Overall, while LTP is not the only mechanism involved in long-term memory formation, it is an important one for complex forms of memory, and factors such as attention, emotion, and relevance can influence the ease and strength of memory formation.
Agnieszka Matylda Schlichtinger
According to my findings all memories are equally strong, i.e. no single axon-branch to dendrite mini-synapse which is involved in memory formation is larger or stronger than any other one, however, access to a memory engram which is a pattern of the same cortical neurons, can be wider, or more easily encountered, making the memory appear to be stronger.
Each time the memory engram is reactivated, new bindings are made that are contextually relevant, so you could say that each relevant context in which the memory is revisited, makes it more accessible.
I would hesitate to say stronger.
A similar error in description of memory is the distortion afforded by an emotional factor. I.e. when a person has a strong emotion associated with a memory, or pain, or any extreme feeling positive or negative, it is assumed to be a stronger memory.
The memory itself is not stronger, however each time it is revisited, it becomes more accessible by the context in which it is revisited.
Jerry Waese
Your findings are largely correct. There is no evidence that any single axon-branch to dendrite mini-synapse involved in memory formation is larger or stronger than any other one. Memories are not stored in any single location in the brain, but rather involve patterns of neural activity distributed across multiple brain regions.
Access to a memory engram can indeed be wider or more easily encountered, making the memory more accessible. Each time a memory is reactivated, new bindings are made that are contextually relevant, which can make the memory more easily retrieved in the future. This is known as the encoding specificity principle.
However, it is also true that emotions can play a role in memory formation and retrieval. When a person has a strong emotional response to an event, it can lead to more elaborate encoding and stronger associations between different elements of the memory. This can make the memory more easily retrieved in the future, especially in contexts that are emotionally similar to the original event.
So while it is true that the emotional factor does not necessarily make the memory itself stronger, it can influence how easily the memory is retrieved and the context in which it is most easily accessed. Overall, the strength of a memory is a complex and multifaceted phenomenon that involves many different factors, including the context in which it is encoded, the emotional associations it carries, and the frequency and depth of its reactivation over time.
Agnieszka Matylda Schlichtinger Vahid Rakhshan
I am embarrassed to admit that I rapidly prepared the pdf this morning without checking that the word "Cures" managed to replace the word "Cues" - I meant cues, and have no cures to offer from my desk today. I have to learn not to trust my typing.
[edit : I re-uploaded the pdf after changes made and I included your medical terms]
As regards strength of memory. You can have a strong feeling and a compelling emotional response, but all memories are equal in strength, even those that include strong feelings and experiences with powerful life impacts.
Otherwise we are only discussing the relative accessibility of memories - which always amount to the relative abundance of perception cues for the engram to be reactivated. Usually that amounts to a wide variance in contexts where the pattern is encountered.
e.g. changing rooms, changing positions of people in a room, changing colors, fonts, orientations, sitting standing walking etc. Learning something corrective requires ingenuity, and avoiding uncorrectable trigger contexts also helps, as well as having remedies for errors in your back pocket. Otherwise the clarity of memory is a gift.
Also with Hyperthymesia, enlarged thalamic regions are noted, and this suggests companion Cortical neurons are increased as well, which amounts to more Cortical Neurons in the brain altogether and that adds up to more Neurons that can be active synchronously which helps differentiate memories better as well. None of that seems to be a drawback as far as I can tell except that it is slightly non-normative, and the whole thing can easily go to one's head that they are smarter than the other family members etc...
Jerry Waese thanks a lot. I am reading your comments with a lot of interest, but I guess you are answering my other question here, right?
Care to copy your answers there too (so that future readers of that question can read them)?
I mean as answers to this: https://www.researchgate.net/post/What_is_this_curious_form_of_non-updatable_mega_memory
Vahid Rakhshan - feel free to quote me there, I got quite excited about the question and was motivated to tinker again with my project; and to use the results as an answer to your question, and the seed to begin examining other's findings about Hyperthymesia.
that an enlargement of thalamic areas exists is critical. Since each thalamic relay neuron has a primary cortical neuron partner that engages in feedback with it supporting synchronous memory formation and the perceptive reflex.
Therefore more thalmic neurons means more primary cortical neurons and overall more potential cognitive activity - larger more "vivid" engrams etc.
Agnieszka Matylda Schlichtinger Thanks for your response. My arguments and further response:
1. Classical conditioning as well relies on LTP. It is not free of LTP. Also, it usually does not happen in only 1 session.
2. Remembering phone numbers is a form of semantic memory, i.e., declarative memory. It too relies on LTP.
3. Incidental learning is not what you said. It is when one learns something indirectly while not attempting to learn that particular thing (and while doing something else). It is totally different from what you explained.
4. The two last paragraphs of your first comment read:
"While it is true that attention and interest can facilitate memory formation, it is also the case that the brain is constantly processing and encoding information, even when we are not consciously aware of it. This process, known as incidental learning, allows for the acquisition of information without explicit effort or attention. Additionally, factors such as emotional arousal, novelty, and relevance can also enhance memory formation and consolidation, even for information that may not be inherently interesting or significant to the individual.
Overall, while LTP is not the only mechanism involved in long-term memory formation, it is an important one for complex forms of memory, and factors such as attention, emotion, and relevance can influence the ease and strength of memory formation."
I find the above 2 paragraphs (except the word "incidental learning") quite convincing.
You are basically suggesting that despite the lack of any conscious efforts to reinforce the memory engram, the brain is actually trying to reinforce the engram behind the scene without the person's knowledge.
If this is true, well this means that the brain is forming LTP for all such memories in the background through some unconscious repeats and reinforcements. In other words, LTP is still happening.
Thanks a lot.
Agnieszka Matylda Schlichtinger I forgot to add that these items you listed contribute to attention:
"...Additionally, factors such as emotional arousal, novelty, and relevance can also enhance memory formation and consolidation..."
Arousal, novelty, and relevance all increase attention.
Vahid Rakhshan
Thank you for your comments and feedback. I appreciate the opportunity to clarify and further discuss the concepts of memory and learning.
Ad 1. You are correct that classical conditioning does involve LTP, particularly in the formation of new neural connections between the conditioned stimulus (CS) and the unconditioned stimulus (UCS) pathways. However, the point I was trying to make is that LTP is not the only mechanism involved in classical conditioning. There are also other forms of synaptic plasticity, such as long-term depression (LTD), that play a role in modifying neural connections during the process of learning and memory.
Ad 2. I agree with you that remembering phone numbers is a form of declarative memory, specifically semantic memory. As you noted, this type of memory also involves LTP, particularly in the hippocampus and related brain regions. My point was that not all forms of memory formation necessarily require conscious effort or attention, and that the brain can process and encode information even when we are not actively trying to remember it. It should therefore be asked to what extent these mechanisms are actually based on the LTP?
Ad 3. I apologize for any confusion regarding incidental learning. You are correct that incidental learning refers to the acquisition of knowledge or skills that occur unintentionally, while engaged in some other activity. This is different from implicit learning, which refers to learning that occurs without conscious awareness of the information being learned.
Ad 4. I'm glad that you found the last two paragraphs of my initial response convincing.
You also wrote:
"If this is true, well this means that the brain is forming LTP for all such memories in the background through some unconscious repeats and reinforcements. In other words, LTP is still happening.".
And here I have to admit that I just don't know. I am not convinced enough and I would ask you to specify this part if you could. And still also relevant is the question "To what extent is LTP still happening?".
And I will ask additionally: Do you recommend any publications on this topic?
Still returning to your question that started the discussion...
Maybe this is such that it is not possible for humans to memorize many things instantly and permanently without any reinforcement. The brain requires repetition and reinforcement to encode and consolidate memories into long-term storage. However, there are some techniques and strategies that can help people improve their ability to memorize and retain information, such as:
1. Using mnemonics: Mnemonics are memory aids that help people remember information through associations, such as acronyms, rhymes, and visual imagery.
2. Practicing active recall: Active recall involves actively retrieving information from memory, rather than simply reviewing it passively. This can be done through techniques such as flashcards, quizzes, and self-testing.
3. Creating associations: Linking new information to existing knowledge or experiences can help make it more memorable and easier to retrieve later.
4. Using spaced repetition: Spacing out the practice of new information over time, rather than cramming it all at once, can help with long-term retention.
5. Getting enough sleep: Sleep is important for memory consolidation and retention, so getting enough sleep can help improve memory.
Overall, while it may not be possible to memorize many things instantly and permanently without any reinforcement, practicing good memory techniques and strategies can help improve one's ability to memorize and retain information over the long term.
Thank you again for your comments and questions, and I hope this further clarifies the concepts of memory and learning.
As far as I can see, nothing happens to memory while unconscious, either alpha (Awake) or theta (REM Sleep) rhythm is required, which specifically means there must be cortico-thalamic feedback. Using the term "unconscious" for anything other than Non-REM sleep or anaesthetic unconscious state will lead to error in understanding mind.
Even the term "subconscious" should be qualified and defined on the spot as it also suggests incorrect meanings. Every 1/10th of a second mental contents can change, much of what is considered subconscious is just rapidly and partially formed sensation or perception.
Subliminal is a real cortical event quality for those events that happen in intervals that are faster than the alpha rhythm as local clusters of interneurons perform localized voxel processing prior to editing allowing specific cortico-thalamic loops to continue. This is extensive in the 6-layer cortex. Brain processing that is faster than alpha rhythm is all evolved via millions of years of adaptation; it is not acquired by learning or affected by associative memory. Some of this high speed processing uses localized potentialization or temporary cellular memory which is not related to the plasticity or associative processes.
Memory interlinks (associates) all synchronous active mental contents every 1/10th of a second. While unconscious, there are no active mental contents.
Mental contents that are unchanging disappear from the active set which is an adaptation that enables attention to follow changes (threats and opportunities).
Otherwise the term "Attention" is not clearly defined, but we know that Hypothalamic suppression of (top down) thalamic feedback can be controlled by the cortex - so that enables attenuation of uninteresting mental contents, even if they are changing and vying for attention.
Emotional arousal etc. affect the duration of cortico-thalamic feedback - usually this is a Hippocampus effect that can increase loops beyond the standard 3-cycles from the thalamic end. I presume some emotional chemistry gives the Thalamic neurons more vigor. Psychedelics do a similar thing on the cortical side, extending the cortico-thalamic looping -> which produces distorted time effects and richer sensation and ideation/perceptive reflexes.
Thanks a lot Agnieszka Matylda Schlichtinger for your clarifications and interesting points. Let me reply point-by-point:
Ad 1. I understand now and agree that LTD goes hand in hand with LTP. I should have elaborated on LTD in my original question. I considered LTD as a given and hence did not even mention it. So I agree that LTD is intermingled with LTP.
However, this is in all cases of memory formation and not just in conditioning.
Ad 2. Your second point is apparently about implicit memory formation (which also includes classical conditioning). I think implicit learning too depends on LTP.
Article Barco A, Bailey CH, Kandel ER. Common molecular mechanisms i...
Ad 3. Thanks for your clarification. I now understood what you meant in that whole paragraph.
Ad 4. Thanks again. You asked:
And here I have to admit that I just don't know. I am not convinced enough and I would ask you to specify this part if you could. And still also relevant is the question "To what extent is LTP still happening?".
Well, it was my misunderstanding of your statement (when you were talking about implicit learning). It was not as if I believed so or had any evidence for it. I was just recapping my understanding of your words.
Apologies for any confusion I made. :)
But as an answer to your question, actually during memory consolidation, this back-and-forth repeat and reinforcement happen between the hippocampus and the neocortical regions. This happens mostly during sleep. Its purpose is to consolidate the memory temporarily stored in the hippocampus during the day into more long-term and more organized networks of the cortex. So yes, such unconscious reinforcements really happen.
But the issue is we see long-term memories formed in sleep-deprived people, even after only 1 exposure. Sleep-deprived people do not have a great chance of such "system consolidations".
So I guess perhaps there are other forms of unconscious reinforcement going on during the wakeful state.
And another possibility is that LTP might not always need repeats. There are different types of LTP, some of which may occur with a single exposure.
However, it is still LTP.
Agnieszka Matylda Schlichtinger
Regarding your second question: "And still also relevant is the question "To what extent is LTP still happening?"
I am not sure if any long-term memory can be free of LTP/LTD. So my guess is that any persistent memory has LTP in it.
Regarding this question of yours: "And I will ask additionally: Do you recommend any publications on this topic?"
Well perhaps the best publications are neuroscience textbooks like Kandel and Gazzaniga. But review articles and even Wikipedia have very good articles too.
Article Barco A, Bailey CH, Kandel ER. Common molecular mechanisms i...
Article Long-term potentiation and long-term depression: A clinical ...
https://en.wikipedia.org/wiki/Synaptic_plasticity
https://en.wikipedia.org/wiki/Long-term_potentiation
"The pre- and postsynaptic activity required to induce LTP are other criteria by which LTP is classified. Broadly, this allows classification of LTP into Hebbian, non-Hebbian, and anti-Hebbian mechanisms. Borrowing its name from Hebb's postulate, summarized by the maxim that "cells that fire together wire together," Hebbian LTP requires simultaneous pre- and postsynaptic depolarization for its induction.[25] Non-Hebbian LTP is a type of LTP that does not require such simultaneous depolarization of pre- and postsynaptic cells; an example of this occurs in the mossy fiber hippocampal pathway.[26] A special case of non-Hebbian LTP, anti-Hebbian LTP explicitly requires simultaneous presynaptic depolarization and relative postsynaptic hyperpolarization for its induction.[27]"
The above paragraph and other excerpts from the above links answer my original question to a great degree: They show that LTP can happen (at least in some cases) with one exposure, if strong enough.
---------
Regarding your comments on how to improve memory, thanks a lot for sharing them. They are very helpful.
Jerry Waese thanks a lot for all your interesting answers. I would quote and answer some in the other thread and some here:
"According to my findings all memories are equally strong, i.e. no single axon-branch to dendrite mini-synapse which is involved in memory formation is larger or stronger than any other one, however, access to a memory engram which is a pattern of the same cortical neurons, can be wider, or more easily encountered, making the memory appear to be stronger."
Neuroscientists share the same opinion that memory recall (access to engrams) is extremely important. I have heard from various neuroscientists, neurologists, and psychiatrists that when you forget something, you have actually lost access to that memory. But it is still there, somewhere.
Of course, this is to a great extent is not based on evidence. It is mostly a hypothesis rather than some proven scientific fact.
But studies with hypnosis and some brain anomalies (such as disruptions due to tumors or concussions) show that some forgotten memories can be remembered under the right circumstances.
But to say that "all memories are equally strong" does not align much with scientific findings. I mean not only memory retrieval can have different strengths, but also can memory encoding.
It is a very old neuroscientific finding that synapses do have different weights, and it is the topology and number of those synaptic weights that form the memory. And of course, those synaptic weights have been simulated within artificial neural networks.
So I wonder how your findings suggest otherwise.
Regarding your next paragraphs: "Each time the memory engram is reactivated, new bindings are made that are contextually relevant, so you could say that each relevant context in which the memory is revisited, makes it more accessible.
I would hesitate to say stronger.
The memory itself is not stronger, however each time it is revisited, it becomes more accessible by the context in which it is revisited."
I would say both stronger and more accessible:
(1) They really get stronger, even from a cellular morphological standpoint. There, proteins get built up and synapses get broader. More synapses may be created. The number of receptors and the amount of built neurotransmitters increase. etc. So it really gets stronger.
(2) "More accessible" too means "stronger" itself: For an engram to get more accessible, the synapses between the "memory retrieval network" and the memory engram should become stronger.
So if I am missing or if I misunderstood your point, please tell me.
------------------
A similar error in description of memory is the distortion afforded by an emotional factor. I.e. when a person has a strong emotion associated with a memory, or pain, or any extreme feeling positive or negative, it is assumed to be a stronger memory.
I am not sure (1) if this is a distortion or a real effect [it is possible that emotional connotations really strengthen memory], and (2) if this is similar to what you said in the previous paragraphs about all synapses being similar. Care to elaborate?
Agnieszka Matylda Schlichtinger you said to Jerry that "There is no evidence that any single axon-branch to dendrite mini-synapse involved in memory formation is larger or stronger than any other one."
Are you sure? If this is true, the whole notion of synaptic weighting and LTP should be thrown out of the window.
Vahid Rakhshan
I apologize for any confusion my previous response may have caused. Let me clarify:
There is evidence that some synapses are stronger than others, and that synaptic strength can change over time through a process known as synaptic plasticity, which includes long-term potentiation (LTP) and long-term depression (LTD). These changes in synaptic strength are thought to be important for learning and memory.
However, it is not clear that any single axon-branch to dendrite mini-synapse involved in memory formation is larger or stronger than any other one in general. The strength of a synapse is determined by a variety of factors, including the amount of neurotransmitter released, the number and sensitivity of receptors on the postsynaptic membrane, and the properties of the presynaptic and postsynaptic neurons themselves. These factors can vary widely from synapse to synapse, even within the same neural circuit.
So, while there may be differences in synaptic strength between individual synapses, it is not accurate to say that any single synapse is always larger or stronger than any other one, or that the strength of a synapse is solely determined by its size. The idea of synaptic weighting and LTP/LTD remains an important and valid concept in the field of neuroscience.
On the other hand, I am also, among other things, a philosopher, a clinical psychologist and a theoretical physicist. I tend to look at commonly accepted definitions and paradigms from many different perspectives.
I worked as a neuroscientist for 2 years at the Medical University of Wrocław, I previously studied neuroscience at Duke University. I did not find answers to my questions. On the one hand, we are looking for something, while on the other hand, our accepted definitions and paradigms often lead us to contradictions.
What interests me are noumena (things in themselves). I look for them in everything, although they can be a difficulty in typically technical discussions. Nevertheless - I cannot believe in any theory, concept or description if I do not find assumptions deep enough in them. The foundations are often fragile.
Neuroscience is one of my favourite disciplines of knowledge, but I nevertheless think that its foundations are not clear enough. I think the same about theoretical physics.
Hence, I warn you that what I write may be risky at times, but I take part in this discussion because I myself very much want to understand and be able to reflect further.
Jerry Waese you made very interesting points about conscious and unconscious states. Thanks a lot.
I agree that one should clearly define the state of being unconscious. However, when saying "unconscious", I didn't mean "while being unconscious". I was meaning "not by the self-aware (conscious), active, and top-down effort". I think it is standard medical terminology and already defined, agreed upon, and commonly used.
This encompasses a wide array of tasks done by the reptilian brain, limbic brain, and some parts of the neocortex.
For example, we breathe unconsciously, i.e., without active, unconscious, or top-down effort. Our heart beats unconsciously, i.e., without any effort from us. It doesn't mean that our heart beats when we are unconscious. It means that it beats regardless of our conscious mind and top-down attention; hence, unconsciously.
So I didn't mean "while being unconscious" by words like "unconsciously" etc.
------------------
Let me argue or discuss some of the details in your comments. By the way, thanks a lot for these thought-provoking comments.
As far as I can see, nothing happens to memory while unconscious, either alpha (Awake) or theta (REM Sleep) rhythm is required, which specifically means there must be cortico-thalamic feedback. Using the term "unconscious" for anything other than Non-REM sleep or anaesthetic unconscious state will lead to error in understanding mind.
Neuroscience textbooks and articles do not agree with most of this paragraph.
1. Memory can be formed at almost all stages of wakefulness and sleeping if not all of them.
2. Unless in particular cases like seizures, there is always a mixture of different brainwaves. More than one brainwave usually exists, depending not only on the state of mind but also on the region of the brain we are talking about. So it is generally a very complicated (and unknown) mess, rather than the commonly simplified version of dominant waves.
1. Being unconscious does not mean the absence of any brainwaves. Clinically speaking, one is unconscious when she lacks any awareness and can't respond. This includes non-REM sleep, hypnosis, deep sedation, general anesthesia, etc.
2. Even REM sleep can be considered somehow an unconscious state because the person has no sense of agency (unless she is Lucid-dreaming) and also can't respond to external stimuli. She is even paralyzed. For being considered fully conscious, one needs to be able to respond to stimuli. This needs both a sense of agency and motor control.
3. Even in non-REM sleep, one may dream.
4. Brainwaves continue to exist even in seizures, in the deepest forms of general anesthesia, in the deepest states of sleep, or even in the deepest forms of coma.
5. Alpha wave is not for wakefulness; it is for the "awake with closed eyes" state. So if you think it is necessary for memory formation, you should double-think it because the normal wakeful state is characterized mostly by beta.
Jerry Waese
"Even the term "subconscious" should be qualified and defined on the spot as it also suggests incorrect meanings. Every 1/10th of a second mental contents can change, much of what is considered subconscious is just rapidly and partially formed sensation or perception."
I am not sure if the term subconscious is even a scientific term. I think it is used in psychoanalysis and has other meanings that are not directly related to clinical consciousness or to the neuroscientific notion of mind, which is the subtext of our discussion.
By the way, if you are interested, please leave some good comments in this thread too:
https://www.researchgate.net/post/Consciousness_Cognition_and_the_Unconsious_Mind
------------------
I couldn't understand much of the below paragraphs except the sentence in bold:
Subliminal is a real cortical event quality for those events that happen in intervals that are faster than the alpha rhythm as local clusters of interneurons perform localized voxel processing prior to editing allowing specific cortico-thalamic loops to continue. This is extensive in the 6-layer cortex. Brain processing that is faster than alpha rhythm is all evolved via millions of years of adaptation; it is not acquired by learning or affected by associative memory. Some of this high speed processing uses local temporary memory which is not related to the plasticity or associative processes.
Memory interlinks (associates) all synchronous active mental contents every 1/10th of a second. While unconscious, there are no active mental contents.
Mental contents that are unchanging disappear from the active set which is an adaptation that enables attention to follow changes (threats and opportunities).
------------------
Otherwise the term "Attention" is not clearly defined, but we know that Hypothalamic suppression of (top down) thalamic feedback can be controlled by the cortex - so that enables attenuation of uninteresting mental contents, even if they are changing and vying for attention.
Emotional arousal etc. affect the duration of cortico-thalamic feedback - usually this is a Hippocampus effect that can increase loops beyond the standard 3-cycles from the thalamic end. I presume some emotional chemistry gives the Thalamic neurons more vigor. Psychedelics do a similar thing on the cortical side, extending the cortico-thalamic looping -> which produces distorted time effects and richer sensation and ideation/perceptive reflexes.
I enjoyed reading these. I don't know if they are your speculations or evidence-based proven facts. But in any case, they were interesting.
Just to add that, as far as I can remember, attention is a very complicated neocortex matter. Not that it does not rely on bottom-up sensory information (i.e., the thalamocortical pathways), but it heavily relies on the prefrontal and parietal cortexes that process the received sensory information.
So in many cases, it is not the thalamus that is suppressed. The thalamus is relaying information to the cortex as before, but the cortex simply chooses to ignore a part or all of the received information.
So if you are interested in attention and working memory, you might find other brain regions and networks even more useful than the thalamus.
Dear Agnieszka Matylda Schlichtinger Thanks a lot for your nice answer. I will reply tomorrow. This Covid is killing me right now!
Vahid Rakhshan
I composed the following for you
addressing some of the points raised in a few above comments.
Forgetting is not an actual action, even if it is a hypnotic suggestion. There is no muscle to forget. No slate wiper. Even “fuggedaboudit” is really a directive to “stop bugging me” or something else.
We do not forget, as a verb, instead we lose access by happenstance, usually by living in contexts (mentally and physically) that do not provide sufficient cues to reflexively elicit reactivation of existing engrams.
It does not mean that the memory is not personally meaningful if it is not accessible, just that everything salient about it is strange (foreign) to the context (different emotional state, different environmental setting, unrelated recent activity (short term memory))
One problematic mental tactic to avoid certain memories is to build a defense behaviour around them such as by tensing the body, or self harm or some distracting action which becomes associated with the triggers to the memory engrams that are troubling, and if the tactic is effectively distracting the memory blockade works.
This is directly related to the idea of getting attention. i.e. noticing a change, and apprising that instead of this.
As you indicated there are a variety of synapse types in the brain. For basic memory formation and the perceptive reflex only two synapse types are fundamental – regular and mini. 1 - The regular synapse is found at the activating axon terminal end where release of a chemical conveys a signal between a thalamic neuron and a cortical neuron forcing activation and dendritic back propagation (the same synapse type is situated at the end of the cortical neuron’s axon which loops back to the same thalamic neuron with which it is paired – establish a feedback loop). And 2 – The pyramidal neuron axon branch that carries 1/5000th of a full charge (or less) to each of it’s cortical neuron targets where mini-synapses exist on the dendrites of cortical neurons. When mini synapses are active at the same time that the cortical neuron is active, ARC spines form. When a plurality of activated mini synapses with spines on a resting cortical neuron are activated, the resting cortical neuron may become active - For a resting cortical neuron perhaps 11 mini synapses with ARC spines must be activated to make the cortical neuron active; but if the neuron was recently active – perhaps only 7 activating mini-synapses with ARC spines will suffice to make the neuron fire (short term memory operates like this).
Mini-synapses without ARC spines do no cause any activation on resting neurons. The spines at mini-synapses are the real world memory linkages. In the case of mini-synapses with spines the micro-charge from a pyramidal axon branch is a signal, and when enough of these signals arrive the cortical neuron makes itself fire (similar but different from neurotransmitters at a regular synapse). Inside the neuron is some facility for sensing that the threshold has been met and changing the membrane polarity (aka firing).
In consideration of the duration of a living brain at ~100 years, it makes sense for older cortical neurons to require a higher threshold of spine activation to trigger, as the number of spines marking memory engrams is always increasing for each cortical neuron.
For active mini-synapses with spines, yes there is a synaptic weighting, but for the other memory related synapses (eg. the C-T loop) there is only fire and react and no partial signalling.
Age related loss of memory which is not due to plaques, injury or disease is a matter of access that is weakened due to the raised thresholds of activation (number of active spines per neuron). As we age each cortical neuron collects more spines, and if the threshold is not raised, false firing and confused reflexes occur (i.e. unfortunate thought and behaviour).
You said :
My response to that is negative.
With regard to the idea of mini-synapses with spines growing larger and stronger, this makes no sense. There is no physical room for it, and no functional need for it. It was Hebb’s original proposal (and he never asked me to correct him across the barrier of decades) but it is not scientifically proven. Introducing it into the model only adds complexity and confusion.
More accessible with regard to memory only means that there are more triggers that will reflexively reactivate the engram (including thought, speech, and action). The more pervasive network you mention more reliably enables engram reactivation, but it is misleading to use the term stronger. There is no good place for it in the model I present.
Emotional states are resonant (altered C-T feedback Loop duration (usually beyond 3 and less than 30 loops)) states with specific hormonal affects, similar to dreaming and psychedelic use, greater intensity of experience is perceived. That perception becomes part of memory, but the memory is not stronger or weaker than other memory it just contains the more impressive feel as part of the contents. All memory is reflexive, some of the reflex content can be thoughts, speech, actions and even emotional behaviours.
A memory content with violence is not stronger than a tranquil memory, although it will be more disruptive in the stream of consciousness, and it may get more attention and cause more changes. We need to be specific about what we are saying, and strength of memory has no place in this model which works very well without relative strength between memory engrams and the triggers that invoke them.
Although Computer Models may claim that they are based upon human neurons, most of the simulated neurons and synapse arrangements in AI are not like those we have in our brains which are involved in memory formation and perception. However the interneuron assemblies in the six layer cortex which can edit, divert, shunt or silence C-T loop activity, do resemble somewhat the kinds of neurons that the AI industry has run with. Our six layer cortical processing is important but not fundamental to the critical operation of memory formation and perceptive reflex generation.
I cannot address all the places where you have stubbed and will stub your mental toes on this journey through the mind, but one thing to remember is that "what fires together wires together" (that is the good part of the original Hebbian memory model). Firing together is also called synchronous firing. it is something that is marshaled and sustained by C-T feedback at the speed of 10hz more or less (alpha and theta wave action is C-T feedback driven).
All active mental contents sensory or imagined are sustained by C-T feedback, and those are in synchrony enough to perform the miracle of continuously forming memory engrams.
Jerry Waese thanks for your detailed explanations. I can't tell that I understood 100% of them, but those parts I understood were very interesting to me.
I really would love to argue some of your points. But I can't right now. Just one question: Is all of this your own model? Or can you give me some relevant references to read?
I mean these in particular:
With regard to the idea of mini-synapses with spines growing larger and stronger, this makes no sense. There is no physical room for it, and no functional need for it. It was Hebb’s original proposal (and he never asked me to correct him across the barrier of decades) but it is not scientifically proven. Introducing it into the model only adds complexity and confusion.
More accessible with regard to memory only means that there are more triggers that will reflexively reactivate the engram (including thought, speech, and action). The more pervasive network you mention more reliably enables engram reactivation, but it is misleading to use the term stronger. There is no good place for it in the model I present.
my model.
I never met Hebb, but his comments inspired me when I was studying in university in the 70's.
in the 70's I referred to the spines as scars (ARD deposits were not yet known or revealed to me). scars or spines, pretty much the same difference.
nevertheless, if memory formation can occur within 1/10th of a second, and repetition added molecules to the spine, within a few thousand repetitions over just a few days you would have lumps that are larger than the neurons themselves.
I jokingly refer to that as Hebbian plaque - the idea of accretion of plaque to strengthen memory makes no sense, and is not required since recollection is a reflex based upon active mental contents lining up somewhat exactly to reactivate associated mental contents.
All of it is reflexive.
Hebb did not consider the reflexive potential of the spines at the ends of branches of pyramidal neurons.
I am not aware that anyone to date except me has put it forward.
Nor has the centrality of the C-T loop penetrated to the core of anyone else's model.
Because of that I put together my github project and I am just poking the embers of that little fire to see who catches on.
What has fallen into place in the last 4 years for me is that the alpha and theta rhythm and cycle timing are artifacts of the C-T loop, and that they disappear in NON-REM sleep but are otherwise part of all conscious states.
What has also fallen into place is that at least 2 cycles of C-T are required for memory formation - the second is to reactivate the primary cortical neurons so that the pyramidal branches (activated on the first pulse) reach primary cortical neurons in the second pulse. this is more than moot since, if the hypothalamus is suppressing Top Down actives in a section of the thalamus - then the second pulse in that section will not occur, the feedback will be swallowed and suppressed. This becomes important while learning something new, where what was active 1/10th of a second ago is silenced to allow a clearer impression. I ran into that during development of the demo as well.
Too many things about the demo work in the right way to be dismissed, so I am thinking of small tweaks and writing bits and pieces of text and waiting for people to find it and use it.
Thanks Jerry Waese for the nice explanations. I see. I though you are in your 20s. :)
I enjoyed reading your comments, especially the parts related to brain waves. But I am not sure if I can follow some of your arguments and premises. For example, I am not sure how the hypothalamus is supposed to suppress top-down pathways. To the best of my knowledge, the role of the hypothalamus doesn't include that one.
nevertheless, if memory formation can occur within 1/10th of a second, and [if] repetition added molecules to the spine, within a few thousand repetitions over just a few days you would have lumps that are larger than the neurons themselves.
Why should we assume such a linear function between repetition and synaptic structure?
Eating adds molecules to the body. So if your above logic applied, after eating a few thousand food servings, we would become giants. And even we would continue to grow non-stop till the end.
But the body doesn't continue to grow larger and larger till death in such a linear pattern.
Agnieszka Matylda Schlichtinger Thanks a lot for your explanations.
However, it is not clear that any single axon-branch to dendrite mini-synapse involved in memory formation is larger or stronger than any other one in general.
V: I now see what you meant. Thanks for clarifying. I am curious about what you are proposing. I guess I have read somewhere a couple of years ago that it is now found that computations happen not only at the synaptic level but also at much smaller scales, i.e., at dendrites.
I will try to find that paper. https://news.mit.edu/2019/neurons-dendrite-role-computation-0606
If that is the case, perhaps it is also possible that those small branches of a single synapse might have heterogeneous properties. But this is my own guess and speculation. I don't know if there is any evidence for or against it.
-------------
So, while there may be differences in synaptic strength between individual synapses, it is not accurate to say that any single synapse is always larger or stronger than any other one, or that the strength of a synapse is solely determined by its size.
V: I agree but did I ever use the words "always" or "solely"? Even I talked about how more receptors or more neurotransmitters can contribute to LTP.
------------
On the other hand, I am also, among other things, a philosopher, a clinical psychologist and a theoretical physicist.
V: That's very impressive.
What interests me are noumena (things in themselves). I look for them in everything, although they can be a difficulty in typically technical discussions.
V: I really don't know what you mean and how it relates to memory. Do you mean you care about the objective truth out there verified by evidence?
Nevertheless - I cannot believe in any theory, concept or description if I do not find assumptions deep enough in them. The foundations are often fragile.
V: I agree that the assumptions and premises should be sound and logical. I don't know what you mean by "deep enough" but I guess you mean something like very accurate or something like that.
Neuroscience is one of my favourite disciplines of knowledge, but I nevertheless think that its foundations are not clear enough. I think the same about theoretical physics. Hence, I warn you that what I write may be risky at times, but I take part in this discussion because I myself very much want to understand and be able to reflect further.
V: These are empirical sciences. Of course their foundations cannot be accurate enough and will be subject to continuous revisions and corrections. I can't imagine otherwise.
This paper is very relevant to this thread:
https://www.nature.com/articles/s41598-022-10466-8
Vahid Rakhshan I do not fully understand that nature article where they purport to use a living neuron as a Hebbian memory/learning device, and raise the branching dendrite model to a neural network decision tree - "learning" without the production of spines or boutons. Note the test signals (probe charges) were 0.2 ms which is possibly significant for interneuron 6-layer cortical signal processing, but not the slower 10 hz alpha rhythm rate that accompanies memory formation.
Vahid Rakhshan the neuron and fibre and glia packing in the brain is quite tight, so there is little room for intercellular roving maintenance cells.
while in our body, what we eat goes partly to growth and repair and otherwise energy and fluid management, we also excrete waste.
if the spines or boutons kept growing, a roving maintenance cell would have to groom them or plaque would develop around all the mini-synapses, basically the brain would fill up with waste.
there is no evidence of roving maintenance neurons of this sort, nor any proposal of how or when cleanup would be applied. the idea of stronger and weaker mini-synapses is wrong.
This is about the effect of the state of consciousness on the memory. In dream our state of mind shifts and the memory of our REM dreams become less available to our current (daytime) state of mind. On the other hand, it becomes more available to the similar state of mind when falling asleep. This is why I for one can remember easily the dream of last night, while I am falling asleep the next night.
The strong association is something agreeable for ANY memory, including a dream. This is another "independent" variable which exists simultaneously with what I described above (as another independent variable).
https://www.researchgate.net/post/Why_cant_we_easily_remember_our_dreams_Why_can_we_remember_the_last_nights_dream_better_during_falling_asleep
Yes, Vahid Rakhshan , In my terms that is the effect of associative access to the engrams, not necessarily the speed of the original synchronous association, nor the speed at the time of the perceptive reflex of the synchronous association, nor the strength of the memory engrams of the dream - entirely it is about being in a familiar enough context to perceive the memory.
Just the contents have to have a salient correspondence, like the bodily sensations while drifting off.
Another thing happens during "narcosis" dreaming and emotional states the C-T Loop is extended - note that the resonant C-T (cortico-thalamic) loop for normal states of mind is 3 cycles = 3 x 1/10th second:
so the first 1/10th of a second has overlap of the previous interlinking pulse (and is linked into the tail of the previous, the second performs it's own interlink of current synchronous activation, and the third leaves a tail to be linked into the next synchronous activation.
when this sequence is extended, the sense of time passing is lost due to the stacking up of synchronous moments, but the each sensation seems richer. more psychedelic and dreamlike, with chimeric mixes of mental objects.
Emotional sequence extension has an adaptive advantage which my demo helps to reveal:
https://jerrywaese.github.io/perception/
for instance - using a small sample and only 3 Thalamo-Cortical Feedback loop cycles we get a lower gross perceptive reflex reactivation than with 8 loop cycles.
as shown in the illustration attached by extending the sequence to 8 50% more clarity of the original engram is achieved from a small stimulus sample.
Jerry Waese said:
It seems my last post got eliminated, It was so cool and I did not keep an offline copy. here goes take 2...
Dream recall is about the relevance of context for access to memory,
so as you drift into dreamland, you enter a context with many equivalences to the dream state onset, and in that context you can more easily access memory from dreams with similar mental contents. Including the prevalent rates of C-T looping.
I prepared a demo image of normal C-T looping with 3 loops compared to extended looping (narcosis, dreaming, emotional states) in which a more extensive perception is possible from a very small sample. you can see that 50% more of the original image engram emerges from the same sample after 8 loops compared to the standard 3 loops
this provides a sense of sequence to associative memory - every frame has a link to the next (and previous).
when the loop length is extended there is a loss of sense of time passing and the onset of the sense of timelessness. there is also intensification of sensation and ideation, and chimeric mixing of mental objects.
Vahid Rakhshan
thanks for quoting me - that last quote refers to this Demo program
https://jerrywaese.github.io/perception/B5000.html
which any browser will reveal and two settings for the C-T loop length.
it also needs this image to make sense
I guess I am mixed up between discussion threads that are about similar topics
Jerry Waese thanks a lot for your nice software and the image. I will try to play with it and understand what is what.
Quoting Jerry Waese from the other thread:
Vahid Rakhshan after my breakfast I will boot up my dev box and adjust the initialization codes in my demo at https://jerrywaese.github.io/perception/
(a) to increase the number of branches per pyramidal neuron from 2000 to 5000 (for more binding or ARC protein/plasticity), and (b) to increase the number of synchronously active spines from 7-20 or something (- can be managed in the interface (needs no new coding)).
I will report back the findings with illustrations and then expand on how it presents in real life memory.
A greater number of branches means more connection potential overall, but also can lead to mistaken perception, i.e. bringing in loosely associated bits (also it slows down the demo).
A greater threshold on number of synchronous active spines to activate a resting neuron means that a more accurate correspondence to original is expected.
If the results look interesting I will report with screen caps on the question threads you posted.
Quoting Jerry Waese from the other thread:
Vahid Rakhshan Thanks for checking it he gitHub project where I am trying to model neural connectivity in the cortex structurally and functionally.
The Function activatePyramidalsUponBackPropagation() takes field energy of activated corticalNeurons and then drives activateAxonBranchesOf(id_of_Pyramidal_neuron)
where for each branch a different result follows
if the target Cortical neuron is active then an ARC protein spine is crated (a new long term memory element is added)
otherwise It leaves a local charge on the resting neuron mini-synapse in the function called tryPerception();
There if the threshold is met for active spines the resting Cortical Neuron is activated.
if the resting neuron was recently active fewer than the usual threshold number of active spines are required to reactivate (short term memory effect).
The demo uses images in the HTML5 Canvas security web model so it needs a web server to host those images. I accepted this as a reasonable trade off for keeping the code small without using any 3rd party libraries to develop the core concepts.