Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Featured Are memories made of this or that?

  1. Jul 24, 2016 #1
    I've been reading a little about memory - the various kinds and how it works. Something immediately springs to mind and I haven't found a clear answer. But then I'm not sure if I am totally missing the point and misunderstanding things.

    When we perceive a scene, we detect sensory stimuli which are passed through processing routines in the brain. However these representations seem to be purely mechanical in nature, and it isn't until we become consciously aware of a scene that it gains its experiential quality.

    For example, our eyes respond to lightwaves and pass those signals into our brain, but colour is some kind of internal construction. Indeed, our visual experience is a construction utilising the raw sensory input.

    When I look at a scene and then later recall that scene, I am accessing a memory of that scene. But what is it that is stored in memory for me to access? Is it the sensory input or is it the constructed visual experience?

    That is, do we construct the recalled visual scene in the same way that we construct the original experience (from neural representations of the raw sensory input), or do we recall the constructed scene?
  2. jcsd
  3. Jul 24, 2016 #2

    jim mcnamara

    User Avatar

    Staff: Mentor

    Just as a passing comment: what your eye actually images is a fraction of what is available. Your brain "fills in" details that are not in the current image.
    So, a lot of your mental image is a construction to start with (this is how I construe your meaning of construction), so a lot of your recall is going to be based on construction or "fill".

    Example: you see a really attractive member of the opposite sex. What do you focus on? The mud puddle next to the person, the adjacent brick wall? Answer:
    Probably not. This is a partial explanation of why several different eye witnesses' accounts of an incident can appear to be way out of whack with each other.

    Eye tracking data and mental image (google books so it may have issues for you):
    https://books.google.com/books?id=2...l image from data coming from the eye&f=false
  4. Jul 24, 2016 #3


    User Avatar
    Gold Member

    I have memories of things never seen, things I have dreamed up in my mind that have never existed in real life (as far as I know) I generally find that if I look hard and long someone has talked about or done most of what I have been thinking, but there are things in my mind that are original thoughts that have never seen the light of day from anyone or anywhere.
    Our minds have the ability to build mental images, that we can remember for all of our lives, seeing an actual event is not absolutely required.

  5. Jul 24, 2016 #4
    I think I am getting at something different. The questions of why we don't have one to one recall, or how we can imagine or dream things are all interesting, but I am curious about a specific physical biological process.

    If I view a scene, the light from that scene activates cells in my retina which pass sensory input in the form of neural signals to my brain. Now I know (but don't necessarily understand) about the dorsal and ventral streams of visual processing and so on. That is the mechanical flow of the signals as they are processed within our brains.

    At some point, "something" happens and we become consciously aware of the scene. However, what we are aware of is not necessarily exactly what was perceived (where by perceived I mean the raw input of sensory representations). For example as noted above, our attention will influence what we are aware of yet all of the scene stimulates our retinal cells. Even things directly in our field of interest may not make it to conscious awareness (for example the famous gorilla on a basketball court on Youtube).

    Now, the process of passing the signals through our visual processing systems is mechanical. Without straying into areas of philosophical concern, as far as I know we don't understand for sure how this process actually generates the conscious experience of a scene.

    On a gross simplification, memory depends upon the strengthening of synapses so that a neuron may more easily fire a connected neuron in future, although I gather short term memory is more of a maintenance of connections while long term memory depends on strengthened synaptic connections (a function that the hippocampi seem critically involved with).

    So, my question relates to that strengthening of synapses process. This is a purely physical process and apparently we can observe some aspects of it in action which is why scientists understand some (or much?) of the process of memory formation and recall.

    However if we aren't clear on how conscious experience arises in a purely physical sense, how is that memories of scenes have some of the same hallmarks as conscious experience? When I recall a scene, I tend to recall the same things of which I was aware originally. I recall the same qualia of experience, such as colours and shades. Similarly I don't remember the same things of which I was not aware originally.

    Thus, my question is simply this. What synaptic connections are strengthened? The ones that underlie the representations of sensory signals as they are processed in the visual system, or the ones that represent conscious experience itself?

    Do we store sensory signals, or do we store conscious experience? Do we recall scenes (memories) using the same process by which we construct conscious experience in the first instance, or do we recall the constructed conscious experience itself?
  6. Jul 24, 2016 #5


    User Avatar

    Staff: Mentor

    Anecdote alert. As you get older, you might find that your memories of past events have changed. I have one story I like to tell, it's a fond memory. Well one day I was going through some old things and found a journal where I had actually written down what had happened right after it happened, it surprised me, it was almost nothing like what I now remembered it to be. I never tried to change it, never tried to embellish or leave out details. But there it was in front of me written down as it really happened, and almost nothing like I remembered it.

    I think what might have happened was as I would tell the story, people would ask about little bits that weren't in the story and I'd stop and think we'll maybe that could've happened, and then next time I told the story, that little bit was added or had changed the story. The mind and memory are very strange.
  7. Jul 24, 2016 #6
    Interesting anecdote, and I've experienced that often. I have come to the conclusion that most of what I recall is a narrative rather than anything exact, and the narrative changes with the retelling. Which is why I've been reading about memory, because introspectively it appears to be a constructive process and I am suspicious that we don't actually store exact representations of conscious experiences. My guess is that we "store" the sensory input and reconstruct memories in a similar vein to constructing conscious experience. But that's a guess. I wondered if scientists have actually investigated memory from that angle.
  8. Jul 24, 2016 #7

    Fervent Freyja

    User Avatar
    Gold Member

    Evo is very right about the mind storing memory strangely. I have also held memories that I have found later to be incorrect (it makes you wonder). I have also caught myself embellishing memories, usually when feeling emotional, quite a few times! I think that this is easy to do with old memories, and we probably adjust the oldest each time we retrieve it. We are able to reanalyze the situation and 'fill-in' details, or we may have new insights about the event when we are older. I do know that some events that I cannot forget have become more vivid and have a wrongful feeling of certainty surrounding them, whereas older versions were probably more foggy.

    The mind does store memory in odd ways. I have memories that seem to have merged over the years, memories where I can find no chronological order, but appear so similar or seem to occur in the same period. For example, I have often been asked how I received the scar on the bridge of my nose. When I was 9 and older, my involuntary, but honest response had probably been, "my stepdad slammed my head into the kitchen table". I hadn't actually witnessed him do it, it had to have come out of nowhere, but for some reason in later years I formed a 'substitute' event where I had been viewing the memory from a third-person perspective. Maybe, to give my mind a better way to store the event, because visually, there wouldn't have been much else to assign that information with- how do you really store the memory of your head moving that quickly towards a table, you can't really see it in that speed later on? The high-speed snapshot wouldn't mean anything without assigning other pieces of information alongside it to give context of what happened. The majority of my memory is stored with moving images, I have difficulty remembering facts unless I assign meaningful movement, that must be why some of my memories are third-person oriented. I wonder if other people have done this too?

    But anyway, over time, I realized that he just likely hit me so hard that my head went into the kitchen table, a place he hadn't intended, but my prior take on the event that he had intended my head go into the table, hence I emphasized slammed to people after the event. Even with that same event, I had believed my aunt had shown up with one of those water-babies (if any of you remember the cute and squishy ones that felt more real because it had the weight of a baby, unlike the air-filled plastic ones) as a gift for me that day. However, years ago, we were talking and she told me that the water-baby hadn't been with her right after it happened, she said she had came over for a different reason (in her defense, I do somewhat remember her behaving as if angry with him) and that she only brought it when she was checking up to see how I was doing days or so afterwards.

    Up until I received that information my mind had merged two separate memories into one. For a long time, I think using the memory of something painful like that happening and contrasting it with playing with my water-baby immediately afterwards served to keep me in a self-pitying state. Probably to assure myself that things like that aren't supposed to happen to children, that it was wrong. The older I get the more I prefer to rely on the objective aspects of the past, memory of actions and re-verifiable facts, speak the most truth for me. It is the only thing we can really use to argue something. The mind seems to store actions or motion to memory more efficiently than sound, words, emotional state, tactile information, etc. She said there was blood, but I cannot remember for sure. I remember my baby sister being there, but can't be certain that she was crying. I may not or may have cried about it. I don't remember the physical pain of it. I remember the shock and finding it hard to breathe. If I wanted to, I could probably look at it again and attempt to piece it together, but really, it won't look anything like it did. I can't remember the color of the table or the furniture arrangements. Knowing the gist of, or ultimate meaning behind our memories are probably the most important. Maybe if there isn't enough information to encode well enough as a memory, then the mind has to resort to creating substitute-like memories to compensate? I do think there are thresholds that have to be met in order for us to place it into long-term memory. That is probably more difficult for the developing brain.

    Maybe I'll hold off on answering psychology, emotional, memory, and likewise questions on here from now on, I just end up making myself cry.
  9. Jul 25, 2016 #8
    If you want to boil it down, you're right about the "reconstruction" of memory argument. Bu this only applies to Homo sapiens sapiens. No other animal has this capacity, that's what makes us unique. Non-human animals do not reconstruct memories and manipulate them in the fashion that humans do. Again, this is what distinguishes us from the animals.
  10. Jul 25, 2016 #9
    Interesting claim. In the absence of verbal reports, how can we be so sure?
  11. Jul 25, 2016 #10
    I think you just answered your own question. It takes the capacity for the mental reconstruction of memories to perform the hiercrchically organized, sequential operations needed to produce a verbal report. You're only going to find that in humans.
  12. Jul 25, 2016 #11


    User Avatar
    Staff Emeritus
    Science Advisor

    Indeed, but the problem is that we don't know how consciousness arises, so we don't know how something goes from being "purely physical" to being a "conscious experience".

    My best guess, based off what little I know about the subject, is that sensory signals are processed by various parts of the brain and these processed signals eventually wind up being either stored or discarded as necessary in different parts of the brain. These signals, combined with all of the other processes that are happening at the same time, constitute "conscious experience". When you recall a memory, you are pulling stored signals from the memory centers of your brain, which are then processed to fill in the details, and those processed signals wind up contributing to your memory of an event.

    What specific information from your senses is stored, how it is stored, how it is recalled, and how it is processed are, unfortunately, all poorly understood at this time.
  13. Jul 25, 2016 #12
    Diracpool, are you sure we aren't talking about two different things? I agree that the use of language and abstract thought might be unique to humans (although I have no idea how "true" that is), but memory serves a broad range of abilities and there are rather a few kinds of memory. You seem to me to be more describing cognitive function and its utilisation of stored concepts than memory storage and retrieval per se, which is what my question was about.

    When I talk about memory "reconstruction", I am referring to a non-conscious, non-directed process. And probably only in certain contexts too, though I am not especially clear about the distinctions. For example procedural memory is not a conscious process, or at least its function in facilitating motor actions seems not to be. Similarly how stored spatial awareness contributes to navigation ability might be largely unconscious as well.

    I suppose I am more referring to what might be called episodic memory? That is, the capacity to retrieve from memory a representation of a prior sensory experience. Wiki seems to suggest that such memory capacity is shared by other animals. https://en.wikipedia.org/wiki/Episodic_memory

    Drakkith, I accept that we don't know much about how conscious experience arises from physical processes, and I suppose that's what prompted my question. If by recalling memories of prior sensory experience, for example a visual scene, we are retrieving stored memories, and those memories broadly consist of strengthened neural connections, then we are I think simply reforming a particular topography of neural arrangements.

    Now, my question is what topography is it that we reform? Is it the topography that represents some stage of the processing of visual input by the visual system, is it the earliest primitives that represent the raw input, or is it the actual state that represents conscious experience itself, whatever that is?

    When scientists say that long term memories are enabled by strengthened neutral connections, they presumably have in mind a specific thing. Wiki says "Long-term memory, on the other hand, is maintained by more stable and permanent changes in neural connections widely spread throughout the brain."

    But what connections? And in what context?

    Another example that touches on this question is hemispatial neglect. In this case, damage to the right hemisphere renders neglect of awareness of the left side of space. From what I've read, the visual system still processes the signals but somehow awareness doesn't happen. That is, the representation of everything on the left just doesn't get to consciousness.

    Bisiach and Luzzatti conducted a nice experiment on this with two patients back in the 70s.


    They found that two patients who were familiar with a specific location could not describe the contents of the left side of this space when recalling their mental imagery. This shows specifically that even recalled sensory representations suffer from neglect.

    That is, they knew all of what was in this location (ie all sides of the space), but in recalling the details, they could not generate a conscious representation of the left hand side, and this was consistent regardless of their (imagined) orientation. When imagining from one direction, they could describe the right hand side but not the left. When imagining the opposite orienttion, they could describe the previously unknown left hand side (which was now the right hand side).

    This suggests to me that stored memories (or more exactly the enhanced synaptic topography that is retrieved) consists of the sensory signals and not the consciously experienced representation, and that memory is constructed in a narrative fashion from neural arrangements in much the same way the original conscious experience was.

    Which if true, then leads me to ask what exactly it is that scientists are observing when they say that strengthened synaptic connections are the basis of memory. If they don't know what physical arrangement gives rise to consciousness, and conscious recall of memory uses the same process, surely the synaptic connections they are observing aren't the actual memories but are rather the raw "data"?

    Or else I'm completely misunderstanding what I am reading (this is highly likely to be the case, hence my question!)
  14. Jul 26, 2016 #13


    User Avatar
    Staff Emeritus
    Science Advisor

    Unfortunately I can't really help you. My knowledge in this area is extremely poor. Sorry!
  15. Jul 26, 2016 #14
    If you take the visual system as an example, memories are stored basically as Hebbian cell assemblies. Raw input from the environment triggers collections of neurons in the primary visual cortex, and the bonding of these assemblies are constructed though the simultaneous firing of neurons which modify the synaptic connections between these neurons. What we call a "memory" is the process of when a new sensory stimulus triggers a certain percentage of the constituent neurons in the original assembly so as to re-trigger the entire assemble. This is wholly a chaotic process and you only need 1-5% of the neurons involved in the original assembly or "frame" as we call it, to re-trigger that percept. Of course this is not a one-dimensional process, memories are episodes, marked by temporal trajectory, or as I think you may have even alluded to, a chaotic itineracy as defined by Tsuda. Walter freeman has dubbed this the "cinematographic hypothesis" of perception and behavior.

    When we train dogs and cats, we do so by either classical or operant conditioning. We are simply training these animals to associate a certain Hebbian assembly with a behavioral routine. This is much different than a brain that can internally reconstruct these "frames" as humans can do. The exact manner in which this is accomplished by humans is the central source of my research in neuroscience, and in deference to the rules laid out by this forum, I don't really want to discuss it. Although, I probably could get away with it because my work is published in peer-reviewed journals, I value my anonymity more :oldsmile:

    If you want further elaboration, though, I suppose we can find a way to talk around this.
  16. Jul 26, 2016 #15
    Ah. So in the context of my original question, and considering visual experience, strengthened synaptic connections caused from the initial input mean that those same connections (or topological formation) may be later triggered by a stimulus that in fact only needs to fire a small subset of the total assembly. Once firing begins, the entire assembly (what you call a "frame"?) is triggered. Of obvious interest there is that a single experience of a novel scene is enough to cause this potential. Does that mean that this synaptic weighting is really quite sensitive such that a single pass sufficiently raises the potential for later reformation of the assembly (or at least, an approximation of it...)?

    Going back to my question then, it is the actual input that is "stored" in the form of potential neural assemblies. Once triggered, these assemblies somehow contribute to a conscious experience of a memory. That is, we don't encode conscious experience as a memory, we encode the data which is then handled by whatever process causes conscious experience. If it were the conscious experience being encoded, we would know how consciousness arises physically.

    I'm still not clear about why you mix experiential or sensory memories (which seem to me to rather automatically generated, the capacity for which is likely shared by all mammals for example) with learned behaviours. It seems to me we are conflating the advanced cognition of humans and how that utilises stored percepts with the fundamental matter of using sensory data to construct awareness and permit flexible behaviour. Can you explain this better for me?
    Last edited: Jul 26, 2016
  17. Jul 26, 2016 #16
    Absolutely. A significant sensory event can become conditioned strongly to the reward-punishment system or what has been recently labeled as the "salience network" after just a single pass, as you put it. Imagine touching a hot pan or an encounter with a charging lion. It only takes once.

    Well, physical memories are stored as perceptual frames in the cortex that can be manipulated by the frontal cortex "later on" to produce consciousness and cognition. The physical trace of the memory is stored as a Hebbian cell-assembly but it needs to be engaged in a chaotic trajectory to recreate an episodic memory, which by the way, is not free from noise or confounding by spurious influences.

    Again, this goes into my personal theories and research, but what I can tell you is that, at least in the anthroplogical community I deal with, there are two camps: one is the "continuity" camp that think that human intellect is just a linear extension of all other primate intellect, descending down to even lesser mammals, and then there is the "discontinuity" camp that believes that the human brain is special and processes at least cognitive information fundamentally different from all other mammals. I happen to subscribe to the latter camp. Hierarchically produced sequential behavior in animals as triggered through some salience signal such as cell-assembly "frame" is very easily modeled and is the best model we can come up with for a non-human animal. There is no compelling evidence that any non-human animal has the capacity for the internal self-construction of novel frames as the human has.
    Last edited: Jul 26, 2016
  18. Jul 26, 2016 #17
    I think I'd sort of fall into the first camp to an extent in that I tend to think that the core experience of perception, awareness/consciousness and adaptation of behaviour therefrom is shared by all higher animals to a greater or lesser extent. For example corvids have been convincingly shown to have considerable cognitive capacity for novel behaviour and problem solving. But there's no doubt "something else" happens in the human brain...

    I still think we are talking different things. Using the memory of a charging lion to inform later behaviour is one thing, actually seeing and remembering that I saw the charging lion seems to me to be another.
  19. Jul 26, 2016 #18
    Well, I explained how that works in post #14. What's the question?

    I order for you to convincingly back up that claim/argument, you'd have to offer up an operational definition of what "cognitive capacity" is, not only functionally, but structural-functionally and physiologically, which I don't think you're going to be able to do. If you were able to, my guess is that you'd see that, whatever cognitive capacity corvids have, it is not even remotely comparable to the functional machinery that humans use in their production for their cognition. In humans, this happens primarily in the isocortex or neocortex, which birds such as the corvid don't even possess. Irene Pepperberg made a similar claim for the intelligence of the African Grey Parrot with pretty much having no knowledge at all about how the brain works. I wrote an opinion article several years ago deconstructing that claim. The nuclear hyperstraitum that forms the most phylogentically recent region of the bird brain is derived from the emryologic ventral pallium, a much more archaic structure than the dorsal pallium which the more laminar 6-layered mammalian neocortex derives from, and which houses the human cognitive apparatus. These structures process information in fundamentally different ways and to try to compare them apples to apples is a naive effort.
  20. Jul 26, 2016 #19
    OK, maybe it's just my misunderstanding?

    In post #14 you explained how memories are "stored", which effectively answered my original question. However, on my reading of your explanation, I gather that memories are re-triggered assemblies that were initially created during the original experience. And that initial creation process is not a directed thing as such, rather it happens biologically through the cellular processes of the brain.

    My simplistic take then is that experiencing a scene requires maintenance of the resulting neuron assemblies over time (short term memory) and that this maintenance process facilitates the strengthened synaptic connections which enable the assembly to be later re-triggered (long term memory).

    Re-triggering occurs, as you say, "when a new sensory stimulus triggers a certain percentage of the constituent neurons in the original assembly" and due to Hebbian association the entire assembly is re-triggered. In consciousness terms, we mentally "see" the original scene to some degree of fidelity.

    Subsequently associating that assembly (memory) with various behavioural routines through conditioning is a value-add process. We cannot make that association if the memory doesn't exist. So creating the memory comes first and results from an auto-associative process (I think this is what is referred to as Hebbian learning?).

    What I am getting at then is that recalling a memory seems a relatively simple thing - retrigger the assembly and "experience" it. The raw response - the Hebbian assembly being activated (and presumably resulting in a conscious inner image) - simply happens as a result of the stimulus.

    Presumably, all of this is the same in the brain of a dog, a gorilla, or a human being? That is, the brain of a dog also generates Hebbian assemblies in response to sensory input, and these assemblies can be later reactivated via an appropriate trigger. I'd assume then, even without conditioning, any such trigger will cause the dog to experience its memory as an inner mental image in the same way we do. I would agree that the dog probably can't retrieve the memory at will for abstract manipulation but that doesn't mean it can't have an inner experience of a memory does it? Unless we are denying animals inner experience. And here we are straying from the accepted format of this forum, so we can't go there.

    On the matter of animal cognition.

    I didn't claim other animals have anything remotely as functional as human beings have, rather I suggested that other animals do have cognitive capacity of varying levels of functionality and that our own capacities, while substantially more advanced, nonetheless lie on a continuum. My best guess (again a largely uninformed one I admit) is that we tend to mistake cognition for more than it is. Evolution has solved tricky problems in various ways that converge on a singular solution (eg the eye), so it seems reasonable to me that a corvid's brain might enable cognition that doesn't differ substantially in kind even if it does in functional capacity.

    Still, corvids do seem to my somewhat uneducated eye capable of quite striking feats of cognition for animals without language capacity.

  21. Jul 26, 2016 #20
    By the way, thanks so much for your comments. You've largely answered my original question and given me much food for thought! :smile:
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted

Similar Threads - memories made Date
Our memory comes from an ancient virus Mar 2, 2018
Memories of being in the womb? Apr 12, 2017
Can plants memorize? Feb 16, 2017
Medical Temporary Memory Blocks and Facilitating Recall Actively Oct 18, 2016