Volatility of Human Memory

prostoalex writes "Scientific Americans looks into the human brain, trying to figure out why some events just tend to stick in our memories forever, while the others are gone: "How does a gene "know" when to strengthen a synapse permanently and when to let a fleeting moment fade unrecorded? And how do the proteins encoded by the gene "know" which of thousands of synapses to strengthen? The same questions have implications for understanding fetal brain development, a time when the brain is deciding which synaptic connections to keep and which to discard. In studying that phenomenon, my lab came up with an intriguing solution to one of these mysteries of memory.""

Comments

This is kinda interesting

[gtrubetskoy]

From the recently noted on slashdot Edge poll What do You believe is true even though you cannot prove it, I remember this bit by Terrence Sejnowski caught my attention (I'm pasting it here cause I can't figure out how to link to that specific part of the page):

How do we remember the past? There are many answers to this question, depending on whether you are an historian, artist or scientist. As a scientist I have wanted to know where in the brain memories are stored and how they are storedthe genetic and neural mechanisms. Although neuroscientists have made tremendous progress in uncovering neural mechanisms for learning, I believe, but cannot prove, that we are all looking in the wrong place for long-term memory.

I have been puzzled by my ability to remember my childhood, despite the fact that most of the molecules in my body today are not the same ones I had as a childin particular, the molecules that make up my brain are constantly turning over, being replaced with newly minted molecules. Perhaps memories only seem to be stable. Rehearsal strengthens memories, and can even alter them. However, I have detailed memories of specific places where I lived 50 years ago that I doubt I ever rehearsed but can be easily verified, so the stability of long-term memories is a real problem.

Textbooks in neuroscience, including one that I coauthored, say that memories are stored at synapses between neurons in the brain, of which there are many. In neural network models of memory, information can be stored by selectively altering the strengths of the synapses, and "spike-time dependent plasticity" at synapses in the cerebral cortex has been found with these properties. This is a hot area of research, but all we need to know here is that patterns of neural activity can indeed modify a lot of molecular machinery inside a neuron.

If memories are stored as changes to molecules inside cells, which are constantly being replaced, how can a memory remain stable over 50 years? My hunch is that everyone is looking in the wrong place: that the substrate of really old memories is located not inside cells, but outside cells, in the extracellular space. The space between cells is not empty, but filled with a matrix of tough material that is difficult to dissolve and turns over very slowly if at all. The extracellular matrix connects cells and maintains the shape of the cell mass. This is why scars on your body haven't changed much after decades of slougare contained in the endoskeleton that connects cells to each other. The intracellular machinery holds memories temporarily and decides what to permanently store in the matrix, perhaps while you are sleeping. It might be possible someday to stain this memory endoskeleton and see what memories look like.what makes you a unique individualhing off skin cells.

My intuition is based on a set of classic experiments on the neuromuscular junction between a motor neuron and a muscle cell, a giant synapse that activates the muscle. The specialized extracellular matrix at the neuromuscular junction, called the basal lamina, consists of proteoglycans, glycoproteins, including collagen, and adhesion molecules such as laminin and fibronectin. If the nerve that activates a muscle is crushed, the nerve fiber grows back to the junction and forms a specialized nerve terminal ending. This occurs even if the muscle cell is also killed. The memory of the contact is preserved by the basal lamina at the junction. Similar material exists at synapses in the brain, which could permanently maintain overall connectivity despite the coming and going of molecules inside neurons.

How could we prove that the extracellular matrix really is responsible for long-term memories? One way to disprove it would be to disrupt the extracellular matrix and see if the memories remain. This can be done with enzymes or by knocking out one or more key molecules with techniques from molecular genetics. If I am right, then all of your memories

Re:This is kinda interesting

[FleaPlus]

That's bullshit.

I wouldn't be so quick to dismiss it. Terry Sejnowski is probably one of the most prominent neuroscientists alive today, and generally knows what he's talking about.

You do not need constancy of material/molecules to keep a memory - in a sense you can exchange a building brick by brick, one at a time, with new bricks, and maintain your building like new, for millenia.

This is true, and undoubtedly works well for short-term and medium-term memories. However, all of this exchange takes energy, and if there's a more energy efficient way of doing things (such as, perhaps, storing memories in the extracellular matrix), evolution would tend to select in favor of it.

Re:This is kinda interesting

[nucal]

The idea that the extracellular matrix might control neuron plasticity is not all that far fetched - there are many studies showing that cell function is controlled by the extracellular environment.

Another aspect to consider is that diseases such as Alzheimers are associated with the accumulation of misfolded proteins (plaques) in the extracellular environment. Although the prevailing idea is that these plaques might be toxic or the residue of dead cells, it's not impossible to think that plaques could also "de-program" neurons by altering the normal extracellular environment.

Re:This is kinda interesting

[Tlosk]

While I belive Sejnowski is absolutely correct that there's probably a lot worth investigating in the extracellular material, one possibility is that very long term memories are illusory.

Rarely are these long term memories of the same quality as very recent memories, and I don't just mean of strength, but that they are qualitatively different. That you no longer have access to what one might call witness memory, where if someone asks you questions about the event you can search the myriad details of the event to find the answer.

Given that the bulk of our early memories are lost over time, what's special about that handful of memories that we do hold onto and that are veridical? I suspect that most of this subset of retained memories are not the original memories but rather memories of the memories.

Personally, when I go over the longest memories I still hold onto, they are almost all experiences that I at some point either told someone else about, thought about, or had cause to remember at some point in the past. Each time you do this the memory is copied to other areas (whatever those might be, we still don't have a good grasp on this). And most of a given memory that I now have owes its features to the nature of the account I gave earlier.

For example say someone remembers the experience of riding on their grandmother's lap on a train when he goes to visit her at the age of three. Shortly after that he will have all sorts of specific stored information relating to that particular event. If the event is never revisited it will likely be almost entirely lost, but if several years later he tells someone else about the experience, a memory of the event recounted still many years further down the road would depend heavily on what exactly the person shared during that earlier recounting. That is, the person is no longer remembering the event, but rather recalling the earlier recounting.

Oh if you're cued well enough you can remember all sorts of things from way back, but they are so fragmentary that it's probably just the distributed nature of memory that saves them from complete loss most of the time. There will always be a few bits and pieces floating around in there.

Pain for me

[purduephotog]

I remember back to when I was only 2 years old- I had had surgery on ... well, we'll call it a sensitive part of the body.

Now I don't remember the surgery, and I don't remember the antics I pulled at showing nurses why I was in the hospital... but I *do* remember the first time I had to goto the bathroom after surgery.

That memory is so seared into my brain I can even recall I was high enough to look out a window over the cityscape, and that there was a bricked church in the background and the window had blinds (the black slatted ones) on it.

And I remember so much so terribly much pain I don't know how I survived it.

My parents tell me that after that brief moment of screaming I was OK... and I don't remember anything else of that event save for that moment.

And just for comparison (of a little kid) I've had 18 kidney stones... I have a good memory for pain. But that memory makes me cringe and shiver every time I have it.

Does this shed any more light on coding solutions?

[Sowelu]

I've been tracking the periphery of AI for quite a while. Even though directly emulating the human brain is probably not the best solution for artificial intelligence, has this research opened any new doors lately?

Fake memories

[D H NG]

Sometimes our brains can be tricked into remembering things that did not happen. Elizabeth Loftus had done much research in the area of misinformation effect, which actually has legal repercussions.

or, 'Potential Programmability of Human Memory'

[brian.glanz]

A leitmotif the article turns on is the potential programmability, more than the volatility, of human memory. They discuss how the older view of our memory as volatile and mysterious has been refined, as we've discovered the mechanisms for transition between short and long term memory. From the physiological to the cellular level, the idea here is a familiar one -- we know more than ever, and we're learning faster than we had before, in this case about memory and about learning.

Most intriguing are the material implications of the article -- they find memories transitioning to long term storage when information is reinforced at specific intervals and with specific techniques. This matches some experimental evidence as referred to, like the familiar ideas of studying or preparing in the same location you will test or perform in -- but, its level of specificity begs for more experimentation and refinement of memory management techniques. Learning and memory across the whole human experience can be biologically maximized if we find just the right process -- read that slippery section in x minute increments and take 10 minute brakes between 3 repetitions. Or maybe, do asdf to remember x words by rote for the next 4 hours, and do ;lkj to sufficiently remember x for a month. Without running a cord into your ear, the article is promising for its level of detail in exact ways we might approach finding best practices for our current hardware.

I'm curious generally about how soon articles like this, especially up at the Scientific American level of exposure, translate into experiments at universities (and, self-help books?). I'm tempted to modify my own learning accordingly, n/m waiting.

BG