Header 1

Our future, our universe, and other weighty topics

Saturday, November 5, 2016

How Could 50-Year-Old Memories Be Stored in the Shifting Sands of Synapses?

I may compare the mighty fortress of materialism to a castle that is under siege. The walls that protect the castle are being breached by a wide variety of findings and observations, such as findings about cosmic fine-tuning, and observations such as near-death experiences. So how does a dedicated defender of this materialist paradigm try to defend his besieged castle? He builds walls to prop up the castle's defenses. But sometimes these walls are built of pure speculation, and therefore are no more solid than walls built of gossamer, the stuff that spider webs are built of.

Some examples can be found in the world of neurological theory. One breach in the wall around the castle of materialism is the fact of rapid molecular turnover in synapses. The most popular theory about the storage of memories in the brain has been that memories are stored as what are called synaptic weights. But those weights are built from protein molecules, and those molecules have short lifetimes of only a few weeks. This paper finds that synaptic proteins turn over at a rate of 0.7% per hour, which means 17% per day. Such turnover (which involves the replacement of protein molecules at a rapid rate) should quickly clear any information stored in synapses. 

Based on what we know about molecular turnover, it would seem that the brain should not be able to store memories for longer than a month. But instead humans can remember things well for 50 years or longer. How can that possibly be, if memories are stored in synapses?

The issue is a crucial one, because if your brain is incapable of storing your very long-term memories, it means they must exist elsewhere, presumably in some place (such as a human soul or some spiritual reality or infrastructure) incompatible with materialistic assumptions.

There are a few theories designed to overcome this molecular turnover difficulty, theories that are called synaptic plasticity maintenance theories. But these theories are super-speculative affairs that strain credulity in many ways. One idea is the idea of bistable or bimodal synapses. The speculation goes something like this:

Maybe there is some set of molecules that acts like an “on/off switch,” in the sense that there can be two different states. So when you learn something, maybe some molecules that act like an on/off switch get switched on. Then when some of those molecules start to disappear, because of rapid molecular turnover, maybe there's some kind of “feedback mechanism” that switches the set of molecules back to its original state. So it's kind of like a little square on an SAT test form that is penciled in, and then, after the rain washes away the pencil mark, something acts to restore the little pencil mark that was filled in.

This ornate speculation is extremely unbelievable. There is also no good evidence that it is true, and the evidence actually stands against such an idea. Below is a quote from a scientific paper:

Despite extensive investigation, empirical evidence of a bistable distribution of two distinct synaptic weight states has not, in fact, been obtained....In addition, a demonstration of synaptic bistability would require not only finding two distinct synaptic strength states but also finding that a set of different protocols for LTP induction (e.g., different patterns of stimuli, or localized application of pharmacological agents) commonly switched synaptic weights between the same two stable states. Such a demonstration has not been attempted. In addition, modeling suggests that stochastic fluctuations of macromolecule numbers within a small volume such as a spine head are likely to destabilize steady states of biochemical positive feedback loops, causing randomly timed state switches.... The weight distribution of Song et al... is based on measurements of several hundred excitatory postsynaptic potential (EPSP) amplitudes and appears to particularly disfavor the bimodal hypothesis.

What the above paragraph is basically saying is that there is no good evidence for the idea of bistable or bimodal synapses, and that there is good evidence for rejecting such an idea. There is an additional reason for rejecting the idea, not mentioned in the quote above. It is the fact that complex information like human memories could never be biologically stored in some information storage mechanism based on some binary “on/off” mechanism of storage such as the bistable or bimodal synapses idea imagines. 

Take the simple example of a visual image you see and remember. Each pixel or dot that makes up the image requires a color dot. A color dot could be biologically represented by some synapse strength that could vary from 1 to 16 or 1 to 256, but a color dot could not be biologically represented by some setup consisting of mere “on/off” switches. Computers can store information in binary form, but no plausible mechanism can be described by which a natural biological system could store memories in a binary form.

So the ultra-speculative idea of bistable or bimodal synapses is a bust as an explanation for how your brain might be able to store memories for 50 years despite rapid molecular turnover. Another major attempt to explain such a thing is a cluster theory of synaptic stability. The theory was presented in this paper. The speculative idea is that maybe molecular turnover is much more likely to occur where information has already been stored.

Imagine a very simple case of a 10 by 10 grid of 100 cells, in which some information is stored in the grid. Imagine there is a black square shape stored in the center of this grid, formed by 80 black cells in the grid. Then imagine the grid cells are being randomly overwritten by molecular turnover. But suppose instead of these grid square replacements occurring randomly at all positions, imagine that the grid replacements occur much more near spots that already have a black mark. Then the information might decay less rapidly.

That doesn't sound too unreasonable. But when we take a look at the details of the paper, we find that it makes quite an outrageous assumption. It states, “The primary effect of this implementation is that the insertion probability at a site with many neighbors (within a cluster or on its boundary) is orders of magnitude higher than for a site with a small number of neighbors.” The phrase “orders of magnitude” means something like 100 times, 1000 times, or 10,000 times. To assume that is to “cook the books” in a quite ridiculous way, like some gambler assuming that he will be 1000 times more successful than the gambler next to him at the roulette table. There is no reason why we would see such gigantic discrepancies in the positions where molecular turnover occurred.

The paper shows simulations designed to show that under these absurdly implausible assumptions, information could be preserved despite molecular turnover. The simulations show a little 7 by 7 square grid evolving over 1000 time steps. Information is preserved over 1000 time steps, but is not preserved over 2000 time steps.

There are three problems here: (1) the assumptions about molecular turnover occurring “orders of magnitude” more often near existing storage receptors is absurdly unrealistic and biased; (2) the simulation period is too short, leaving the author with the claim merely that such a “meta-stable network” could last as long as a year (we actually need something that would store information for 50 years); (3) the information being tested in the simulation is too simple, being the type of simplest-possible shape (a square) that works best under such a simulation, rather than a more complicated shape that would tend to not work as well. 

If a similar simulation were attempted with a more complicated shape, such as a Y shape or a P shape, the information would not be well-preserved.  Of course, what we actually store in our memories is vastly more complicated than such test shapes.  

There is another reason why these synaptic plasticity maintenance theories are futile. Even if you were somehow to explain how information could be preserved despite rapid molecular turnover in synapses, you would still have the problem of instability on a larger structural level.

We are told that what are called dendritic spines are storehouses of synaptic strengths. A person believing memories are stored in synapses will sometimes think of these dendritic spines as being rather like words in a paragraph, words storing our memories.

 The little "leaves" shown here are dendritic spines

But how long do these dendritic spines last? In the hippocampus of the brain, they last for less than two months. In the cortex of the brain, they last for less than two years. This study found that dendritic spines in the hippocampus last for only about 30 days. This study found that dendritic spines in the cortex of mice brains have a half-life of only 120 days. 

So it's rather like this. Imagine you are driving on the highway, and you watch a car filled with papers, and papers are blowing out of the car's open windows at a constant rate. That is like the information loss that would seem to be caused by rapid molecular turnover. Then suppose you watch the car crashing into a tree and bursting into flames. That's like the loss of information caused by the short lifetimes of dendritic spines. Now suppose someone creates a theory that maybe someone in the car wrote down the information in all the papers blowing out the windows. That utterly contrived and implausible theory is like the synaptic plasticity maintenance theories I have described. But such theoretical ingenuity is futile, because it cannot explain how the information could be preserved after the car crashes and burns. Similarly, the synaptic plasticity maintenance theories are futile because they can't explain how we could have memories lasting for 50 years despite dendritic spines that last no longer than two years.

Our scientists need to wake up and smell the coffee of their own research findings about the brain. Such findings imply that 50-year-old memories cannot be stored in synapses, and (as discussed here) there's no other plausible storage place where the brain could store them. Our minds must therefore be something much more than just the product of our brains. My 50-year-old memories may be stored in a soul, or some other mysterious consciousness reality, but they cannot be stored in my synapses, which are not a stable platform for permanent information storage. 

Postscript: I mentioned here the low lifetimes of dendritic spines, but didn't mention that synapses themselves don't last very long. Below is a quote from a scientific paper:

A quantitative value has been attached to the synaptic turnover rate by Stettler et al (2006), who examined the appearance and disappearance of axonal boutons in the intact visual cortex in monkeys.. and found the turnover rate to be 7% per week which would give the average synapse a lifetime of a little over 3 months.

 After writing this post, I found the important scientific paper "The demise of the synapse as the locus of memory: A looming paradigm shift?" by Patrick C. Trettenbrein.  This scientist makes quite a few points similar to my points in this post, and then makes this very astonishing confession: "To sum up, it can be said that when it comes to answering the question of how information is carried forward in time in the brain we remain largely clueless."