Thursday, November 22, 2018

Why Most Animal Memory Experiments Tell Us Nothing About Human Memory

Recently the BBC reported a science experiment with the headline “'Memory transplant' achieved in snails.” This was all over the science news on May 14. Scientific American reported it with a headline stating “Memory transferred between snails,” and other sites such as the New York Times site made similar matter-of-fact announcements of a discovery. But you need not think very hard to realize that there's something very fishy about such a story. How could someone possibly get decent evidence about a memory in a snail?

To explain why this story and similar stories do not tell us anything reliable about memory, we should consider the issue of small sample sizes in neuroscience studies. The issue was discussed in a paper in the journal Nature, one entitled Power failure: why small sample size undermines the reliability of neuroscience. The article tells us that neuroscience studies tend to be unreliable because they are using too small a sample size. When there is too small a sample size, there's a too high chance that the effect reported by a study is just a false alarm. An article on this important Nature article states the following:


The group discovered that neuroscience as a field is tremendously underpowered, meaning that most experiments are too small to be likely to find the subtle effects being looked for and the effects that are found are far more likely to be false positives than previously thought. It is likely that many theories that were previously thought to be robust might be far weaker than previously imagined.

I can give a simple example illustrating the problem. Imagine you try to test extrasensory perception (ESP) using a few trials with your friends. You ask them to guess whether you are thinking of a man or a woman. Suppose you try only 10 trials with each friend, and the best result is that one friend guessed correctly 70% of the time. This would be very unconvincing as evidence of anything. There's about a 5 percent chance of getting such a result on any such test, purely by chance; and if you test with five people, you have perhaps 1 chance in 4 that one of them will be able to make 7 such guesses correctly, purely by chance. So having one friend get 7 out of 10 guesses correctly is no real evidence of anything. But if you used a much larger sample size it would be a different situation. For example, if you tried 1000 trials with a friend, and your friend guessed correctly 700 times, that would have a probability of less than 1 in a million. That would be much better evidence.

Now, the problem with many a neuroscience study is that very small sample sizes are being used. Such studies fail to provide convincing evidence for anything. The snail memory test is an example.

The study involved giving shocks to some snails, extracting RNA from their tiny brains, and then injecting that into other snails that had not been shocked. It was reported that such snails had a higher chance of withdrawing into their shells, as if they were afraid and remembered being shocked when they had not. But it might have been that such snails were merely acting randomly, not experiencing any fear memory transferred from the first set of snails. How can you have confidence that mere chance was not involved? You would have to do many trials or use a sample size that guarantees that sufficient trials will occur. This paper states that in order to have moderate confidence in results, getting what is called a statistical power of .8,  there should be at least 15 animals in each group. This statistical power of .8 is a standard for doing good science. 

But judging from the snail paper, the scientists did not do a large number of trials. Judging from the paper, the effect described involved only 7 snails (the number listed on lines 571 -572 of the paper). There is no mention of trying the test more than once on such snails. Such a result is completely unimpressive, and could easily have been achieved by pure chance, without any real “memory transfer” going on. Whether the snail does or does not withdraw into its shell is like a coin flip. It could easily be that by pure chance you might see some number of “into the shell withdrawals” that you interpret as “memory transfer.”

Whether a snail is withdrawing into its shell requires a subjective judgment, where scientists eager to see one result might let their bias influence their judgments about whether the snail withdrew into its shell or not. Also, a snail might withdraw into its shell simply because it has been injected with something, not because it is remembering something. Given such factors and the large chance of a false alarm when dealing with such a small sample size, this “snail memory transfer” experiment offers no compelling evidence for anything like memory transfer. We may also note the idea that RNA is storing long-term memories in animals is entirely implausible, because of RNA's very short lifetime. According to this source, RNA molecules typically last only about two minutes, with 10 to 20 percent lasting between 5 and 10 minutes. And according to this source, if you were to inject RNA into a bloodstream, the RNA molecules would be too large to pass through cell membranes.

The Tonegawa memory research lab at MIT periodically puts out sensational-sounding press releases on its animal experiments with memory. Among the headlines on its site are the following:
  • “Neuroscientists identify two neuron populations that encode happy or fearful memories.”
  • “Scientists identify neurons devoted to social memory.”
  • “Lost memories can be found.”
  • “Researchers find 'lost' memories”
  • “Neuroscientists reverse memories' emotional associations.”
  • “How we recall the past.”
  • “Neuroscientists identify brain circuit necessary for memory formation.”
  • “Neuroscientists plant false memories in the brain.”
  • “Researchers show that memories reside in specific brain cells.”
But when we take a close look at the issue of sample size and statistical power, and the actual experiments that underlie these claims, it seems that few or none of these claims are based on solid, convincing experimental evidence. Although the experiments underlying these claims are very fancy and high-tech, the experimental results seem to involve tiny sample sizes so small that very little of it qualifies as convincing evidence.

A typical experiment goes like this: (1) Some rodents are given electrical shocks; (2) the scientists try to figure out where in the rodent's brain the memory was; (3) the scientists then use an optogenetic switch to “light up” neurons in a similar part of another rodent's brain, one that was not fear trained; (4) a judgment is made on whether the rodent froze when such a thing was done.

Such experiments have the same problems I mentioned above with the snail experiment: the problem of subjective interpretations and alternate explanations. The MIT memory experiments typically involve a judgment of whether a mouse froze. But that may often be a hard judgment to make, particularly in borderline cases. Also, we have no way of telling whether a mouse is freezing because he is remembering something. It could be that the optogenetic zap that the mouse gets is itself sufficient to cause the mouse to freeze, regardless of whether it remembers something. If you're walking along, and someone shoots light or energy into your brain, you might stop merely because of the novel stimulus. A science paper says that it is possible to induce freezing in rodents by stimulating a wide variety of regions. It says, "It is possible to induce freezing by activating a variety of brain areas and projections, including the hippocampus (Liu et al., 2012), lateral, basal and central amygdala (Ciocchi et al., 2010); Johansen et al., 2010; Gore et al., 2015a), periaqueductal gray (Tovote et al., 2016), motor and primary sensory cortices (Kass et al., 2013), prefrontal projections (Rajasethupathy et al., 2015) and retrosplenial cortex (Cowansage et al., 2014).”

But the main problem with such MIT memory experiments is that they involve very small sample sizes, so small that all of the results could easily have happened purely because of chance. Let's look at some sample sizes, remembering that according to this scientific paper, there should be at least 15 animals in each group to have moderate confidence in your results, sufficient to reach the standard of a “statistical power of .8.”.

Let's start with their paper, “Memory retrieval by activating engram cells in mouse models of early Alzheimer’s disease,” which can be accessed from the link above after clicking underneath "Lost memories can be found." The paper states that “No statistical methods were used to predetermine sample size.” That means the authors did not do what they were supposed to have done to make sure their sample size was large enough. When we look at page 8 of the paper, we find that the sample sizes used were merely 8 mice in one group and 9 mice in another group. On page 2 we hear about a group with only 4 mice per group, and on page 4 we hear about a group with only 4 mice per group. Such a paltry sample size does not result in any decent statistical power, and the results cannot be trusted, since they very easily could be false alarms. The study therefore provides no convincing evidence of engram cells.

Another example is this paper by the MIT memory lab, with the grandiose title “Creating a False Memory in the Hippocampus.” When we look at Figure 2 and Figure 3, we see that the sample sizes used were paltry: the different groups of mice had only about 8 or 9 mice per group. Such a paltry sample size does not result in any decent statistical power, and the results cannot be trusted, since they very easily could be false alarms. No convincing evidence has been provided of creating a false memory.

A third example is this paper with the grandiose title “Optogenetic stimulation of a hippocampal engram activates fear memory recall.” Figure 2 tells us that in one of the groups of mice there were only 5 mice, and that in another group there were only 3 mice. Figure 3 tells us that in two other groups of mice there were only 12 mice. Figure 4 tells us that in some other group there was only 5 mice. Such a paltry sample size does not result in any decent statistical power, and the results cannot be trusted, since they very easily could be false alarms. No convincing evidence has been provided of artificially activating a fear memory by the use of optogenetics.

Another example is this paper entitled “Silent memory engrams as the basis for retrograde amnesia.” Figure 1 tells us that the number of mice in particular groups used for the study ranged between 4 and 12. Figures 2 and 3 tell us that the number of mice in particular groups used for the study ranged between 3 and 12. Such a paltry sample size does not result in any decent statistical power, and the results cannot be trusted, since they very easily could be false alarms. Another unsound paper is the 2015 paper "Engram Cells Retain Memory Under Retrograde Amnesia," co-authored by Tonegawa. When we look at the end of the supplemental material, and look at figure s13, we find that the experimenters were using a number of mice that was equal to only 8 in one study group, and 7 in another study group.  Such a paltry sample size does not result in any decent statistical power, and the results cannot be trusted, since they very easily could be false alarms. 

We see the same "low statistical power" problem in this paper claiming an important experimental result regarding memory. The paper states in its Figure 2 that only 6 mice were used for a study group, and 6 mice for the control group. The same problem is shown in Figure 3 and Figure 4 of the paper.  We see the same  "low statistical power" problem in this paper entitled "Selective Erasure of a Fear Memory." The paper states in its Figure 3 that only 6 to 9 mice were used for a study group, That's only about half of the "15 animals per study group" needed for a modestly reliable result.  The same defect is found in this memory research paper and in this memory research paper. 

The term “engram” means a cell or cells that store memories. Decades after the term was created, we still have no convincing evidence for the existence of engram cells. But memory researchers are shameless in using the term “engram” matter-of-factly even though no convincing evidence of an engram has been produced. So, for example, one of the MIT Lab papers may again and again refer to some cells they are studying as “engram cells,” as if they could try to convince us that such cells are actually engram cells by telling us again and again that they are engram cells. Doing this is rather like some ghost researcher matter-of-factly using the term “ghost blob” to refer to particular patches of infrared light that he is studying after using an infrared camera. Just as a blob of infrared light merely tells us only that some patch of air was slightly colder (not that such a blob is a ghost), a scientist observing a mouse freezing is merely entitled to say he saw a mouse freezing (not that the mouse is recalling a fear memory); and a scientist seeing a snail withdrawing into its shell is merely entitled to tell us that he saw a snail withdrawing into its shell (not that the snail was recalling some fear memory).

The relation between the chance of a false alarm and the statistical power of a study is clarified in this paper by R. M. Christley. The paper has an illuminating graph which I present below with some new captions that are a little more clear than the original captions. We see from this graph that if a study has a statistical power of only about .2, then the chance of the study giving a false alarm is something like 1 in 3 if there is a 50% chance of the effect existing, and much higher (such as 50% or greater) if there is less than a 50% chance of the effect existing. But if a study has a statistical power of only about .8, then the chance of the study giving a false alarm is only about 1 in 20 if there is a 50% chance of the effect existing, and much higher if there is less than a 50% chance of the effect existing. Animal studies using much fewer than 15 animals per study (such as those I have discussed) will result in the relatively high chance of false alarms shown in the green line.

false positive

The PLOS paper here analyzed 410 experiments involving fear conditioning with rodents, a large fraction of them memory experiments. The paper found that such experiments had a “mean normalized effect size” of only .29. An experiment with an effect size of only .29 is very weak, with a high chance of a false alarm. Effect size is discussed in detail here, where we learn that with an effect size of only .3, there's typically something like a 40 percent chance of a false alarm.


To determine whether a sample size is large enough, a scientific paper is supposed to do something called a sample size calculation. The PLOS paper here reported that only one of the 410 memory-related neuroscience papers it studied had such a calculation.  The PLOS paper reported that in order to achieve a moderately convincing effect size of .80, an experiment typically needs to have 15 animals per group; but only 12% of the experiments had that many animals per group. Referring to statistical power (a measure of how likely a result is to be real and not a false alarm), the PLOS paper states, “no correlation was observed between textual descriptions of results and power.” In plain English, that means that there's a whole lot of BS flying around when scientists describe their memory experiments, and that countless cases of very weak evidence have been described by scientists as if they were strong evidence.

Our science media shows very little sign of paying any attention to the statistical power of neuroscience research, partially because rigor is unprofitable. A site can make more money by trumpeting borderline weakly-suggestive research as if it were a demonstration of truth, because the more users click on a sensational-sounding headline, the more money the site make from ads. Our neuroscientists show little sign of paying much attention to whether their studies have a decent statistical power. For the neuroscientist, it's all about publishing as many papers as possible, so it's a better career move to do 5 underpowered small-sample studies (each with a high chance of a false alarm) than a single study with an adequate sample size and high statistical power.

In this post I used an assumption (which I got from one estimate) that 15 research animals per study group are needed for a moderately persuasive result. It seems that this assumption may have been too generous. In her post “Why Most Published Neuroscience Findings Are False,” Kelly Zalocusky PhD calculates (using Ioannidis’s data) that the median effect size of neuroscience studies is about .51. She then states the following, talking about statistical power:

To get a power of 0.2, with an effect size of 0.51, the sample size needs to be 12 per group. This fits well with my intuition of sample sizes in (behavioral) neuroscience, and might actually be a little generous. To bump our power up to 0.5, we would need an n of 31 per group. A power of 0.8 would require 60 per group.

If we describe a power of .5 as being moderately convincing, it therefore seems that 31 animals per study group is needed for a neuroscience study to be moderately convincing. But most experimental neuroscience studies involving rodents and memory use fewer than 15 animals per study group. 

Zalocusky states the following:

If our intuitions about our research are true, fellow graduate students, then fully 70% of published positive findings are “false positives.” This result furthermore assumes no bias, perfect use of statistics, and a complete lack of “many groups” effect. (The “many groups” effect means that many groups might work on the same question. 19 out of 20 find nothing, and the 1 “lucky” group that finds something actually publishes). Meaning—this estimate is likely to be hugely optimistic.

All of these things make it rather clear that a large fraction or most animal memory experiments are dubious.  There is another reason why the great majority of these experiments tell us nothing about human memory.  It is that most such experiments involve rodents, and given the vast differences between men and rodents, nothing reliable about human memory can be determined by studying rodent memory. 

Postscript: The paper here is another example of a memory experiment failing to actually prove anything because of its too-small-sample size.  Widely reported in the press with headlines suggesting scientists had added memories to mice while the mice slept, the study says, "We induced an explicit memory trace, leading to a goal-directed behavior toward the place field."  Typically this type of study will be behind a pay wall, allowing the scientists to hide their too-small sample sizes where the public won't be able to see them without paying. But luckily www.researchgate.net often publishes the graphs from such studies, where anyone can see them. In this case the graph explanation allows us to see the scientists were using only 5 to 7 animals per study group, which means that the reported result isn't strong evidence for anything, being the type of result we might easily get from chance effects. 

Post-Postscript: The latest example of a memory experiment failing to actually prove anything (because of its too-small-sample size) is a study in Nature that has been hyped with headlines such as "Artificial memory created."  The study has the inaccurate title, "Memory formation in the absence of experience." The study fails to prove any such thing occurred. When we look at the number of animals involved, we often find that the study fails to meet the minimum standard of 15 animals per study group.  In Figure 1 we learn that two of the study groups consisted of only 8 mice. In Figure 2 we learn that two of the study groups consisted of only 10 mice.  In Figure 3 we learn that one of the study groups consisted of only 7 mice.  Moreover, the methodology used in the study is so convoluted that it fails to provide clear and convincing evidence for anything interesting.  The only evidence of memory recall is that the mice supposedly avoided some area,  something that might have occurred for any number of reasons other than a recall of some memory.  A robust test of an artificial memory would test an actual acquired skill, such as the ability to navigate a maze in a certain time. 

Sunday, November 18, 2018

Brain Dogmas Versus Case Histories That Refute Them

Our neuroscientists have the bad habit of frequently spouting dogmas that have not been established by observations. We have all heard these dogmas stated hundreds of times, such as when neuroscientists claim that memories are stored in brains, and that our minds are produced by our brains. There are actually many observations and facts that contradict such dogmas, such as the fact that many people report their minds and memories still working during a near death experience in which their brains shut down electrically (as the brain does soon after cardiac arrest).

One of the most dramatic type of observations conflicting with neuroscience dogmas is the fact that memory and intelligence is well-preserved after the operation called hemispherectomy. Hemispherectomy is the surgical removal of half of the brain. It is performed on children who suffer from severe and frequent epileptic seizures.

In a scientific paper “Discrepancy Between Cerebral Structure and Cognitive Functioning,” we are told that when half of their brains are removed in these operations, “most patients, even adults, do not seem to lose their long-term memory such as episodic (autobiographic) memories.” The paper tells us that Dandy, Bell and Karnosh “stated that their patient's memory seemed unimpaired after hemispherectomy,” the removal of half of their brains. We are also told that Vining and others “were surprised by the apparent retention of memory after the removal of the left or the right hemisphere of their patients.”

On page 59 of the book The Biological Mind, the author states the following:

A group of surgeons at Johns Hopkins Medical School performed fifty-eight hemispherectomy operations on children over a thirty-year period. "We were awed," they wrote later of their experiences, "by the apparent retention of memory after removal of half of the brain, either half, and by the retention of the child's personality and sense of humor." 

In the paper "Neurocognitive outcome after pediatric epilepsy surgery" by Elisabeth M. S. Sherman, we have some discussion of the effects on children of temporal lobectomy (removal of the temporal lobe of the brain) and hemispherectomy, surgically removing half of the brain to stop seizures. We are told this:

After temporal lobectomy, children show few changes in verbal or nonverbal intelligence....Cognitive levels in many children do not appear to be altered significantly by hemispherectomy. Several researchers have also noted increases in the intellectual functioning of some children following this procedure....Explanations for the lack of decline in intellectual function following hemispherectomy have not been well elucidated. 

Referring to a study by Gilliam, the paper states that of 21 children who had parts of their brains removed to treat epilepsy, including 10 who had surgery to remove part of the frontal lobe, "none of the patients with extra-temporal resections had reductions in IQ post-operatively," and that two of the children with frontal lobe resections had "an increase in IQ greater than 10 points following surgery." 

The paper here gives precise before and after IQ scores for more than 50 children who had half of their brains removed in a hemispherectomy operation in the United States.  For one set of 31 patients, the IQ went down by an average of only 5 points. For another set of 15 patients, the IQ went down less than 1 point. For another set of 7 patients the IQ went up by 6 points. 

The paper here (in Figure 4) describes IQ outcomes for 41 children who had half of their brains removed in hemispherectomy operations in Freiburg, Germany. For the vast majority of children, the IQ was about the same after the operation. The number of children who had increased IQs after the operation was greater than the number who had decreased IQs. 

Referring to these kind of surgeries to remove huge amounts of brain tissue, the
paper “Verbal memory after epilepsy surgery in childhood” states, “Group-wise, average normed scores on verbal memory tests were higher after epilepsy surgery than before, corroborating earlier reports.” 

Some try to explain these results as some kind of special ability of the child brain to recover. But there are similar results even for adult patients. The page here mentions 41 adult patients who had a hemispherectomy. It says, “Forty-one patients underwent additional formal IQ testing postsurgery, and the investigators observed overall stability or improvement in these patients,” and notes that “significant functional impairment has been rare.”

Of these cases of successful hemispherectomy, perhaps none is more astonishing than a case of a boy named Alex who did not start speaking until the left half of his brain was removed. A scientific paper describing the case says that Alex “failed to develop speech throughout early boyhood.” He could apparently say only one word (“mumma”) before his operation to cure epilepsy seizures. But then following a hemispherectomy (also called a hemidecortication) in which half of his brain was removed at age 8.5, “and withdrawal of anticonvulsants when he was more than 9 years old, Alex suddenly began to acquire speech.” We are told, “His most recent scores on tests of receptive and expressive language place him at an age equivalent of 8–10 years,” and that by age 10 he could “converse with copious and appropriate speech, involving some fairly long words.” Astonishingly, the boy who could not speak with a full brain could speak well after half of his brain was removed. The half of the brain removed was the left half – the very half that scientists tell us is the half that has more to do with language than the right half. 

We learn of quite a few such cases in the scientific paper "Long-Term Memory: Scaling of Information to Brain Size" by Donald R. Forsdyke of Queens University in Canada.  He quotes the physician John Lorber on an honors student with an IQ of 126:

Instead of the normal 4.5 centimetre thickness of brain tissue between the ventricles and the cortical surface, there was just a thin layer of mantle measuring a millimeter or so. The cranium is filled mainly with cerebrospinal fluid. … I can’t say whether the mathematics student has a brain weighing 50 grams or 150 grams, but it’s clear that it is nowhere near the normal 1.5 kilograms.

Forsdyke notes two similar cases in more recent years, one from France and another from Brazil.  

Cases like this make a mockery of scientist claims to understand the human brain. When scientists discuss scientific knowledge relating to memory, they almost never discuss the most relevant thing they could discuss, the cases of high brain function after hemispherectomy operations in which half of the brain is removed. Instead the scientists cherry-pick information, and describe a few experiments and facts carefully selected to support their dogmas, such as the dogma that brains store memories, and brains make minds. They also fail to discuss the extremely relevant research of John Lorber, who documented many cases of high-functioning humans who had lost almost all of their brain due to hydroencephaly. 


cherry picking

A scientist discussing memory will typically refer us to experiments involving rodents. Such experiments are almost always studies with low statistical power, because the experimenter failed to use at least 15 animals per study group, the minimum needed for a moderately reliable result with a low risk of a false alarm. There will be typically be some graph showing some measurement of what is called freezing behavior, when a rodent stops moving. The experimenter will claim that this shows something was going on in regard to memory, although it probably does not show such a thing, because all measurements of a rodent's degree of freezing are subjective judgments in which an experimenter's bias might have influenced things. There will often be claims that a fear memory was regenerated by electrically zapping some part of the brain where the experimenter thought the memory was stored. Such claims have little force because it is known that there are many parts of a rodent's brain that will cause a rodent to stop moving when such parts are electrically stimulated. And, of course, rodent experiments prove nothing about human memory, because humans are not rodents.

When a scientist discusses memory research, he will typically discuss the case of patient HM, a patient who was bad at forming new memories after damage to the tiny brain region called the hippocampus. Again and again, writers will speak as if this proves the hippocampus is crucial to memory. It certainly does not. The same very rare effect of having a problem in forming new memories cropped up (as reported here) in a man who underwent a dental operation (a root canal). The man had no brain damage, but after the root canal he was reportedly unable to form new memories. Such cases are baffling, and the fact that they can come up with or without brain damage tells us no clear tale about whether the hippocampus is crucial for memory. The hemispherectomy cases suggest that the hippocampus is not crucial for memory, for each patient who had a hemispherectomy lost one of their two hippocampuses, and overall there was little permanent effect on the ability to form new memories.

A scientific paper tells us that “lesions of the rodent hippocampus do not produce reliable anterograde amnesia for context fear,” meaning rodents with a damaged hippocampus can still produce new memories. The paper also tells us, “These data suggest that the hippocampus is normally involved in context conditioning but is not necessary for learning to occur.” So it seems that the main claim that neuroscientists cite to persuade us that they have some understanding of a neural basis for memory (the claim that the hippocampus is “crucial” for memory) is really a factoid that is not actually well established. 

Postscript: The case of patient HM has been cited innumerable times by those eager to suggest that memories are brain based. Such persons usually tell us that patient HM was someone unable to form any new memories.  But a 14-year follow-up study of patient HM (whose memory problems started in 1953) actually tells us that HM was able to form some new memories. The study says this on page 217:

In February 1968, when shown the head on a Kennedy half-dollar, he said, correctly, that the person portrayed on the coin was President Kennedy. When asked him whether President Kennedy was dead or alive, and he answered, without hesitation, that Kennedy had been assassinated...In a similar way, he recalled various other public events, such as the death of Pope John (soon after the event), and recognized the name of one of the astronauts, but his performance in these respects was quite variable. 

The study also says that patient HM was able to learn a maze (although learning it only very slowly),  and was able eventually to walk the maze three times in a row without error. 

Saturday, November 3, 2018

Vacillating Disarray of the Memory Trace Theorists

In a February post entitled "Turmoil of the Baffled Engram Theorists," I discussed a Science News article that showed the theoretical disarray of engram theorists, scientists who speculate about a physical brain basis for human memories. Three scientific papers in recent years suggest that claims that human memories are stored in brains do not have a solid theoretical basis well-substantiated by experiments.

One paper was entitled “The mysteries of remote memory,” and was published in the Philosophical Transactions of the Royal Society B. Speaking of long-lasting memories, the authors told us that “our current knowledge of how such memories are stored in the brain and retrieved, as well as the dynamics of the circuits involved, remains scarce.” Using the term “engrams” to mean the hypothesis that there are cells in the brain that store memories, the authors state “what and where engrams are implicated in remote memory storage and how they change over time have received little experimental attention thus far.” The authors also frankly tell us that “From engrams to spines surprisingly little evidence exists in the literature on the grounds of remote information processing, maintenance and storage to account for the lifelong and persistent nature of the mnemonic signal.” This type of candor is a refreshing contrast from the click-bait hype about memory research that we get in the science news, where dubious studies using insufficient experimental groups are often trumpeted as scientific breakthroughs.

One of the ideas about a brain storage of memory is that memories get stored in dendritic spines, little bumps on dendrites. But this study found that dendritic spines in the hippocampus last for only about 30 days. And this study found that dendritic spines in the cortex of mice brains have a half-life of only 120 days. So such dendritic spines don't last enough to store memories that last for years. The “Mysteries of remote memory” paper mentions a study that found that studied the persistence of dendritic spines, and found a “near full erasure of the synaptic connectivity pattern within 15 days post-learning.” The paper says “these incongruent findings point to the need for an alternative explanation to spine dynamics for remote memory stability.” In other words, we can't explain dendritic spines as some physical basis for long-term memory. Referring to the often stated speculations that memories start out in the hippocampus and are transferred to the cortex, the paper states unequivocal experimental evidence in support of it is lacking.” The paper then spends some time talking about the very speculative possibility that DNA methylation has something to do with long-term memory, an idea for which there is no evidence. 

The overall impression we get from reading such a paper is one of uncertainty and disarray, as if no clear idea was emerging from brain studies on the long-term storage of memory. We get such an impression even more strongly from two other papers on this topic published in recent years. One is the 2016 paper “What Is Memory? The Present State of the Engram.” Very oddly for a scientific paper, the paper consists of ten sections, each written by a different author or authors. We get conflicting theories as to how a brain might store memories, with little agreement between the sections.

The 2017 paper here is entitled, "On the research of time past: the hunt for the substrate of memory." It is a portrait of memory theorists in disarray, presenting no one theory about how memory might be stored in a brain, and instead suggesting seven or more possibilities, none of which is plausible. The paper is all over the map in its speculations, like someone shooting a gun in all different directions.



The paper tells us on its sixth page that “Engram-labeling studies have shown that certain populations of neurons encode specific memories in mice.” This is not correct. The studies in question are typically small-sample studies with very low statistical power, in which there is a high chance of false alarms. In a typical such study, an experimenter will zap some small portion of a mouse's brain, and claim that he has elicited a fear memory stored in that part, producing a freezing effect in the mouse. But such freezing effects (which require subjective judgments) mean little, since it is known that there are many parts of a mouse's brain that can be stimulated to produce a freezing effect in a mouse.

The paper tells us on page 9 that “synaptic weight changes can now be excluded as a means of information storage.” But for many years neuroscientists have been pushing the very dubious dogma that memories are stored by changes in synaptic weights. As discussed here, this idea never made any sense, for there has never been a proven case of any information that was ever stored as changes in the weights of something, and also synapses are too volatile to store memories that last for decades, for reasons discussed later in this post. 

The authors then proceed to discuss a wide variety of possibilities for how memory might be stored in a brain.  These include the following (the quotes below in italics are from the paper):

Theory 1: “The particularly longlived proteins associated with DNA (i.e., nucleoporins and histones).” This is not a good option because a scientific paper tells us that the half-life of histones in the brain is only about 223 days, meaning that every 223 days half of the histone molecules will be replaced. So histones are not suitable for storing memories lasting decades.

Theory 2: “Some have suggested that DNA (or the epigenetic modifications on it) is the most suitable candidate for memory, being the cellular storage mechanism for other (lifelong) information.” For why this speculation is untenable, see the section entitled “Why Very Long-Term Memories Cannot Be Stored in the Cell Nucleus” in this post. Human DNA molecules have been exhaustively analyzed by multi-year projects such as the Human Genome Project and the ENCODE project, and no evidence has been found of human memories stored in DNA. See the "Why Long-Term Memories Cannot Be Stored in the DNA Methylome" section of this post for why DNA methylation is also an unsuitable possibility for memory storage. If DNA molecules stored memories, we would find that the DNA molecules in the brain of a dead person would vary a lot, with one DNA molecule in the brain being very different from another. Instead, all DNA molecules in the brain are basically the same, and are the same as DNA molecules in other parts of the body. 

Theory 3: “The late Roger Tsien recently proposed the notion that perineuronal nets, the extracellular matrices around neurons and synapses, provide the architecture for information.” Wikipedia tells us that these perineuronal nets are “composed of chondroitin sulfate proteoglycans,” but this paper tells us that the half-life of such molecules is only 10 days, making them unsuitable for a storage of memories lasting a lifetime. The idea behind this perneuronal nets speculation is that memory may be stored in a pattern of holes, like punch cards. The idea is absurd. IBM punchcards worked back in the 1960's because they worked with a punchcard reader which shined light through the punch cards. The brain has nothing like a punchcard reader to read information if it had information stored in such a way, and such a system only works with flat two-dimensional surfaces, not three-dimensional surfaces like that in the brain. There are two research papers that claim to have a result suggesting that perineuronal nets may be important in memory, but both do nothing to establish such a claim, because they both used fewer than 10 animals in some of their study groups (for a moderately reliable research result, 15 is the minimum number of animals per study group).

Theory 4: “Structures composed of short-lived components could constitute a long-term memory if the configurations were preserved by normal homeostatic replenishment.” To the contrary, there is certainly no “normal” bodily mechanism that might allow “structures composed of short-lived components” to store memories for decades.

Theory 5: “Other theories involve information storage or processing in microtubules, the long polarized helices of tubulin subunits that compose the cellular skeleton.” There is no evidence for such theories, and there is a good reason for rejecting them: the fact that there is high molecular turnover in microtubules. This paper tells us, “Neurons possess more stable microtubules compared to other cell types...These stable microtubules have half-lives of several hours and co-exist with dynamic microtubules with half-lives of several minutes.” So microtubules in the brain last less than a week, and are not any place that memories lasting decades could be stored.

Theory 6: “The physical connectivity within neural ensembles is a plausible new candidate substrate for memory information storage, with many merits, including robustness to insult, bioenergetic efficiency, stability of information storage in a potentially binary format, and a high capacity for informational content.” This is a completely different idea from all the previous things discussed, and the paper does not discuss it truthfully. Far from being a "plausible" idea having “many merits,” the idea has no merits. No one has any plausible idea as to how mere “physical connectivity” could be storing the complex things human remember. The paper refers us to the previously mentioned “What Is Memory? The Present State of the Engram” paper, but the sketch of the idea given in that paper does not present the idea in a coherent manner. We see a diagram showing what looks like a necklace of green beads as a representation of how "physical connectivity" could supposedly store information.  That's not a way to store complex information like humans learn. If there is “stability of information storage” in the connections between neurons, it is not the type of stability that would allow memories to be stored for years, let alone decades. The proteins in synapses have an average half-life of less than a week, and synapses themselves have a lifetime of less than a year. The research of Stettler suggests that synaptic connections do not last longer than about three months. In a paper he stated the following, referring to remodeling which would break any "connectivity pattern":

Depending on whether the population of boutons is homogeneous or not, the amount of bouton turnover (7% per week) has different implications for the stability of the synaptic connection network. If all boutons have the same replacement probability per unit time, synaptic connectivity would become largely remodeled after about 14 weeks.

A very important recent scientific paper is the paper “Synaptic Tenacity or Lack Thereof: Spontaneous Remodeling of Synapses.” The paper used the term “synaptic tenacity” for the idea that synapses (brain connections) are relatively stable. Making a devastating case against such a claim, the paper stated the following:

The aim of this Opinion is to discuss challenges to the notion of synaptic tenacity that come from general biological considerations and experimental findings. Such findings collectively suggest that synaptic tenacity is inherently limited, since synapses do change spontaneously and to a fairly large extent....It is probably unrealistic to expect that synapses maintain their particular contents and, by extension, their functional properties with pinpoint precision. This expectation is further challenged by the fact that synapses are not rigid structures but rather are dynamic assemblies of molecules (and organelles) that continuously migrate into, out of, and between neighboring assemblies through lateral diffusion, active trafficking, endocytosis, and exocytosis....Closer examination reveals, however, that properties of individual synapses, such as spine volume, presynaptic bouton volume, synaptic vesicle number, active zone (AZ) molecule content, and PSD protein content fluctuate considerably over these timescales....Imaging studies in primary culture indicate that synaptic configurations erode significantly over timescales of a few days....The findings summarized above indicate that synaptic tenacity is inherently limited or, using the terminology of Rumpel, Loewenstein, and others, that synapses are intrinsically ‘volatile’....When it comes to cognitive functions, long-term memory is one area where the notion of synaptic volatility raises perhaps some of the most challenging questions. In light of findings discussed in this Opinion article, and possibly others, age-old notions concerning relationships between histories of ‘elementary brain-processes’, connection strengths, and memory traces might need to be revisited; put differently (to paraphrase James, modern science might need to improve on this explanation.

The evidence presented by this “Spontaneous Remodeling of Synapses” paper is devastating evidence against the predominant theory of the brain storage of memories (that memories are stored by changes in synapse strength), and also Theory 6 mentioned above, that memories are stored by some “connectivity pattern” created by synaptic connections. Neither theory can be true if synapses have the kind of volatility described in the paper.

In a 2010 book two neuroscientists state that they are “profoundly skeptical” about the main theory of a physical storage of memory, but suggest that they have nothing like a substitute theory to offer:

We take up the question that will have been pressing on the minds of many readers ever since it became clear that we are profoundly skeptical about the hypothesis that the physical basis of memory is some form of synaptic plasticity, the only hypothesis that has ever been seriously considered by the neuroscience community. The obvious question is: Well, if it’s not synaptic plasticity, what is it? Here, we refuse to be drawn. We do not think we know what the mechanism of an addressable read/write memory is, and we have no faith in our ability to conjecture a correct answer.

I submit that the reason for such hesitancy is that there is no theory of a physical storage of memory that can stand up to careful scrutiny, no theory that can explain both memories that can form instantly and the fact that we can instantly retrieve memories of things learned or experienced long ago. Deprived of any credible neural explanation, we should conclude that human episodic and conceptual memory must be a spiritual or psychic or soul phenomenon, not a neural phenomenon.

One of the many reasons for rejecting claims that memories are stored in brains discussed at this site is the essentially instantaneous speed at which humans are able to remember very rarely recalled pieces of information. You say to me, “Dizzy Dean,” and I may in less than two seconds start saying, “eccentric St. Louis Cardinals pitcher, in the 1930's,” even though I haven't thought or read about Dizzy Dean in decades. How could a brain achieve this effect through reading just the right little spot where the information was stored, which would be like instantly finding a needle in a mountain-sized haystack, given a million little spots in the brain where a memory might be stored?

One major reason why it seems hard to believe that the brain could achieve instant recall is that neurons are slow. Information is passed around in a brain at the slow speed of about 100 meters per second, which is only the tiniest fraction of the speed at which electrical signals move about in a computer. Based on this fact we should consider a brain absolutely incapable of performing memory recall as quickly as humans do.

It is often claimed that the brain “makes up” for its slow speed of nerve transmission by being “massively parallel.” The claim that the brain is massively parallel is false. In the computer world, a computer system is massively parallel if it consists of multiple CPUs or central processing units, each of which is running a computer program. We know of nothing in the brain that acts like a computer CPU, nor do we know of anything like a program that the brain runs to compute. It is false to claim that each neuron is like its own CPU. Neurons do not run any program like a computer CPU does. So we cannot at all overcome the problem of low signal transmission speeds in the brain by claiming that the brain is “massively parallel.” It is true that the brain consists of many neurons working together, but that does not make the brain massively parallel. My television set has many transistors working together, but my television set is not massively parallel; and my arms have many cells working together, but that does not make my arms massively parallel. The brain does does not consist of multiple programs running together on multiple processors, and we don't even know of a single program running anywhere in the brain.

Let us imagine some weird group of conspiracy theorists who maintain that secret information from the World War II era is stored in the leaves lying about in modern Germany. If the theorists were to maintain that such information is written on individual leaves, you could easily show the absurdity of the theory by pointing out that leaves don't last much longer than a year. If the theorists were to maintain instead that it was the arrangements of the leaves that stored the information, with one type of leaf pile representing one thing and another type of leaf pile representing something else, you could also show the error of the theory by pointing out that leaf piles are unstable and ever-changing, being blown around by the wind. Similarly, the protein turnover, synaptic volatility and synaptic remodeling discussed in the "synaptic remodeling" paper above are constraining effects as powerful as the short lifetimes of leaves and the instability and impermanence of leaf piles. Anyone claiming that memories persist in synapses for 50 years is advancing a claim as unbelievable as the 50-year "leaf storage" claims of such conspiracy theorists.