Quanta Magazine is a widely-read online magazine with slick graphics. On topics of science the magazine again and again is guilty of the most glaring failures. Quanta Magazine often has assigned its online articles about great biology mysteries (involving riddles a thousand miles over the heads of PhDs) to writers identified as "writing interns." The articles at Quanta Magazine often contain misleading prose, groundless boasts or glaring falsehoods. I discuss some examples of such poor journalism in my posts here and here and here and here.
The latest example of false news in Quanta Magazine is an article with the bogus headline "A New Type of Neuroplasticity Rewires the Brain After a Single Experience." The claim is BS, pure baloney. Anyone familiar with the structure of the brain should instantly realize what nonsense this headline is. Neurons have fixed positions in the brain. Synapses are like roots in a dense forest, roots that lock trees into fixed positions in the forest (but instead of locking trees into their positions in a forest, synapses help lock neurons into their positions in the brain). Synapses are things almost as slow-changing as the roots of trees in a forest. So physically the idea that a brain could be instantly rewired is nonsense.
The article starts out with this untrue claim: "Every experience we have changes our brain, the way a ceramicist reshapes a slab of clay." To the contrary, an experience does not change a brain. The analogy that the brain is like a lump of clay onto which impressions are written by experiences (like letters being written by the earliest cuneiform writers in Mesopotamia) is an extremely misleading analogy with no evidence to support it. The brain has nothing like a stylus that could write such impressions. And no trace of any such impressions can be found. Microscopic examination of brain tissue (which has been done very abundantly) has never produced the slightest trace of anything anyone learned.
We have this vacuous attempt to explain memory, not corresponding to any physical reality in the brain: "This plasticity, the quality of being easily reshaped, makes the brain really good at learning — a quintessential process that allows us to remember the plotline of a novel, navigate a new city, pick up a new language, and avoid touching a hot stove." It is not correct that brains are "easily reshaped," and it is not correct to suggest that brain structure changes after learning. Scan a brain before and after 8 hours of school learning, and you will see no difference.
The writer then tells us a myth with no basis in fact, stating this:
"Recently, neuroscientists described a new form of neuroplasticity that might be helping the brain learn across a timescale of several seconds — long enough to capture the behavioral process of learning from a single experience. In two recent reviews, published in The Journal of Neuroscience
and Nature Neuroscience, they describe 'behavioral timescale synaptic plasticity,' or BTSP. This type of learning in the hippocampus, the brain’s memory hub, is caused by an electrical change that affects multiple neurons at once and unfolds across several seconds."The claim that something called "behavioral timescale synaptic plasticity," or BTSP is a "type of learning" is a claim without any basis in fact. Before looking at one of the reviews cited in the quote above (the one that is not behind a paywall), I must give some prefatory description of the social construction of discovery legends in today's neuroscience.
Research in cognitive neuroscience research is dominated by low-quality studies. The study here concludes, "Our results indicate that the median statistical power in neuroscience is 21%." This is an abysmal number, an appalling figure. It has long been said that in experimental research, the goal should be a statistical power of 80%, which roughly corresponds to a likelihood of 80% that the result will be replicated. A study with a statistical power of 21% is a low quality study that is likely to be announcing a false alarm. When a research field has a median statistical power of 21%, that means half of the studies have a statistical power of 21% or less. If such an estimation is correct, it means the great majority of neuroscience studies report results that are unreliable or untrue.
A neuroscientist wishing to gain fame and funding may do some low-quality research, and claim his research is a discovery of some new effect for the first time. The neuroscientist may coin some name for this alleged effect, perhaps using some acronym. Whether that name is forgotten and never repeated may depend on whether other neuroscientists are willing to repeat the observational claim, and whether other scientists are willing to try to replicate the effect. If a claim of a discovery "presses the buttons" of neuroscientists by claiming something neuroscientists are eager to see, the original observer's discovery claims may be repeated by other scientists. But it will often be the case that there is no sound warrant for either the original observational claim nor similar claims by other scientists. An original low-quality paper reporting some effect may use poor experimental methods, and equally poor experimental methods may be used by others who claim to see the same effect.
Consequently we should never "take it for granted" that something is true, just because some scientific paper says that scientist X claimed to see such a thing, and that also scientist Y and scientist Z claimed to see it. The social construction of groundless triumphal legends is extremely common in neuroscience literature. The standards for getting a neuroscience paper published are low. Junk research is published every week, and low-quality experimental papers are published every month. So you must always go back to the papers being cited, and look at them critically, and ask: was their ever any decent evidence observed here?
Let's do that with one of the two papers cited by the Quanta Magazine article above, the one not behind a paywall. The paper is an extremely misleading review article entitled "Behavioral Timescale Synaptic Plasticity: A Burst in the Field of Learning and Memory" which you can read here. The paper is short on descriptions of reliable observations of things naturally occurring and very long on triumphal narrations, most of which are groundless boasts hailing supposedly marvelous accomplishments of the most poorly designed and low-quality scientific studies. The paper's chief triumphal narration is the claim that something called behavioral timescale synaptic plasticity (BTSD) was discovered in 2017 by Katie C. Bittner and others.
We hear these claims by the "burst in the field" paper about this BTSD:
- "It is triggered by occasional dendritic plateau potentials associated with a burst of firing in the soma." Neurons fire unpredictably at a rate between about 1 time per second and 100 times per second, with their firing rates varying unpredictably. So anyone analyzing noisy, variable data on neuron firings might be able to find "occasional dendritic plateau potentials associated with a burst of firing in the soma," even if there is no such thing as BTSP. Similarly, anyone analyzing cloud formations sufficiently will be able to find occasional cloud clumps with this or that shape.
- "BTSP operates on the timescale of seconds rather than milliseconds and can therefore support associative learning over temporal delays relevant to behavior." If this alleged BTSP occurs quickly, that is no reason at all for thinking it has anything to do with any type of learning.
- "It leads to large changes in synaptic strength, enabling fast remodeling of neuronal representations that may support one-shot learning." There is no robust evidence for representations in either neurons or synapses. The size and strength of synapses vary randomly over days and weeks. An increase in the strength of some synapses is never evidence that anything is being represented. And you cannot actually have "large changes in synaptic strength" being produced by anything operating on a timescale of seconds. There is no robust evidence that "large changes in synaptic strength" ever naturally occur on a timescale of seconds.
The paper claims this: "A different kind of plasticity called behavioral timescale synaptic plasticity (BTSP) has recently been uncovered in area CA1 of the hippocampus (Bittner et al., 2017; Milstein et al., 2021) and has properties that appear to solve many of the aforementioned limitations of Hebbian plasticity (Magee and Grienberger, 2020)." Bittner's 2017 paper supposedly first observing this BTSP alleged effect is behind a paywall. But we can look at this 2021 paper by Milstein, co-authored by Bittner. It is a very low-quality paper that you can read here, one with the misleading title "Bidirectional synaptic plasticity rapidly
modifies hippocampal representations." There is no robust evidence for the representations claimed.
This 2021 paper co-authored by Milstein and Bittner starts out by reciting many an unfounded legend and dubious dogma of neuroscientists. Then we have in Figure 1 some actual fresh observational data. The data is nothing remotely resembling compelling observational data. We have data from a single mouse that was on a treadmill. The graphs are not really hard data, because we have references to "place fields" that are social constructs of neuroscientists. Here the paper makes some claimed observational result that no one should take seriously or pay attention to unless it involved results with at least 15 or 20 animals per study group. But the study group consists of a measly one animal. The study group size is not even 10% of what it should be for reliable evidence to be claimed.
Figure 2 is just as laughable as evidence. Since it has a caption of "mouse running," it seemingly also is data from a single mouse. Figure 3 fails to mention any study group size larger than 1. We have a graph mentioning "synaptic weights," one seeming to show some kind of increase. But the graph has no scale. No actual measurement of synaptic weights is occurring. Why of course -- synapses are things too tiny to be weighed with any accuracy.
Later in the paper we have an indication that no reliable measurement of synaptic weights was occurring. We read, "We modeled changes in synaptic weights as a function of the time-varying amplitudes of these two biochemical intermediate signals, ET and IS." So apparently something else easier to measure was being measured, and the authors were engaging in the very dubious business of claiming that this other thing was some indication of synaptic weights. That sounds rather like someone trying to deduce the weight of someone's meals by how much they spent on groceries this week -- not a reliable way of doing things.
Nothing reliable is being done here to show that this claimed "Behavioral Timescale Synaptic Plasticity" naturally exists, or that it can produce rapid changes in synaptic strength. And even if you were to show such a thing, that would do nothing to explain instant learning, since changes in synaptic strength are not credible explanations of how newly learned information could be stored. For further evidence of the low quality of Wilstein and Bittner's 2021 paper "Bidirectional synaptic plasticity rapidly modifies hippocampal representations," we need merely search for whether a sample size calculation was done. The paper confesses, "Sample sizes were not determined by statistical methods." Why of course. Since laughable, ridiculous sample sizes such as only one mouse were used (rather than decent study group sizes such as 15 or 20 mice per study group), the authors did not do a sample size calculation, which would have revealed some ridiculously low statistical power way, way below 25%.
By citing Wilstein and Bittner's 2021 very low-quality paper "Bidirectional synaptic plasticity rapidly modifies hippocampal representations," the review article "Behavioral Timescale Synaptic Plasticity: A Burst in the Field of Learning and Memory" has given us another example of what constantly goes on in the dysfunctional world of neuroscience research: paper authors citing very low-quality research as evidence of some effect they are arguing for, with the authors seeming to apply no critical scrutiny before citing a paper.
The Quanta Magazine article and some of the papers cited by the papers (mentioned above) refer to a 2015 paper co-authored by Bittner and Jeffrey C. Magee, entitled "Conjunctive input processing drives feature selectivity in hippocampal CA1 neurons." While the study makes reference to a "normative" pool of 21 mice, the study's claims of detecting something are based on a way-too-small study group size of only 6 mice. It's another very low-quality paper failing to provide any decent evidence of "Behavioral Timescale Synaptic Plasticity." The authors confess that "no statistical methods were used to predetermine sample sizes," which is always a damning confession in a scientific paper of this type, kind of a "we were too lazy to act like good scientists" confession. But (paying no attention to quality factors) the Quanta Magazine senselessly treats the study as if was something important, and has a big photo of Magee. This is typical for Quanta Magazine, which seems to never pay any attention to whether neuroscience studies are meeting the hallmarks of robust, well-designed science.
The paper "Behavioral Timescale Synaptic Plasticity: A Burst in the Field of Learning and Memory" has graphs from the year 2025 paper "Synaptic plasticity rules driving representational shifting in the hippocampus" that you can read here. That paper mainly refers vaguely to "mice" without mentioning exact study group sizes. But occasionally the paper does mention the exact study group sizes, which were way-too-small study group sizes such as only 4 mice, only 7 mice and only 9 mice. We read, "CA1 recordings were done in 4 mice, CA3 recordings in 7 mice and optogenetic experiments in 9 mice." These study group sizes were way-too-small for the paper to be taken as serious evidence of anything. No paper like this should be taken seriously unless 15 or 20 animals per study group were used. We have the damning confession in the paper that "No statistical method was used to predetermine sample size." If the paper authors had acted like good scientists by doing such a calculation, they would have found out how inadequate were the study group sizes they used. We also read, "Investigators were not blinded to CA1 or CA3 groups." This is a crucial defect for a paper like this. We have here a very low-quality example of a Questionable Research Practices study, one that fails to provide any good evidence for "Behavioral Timescale Synaptic Plasticity."
So it seems the review article "Behavioral Timescale Synaptic Plasticity: A Burst in the Field of Learning and Memory" (which you can read here) is a paper that fails to cite any convincing studies showing any such phenomenon as "Behavioral Timescale Synaptic Plasticity." I reach this conclusion not based on a readership of all papers cited by that paper, but by looking at the studies discussed above, which were all very low-quality papers very badly guilty of Questionable Research Practices. The paper provides no robust evidence that scientists have demonstrated any such thing as any natural ability by which synapses could be instantly or very quickly strengthened.
The review article "Behavioral Timescale Synaptic Plasticity: A Burst in the Field of Learning and Memory" is a mere review article, not a systematic review. In scientific literature, systematic review articles are articles following a clear methodology in regard to which papers are to be cited as evidence, a quality filter clearly defined in the paper. A mere review article involves citing any papers that the authors wish to cite, without the papers being subjected to a quality filter stated in the paper. In today's neuroscience literature there is a plethora of misleading review articles citing poor-quality papers.
We get an indication in the review article "Behavioral Timescale Synaptic Plasticity: A Burst in the Field of Learning and Memory" that what is typically going on in the results it reports are not natural occurrences, but instead artificial occurrences produced by scientists doing special fiddling. In the article we read this: "The main approach, introduced by the Magee lab (Bittner
et al., 2017), is to artificially induce BTSP with a long-lasting
high–amplitude depolarization of the soma (typically a 300 ms
600 pA current injection) triggering a somatic CS (which is
assumed to reflect a dendritic plateau), preceded or followed
by synaptic activity within a few seconds time window
(Fig. 3B)."
So mainly what is being reported are artificial results produced by current injections -- experimenters zapping brains with electricity or electrical currents. The paper then says that this can be either done "in vitro" (that is, using tissue detached from an organism's body) or "in vivo" (observing something inside a living organism). But we are told that the "in vivo" observations require artificial experimenter manipulations such as "current injections" or "optogenetic stimulations." Both are artificial types of brain zapping. Observations requiring such energy injections by experimenters are not evidence of something naturally occurring in the brain.
I propose the term "electromisrepresentation" to describe misleading narratives of this type. We can define electromisrepresentation as the artificial production of brain effects by methods such as electrical stimulation, combined with a misleading narrative trying to suggest that the resulting effects can explain natural human capabilities. Electromisrepresentation has massively occurred in discussion of so-called long-term potentiation or LTP.
The Quanta Magazine article based on this scientific paper is also very misleading bunk, an article that attempts to persuade us of the existence of something for which there is no robust experimental evidence. A very bad example of groundless narration, the paper is full of untrue statements claiming magnificent accomplishments from scientists who actually ran very low-quality studies deserving mainly scorn because of multiple methodological sins such as the use of way-too-small study group sizes, the lack of a blinding protocol, the lack of pre-registration, and an abundance of unreliable claims about the physical state of things (synapses) too small to have their physical state reliably measured. Occasionally the Quanta Magazine article has an indication of what baloney it is shoveling, such as when it says, "There’s still much unknown about BTSP, especially the mechanism, which Madar said is 'quite speculative.' ”
But that's par for the course in the untrustworthy world of today's neuroscience research, where neuroscientists these days boast like crazy about doing all kinds of wonderful things that were not actually done, because decent scientific procedures were never followed, and the experiments were so poorly designed and guilty of so many defects.
There are two very strong reasons for rejecting all claims that there is good evidence that there are ever any quick natural increases in synaptic strength:
(1) The intrinsic unreliability of all attempts to measure the strength of synapses, given the incredibly small size of synapses, which makes all attempts to measure their strength dubious and unreliable. The largest parts of synapses (their clefts) are about 500 to 1000 times smaller than the largest part of a neuron (its soma or main body).
(2) The intrinsic implausibility of any claims that synapses could naturally be quickly strengthened, given the fact that any synapse strengthening would require new protein synthesis, a process that takes minutes or hours.
Scientists have never produced any credible tale to explain how either instant learning or learning of any type could occur in a brain, which has no components having any resemblance to a system for storing learned information. You would never show information storage or a storage of learned data or experiences by merely showing an increase in the strength of something. No well-designed and robust scientific studies have ever produced any compelling evidence that learned information has been physically stored in brains. Claims about LTP arose from the type of artificial brain tissue zapping described above, with researchers ignoring that what was occurring did not correspond to natural events in the brain. Microscopic examination of brain tissue has never yielded the slightest trace of anything anyone learned or experienced -- not a single sentence, not a single word, not even a single character or letter or number or even a single pixel of anything anyone saw.
Scientists have reliably determined that synapses are built of proteins that have average lifespans of only a few weeks, roughly a thousandth (1/1000, that is .001) of the maximum amount of time that humans can remember things, which is about sixty years. Besides utterly failing to explain how a brain could do memory storage, and how memories could persist for decades, scientists have utterly failed to explain how a brain could do instant memory retrieval, or memory retrieval of any type. We know the types of things that allow for instant retrieval of stored information: things such as addresses, indexing and sorting. No such things exist in the brain. The world of neuroscientist claims about memory is a world of fantasy and pareidolia, in which neuroscientists eagerly hoping to see things claim to see the faintest evidence of such things, like some wanderer in the desert eagerly scanning the far horizon five miles away and claiming to see water on the far horizon (although only a mirage is there, or only "see what you yearn to see" pareidolia is occurring).
No comments:
Post a Comment