Talk:Long-term potentiation/Archive 1

Drug usage to stimulate and increase learning

I was reading some information on drugs used to increase the learning inside of animals that has to do with LTP. I don't exactly know too much about this subject since it seems somewhat brand new to me on a biochemical level.

Does anyone have information relating to the withdrawal effects on subjects that had taken in the drug? Did they show a lack of an ability to learn afterwards? What side effects did such a drug have? --Cyberman 17:56, 10 October 2005 (UTC)

Tetanic stimulation

What is "tetanic stimulation"? Is it just a high-frequency sequence of stimulations? AxelBoldt 21:41, 9 Sep 2004 (UTC)

Yes. --Diberri | Talk 22:05, Sep 9, 2004 (UTC)

Ok, what kind of frequencies are we talking about here? AxelBoldt 04:52, 11 Sep 2004 (UTC)

IIRC it's often a 5 x 100 Hz protocol (five 100-or-so-millisecond pulses at 100 Hz). I'll try to find a reference and post it here. --Diberri | Talk 23:41, Sep 11, 2004 (UTC)

Rewrite

I've just done a massive rewrite of the page in an effort to bring everything up-to-date with current studies of NMDA-dependent hippocampal LTP. In doing so, I've referred to journal articles using the syntax (PubMed ID), but haven't had time to apply the suggestions of Wikipedia:Footnotes. (If one checks the page history, they'll find that I did try to add footnote links, but I used the wrong CSS ID in the <sup> tags, so I got frustrated and gave up, at least temporarily :-) Incidentally, if anyone wants to tackle the task of adding ap

propriate footnotes (and citations under the "References" section), by all means feel free -- I'd very much appreciate it! For now I hope that readers find the rewrite more useful than the previous version. Best, David Iberri | Talk 00:26, Nov 16, 2004 (UTC)

Hi David, thanks for letting me know about the rewrite. I read it and learned a lot, again. Here are a couple notes that came to mind. I don't really know enought to fix these myself.

• In the earliest experiment showing LTP, was it done in vivo or in vitro? Generally, a section about how one conducts LTP experiments today in vivo and in vitro would be nice.
• Like I mentioned above, it might be good to give the frequencies typically used in experiments.
• In the section about non-associative LTP, it is mentioned in passing that associative LTP requires two stimuli, but the section about associative LTP does not make that clear.
• In the section about properties, are we really only talking about NMDAR-dependent LTP? If not, it would be good to give the section with the high level properties first, before going into biochemical details.
• I assume CREB only acts postsynaptically? This should be made clear. NO acts on both cells?
• The pictures of the synapses would be clearer if presynaptic and postsynaptic cells were labeled.
• The picture of the brain is a bit unclear: where's front and back?

Cheers, AxelBoldt 01:30, 29 Nov 2004 (UTC)

If I may be of help with your some of your questions (Synaptidude).

1. Bliss and Lomo did their original experiments in vivo in the dentate gyrus. I can't remember if if was rat, guinea pig or rabbit. (I checked, it was rabbit)
2. Typically 100 Hz tetanus for 1 second is the "standard" LTP inducting stimulus (often applied multiple times). Theta burst where a series of short tetani spaced to take advantage of inhibitory refractoriness is another way. I'd be willing to bet that today the most common method of inducding LTP is the so-called "pairing" protocol, where the postsynaptic cell is artificially depolarized while low frequency presynaptic stimuli are applied (pairing postsynaptic depolarization with presynaptic stimuli).
3. I've not yet read the section on non-associative LTP, but I'm assuming this means the LTP that has been seen by ultra-high frequency tetanus. This is believed to be mediated by postsynaptic voltage-dependent calcium channels, and thus requires only postsynaptic depolarization.
4. NMDA-dependent LTP is the "primary" form of LTP. Almost all the known LTP in the brain is NMDA-dependent. Non-associative LTP is not NMDA dependent (although there is some controversy here) and makes only a tiny contribution to changes in synaptic strength. Mossy fiber LTP is entirely presynaptic (again, some controversy, but I'd bet heavily on this) and not NMDAR dependent, but it occurs only in anatomically restricted (and synapse type restricted) areas.
5. I'm going to leave CREB alone. I'm not a big supporter of the CREB hypothesis. I believe the conclusions on this have outrun the actual supporting data.

Question

Hi, i am layman and i am not sure whether i understand this sentence right:

Though a single presentation of the stimulus is USUALLY not sufficient to induce LTP, repeated presentations cause the postsynaptic cell to be progressively depolarized.

Does it mean that that some things are written to our memory completely by chance? that some quick glance can be stored in our memory just because of rigtht polarization of the postsynaptic cell? and on the other hand something we want to remember may not be joined by right polarization and therefore we just forget it?

thank you for a respond tomas

Whether or not chance has any effect on LTP induction is largely a philosophical question. In science, when the parameters aren't sufficiently understood, variations are often attributed to chance for convenience. I suspect this is because scientists are wary of violating Occam's Razor. Anyway, by usually I was trying to account for some experiments that have elicited LTP with a single stimulus, while the vast majority of trials require multiple high frequency stimuli; I didn't intend to imply that chance alone can induce LTP. At most, chance plays a part in LTP induction, but again that's largely because many (if not most) of the details of LTP are as yet unknown.

As for a quick glance coupled with the right postsynaptic polarization leading to learning, remember that LTP and learning are not equivalent. Learning is behavioral (and thus deals with the whole organism) while LTP is cellular (and thus deals only with cells). A glance is the result of millions upon millions of cells in the visual cortex and elsewhere processing visual stimuli. The brain's got ways to reduce the contribution of single cells to memory formation. (Some of these ways prevent us from storing things like frame-by-frame memories. Just imaging how devastating things would be if the whole of your memory was consumed by the individual visual frames you perceive throughout your day -- you'd never be able to commit any useful information to memory. Hardly an adaptive strategy!) So a single postsynaptic cell being polarized just right wouldn't have a significant impact on whether a particular split-second of our vision gets stored away. (At least in humans. In other animals with simpler nervous systems, LTP in a single cell can have remarkably significant effects; cf Aplysia.)

Now if we talk about single cells rather than organisms, and LTP (cellular) rather than learning (behavioral), then things are a bit different. Single synapses, and thus single cells are very important and the polarization of an individual cell is a big contributor to LTP. But even then, you'd be hard-pressed to find a reputable source that would attribute LTP in single synapses to chance alone. Best wishes, David Iberri | Talk 18:21, Dec 27, 2004 (UTC)

Question II

thanks for a quick respond- i am writing some school work about memory and i have got a few questions about LTP( maybe answers are written in your article but i am not from english speaking country and therfore i do not understand everything:-(

1. does LTP have proven and clear connection with learning and information storing process? am i right if i say that increased calcium level in dendrite results (after many chemical reactions, and if everything goes right) into creation of neuron network responsible for remembering of some piece of information ???

2. this neuron network can be located anywhere in the brain according to character of stored information-but if i understood it right not all parts of brain have LTP ability

3. how are those neurons in this network connected? and how are new neurons representing new information related to original added to this network? does LTP still plays role here?

thanks for your understanding. i know that my questions may look simple but i am only high school student and so i am doing as much as i can, to get into this problem...

so if you have some time please answer my questions it will help me a lot:-) THANK YOU VERY MUCH --R0SI 21:51, 27 Dec 2004 (UTC)

You're very welcome for the responses. I'm just one of many Wikipedians who is more than happy to help. As for not understanding everything in the article, I'm partially to blame; it's written from too technical a perspective, and that should be cleaned up when possible. When I get some more time I'll be sure to clarify things as best I can. (Thankfully, this being a wiki, the onus isn't completely on me to make sure the article is in tip-top shape, though :-) For now, I'll do my best to answer your questions:

1. I tried to address this a bit in the article under the section LTP and behavioral memory. As far as I know, no one has yet demonstrated a causal role between LTP and behavioral learning. I'm making a point to say behavioral learning here, not just any type of learning, because in the laboratory, scientists can induce LTP to cause some types of learning. But nobody's shown that LTP is the way organisms learn on their own.
   
   There have been plenty of correlational studies looking at LTP's causal role in behavioral learning, though. In these experiments, scientists often give animals drugs that alter LTP before teaching them a task. Then they compare the changes in LTP to the changes in learning. For example, blocking the NMDA receptor (which in turn blocks LTP) impairs spatial learning in rats. So when LTP is impaired, spatial learning is impaired too. It's tempting to say that LTP therefore causes memory formation, but that would be incorrect. A common saying is that "correlation does not imply causation," meaning that just because two things happen together doesn't mean that one causes the other (see spurious relationship for more on that). So no, there hasn't yet been any "proven and clear connection with learning" unfortunately.
   
   As for the question about calcium, are you asking if calcium results in the formation of new neurons in a network capable of storing memories? If so, there's very little new neuron growth in humans after about 8 years old, so you can see that new cell growth couldn't be the way memories are stored. Current thinking is (roughly) that memories are stored in synapses, which are dynamic (i.e. they be created/destroyed/modified relatively quickly), not neurons themselves. So you might say that calcium induces LTP and (presumably) ultimately alters synapses in existing networks for memory storage.
    
2. You're right, not all areas of the brain are capable of LTP. But LTP has been demonstrated in many areas of the brain implicated in learning (e.g. amygdala, hippocampus, cortex). One popular working model of learning says that the hippocampus is the gateway to long-term memory; once the hippocampus has registered a memory destined for LTM, it propagates that memory to relevant portions of the cortex. So for example, when a rat learns how to navigate a maze, it might keep a visual representation of the maze in the visual cortex, a memory of what sounds it encountered when navigating the maze in its auditory cortex, etc. (I presume that's what you mean by the brain storing memories in regions "according to character of stored information"). When the memory is recalled, all of these various cortical areas are activated to reconstruct the rat's memory of the maze.
    
3. I'm not sure I understand your question about how neurons are connected. Are you talking about neurons in the cortex where these long-term memories are stored? If so, nothing's changed here; the connection is still made through chemical synapses. The changes that have occurred in these neurons is attributed to LTP, so you'd expect to find enhanced activity at existing synapses, growth of new synapses between existing cells, etc.

Ideally most of this would be covered at Memory#The physiology of memory or even branched off into its own article, but no one has got around to it yet. Until that day (and even after!) feel free to keep the questions coming, and please don't think you're asking anything invalid or too simple. Your questions are absolutely always welcome. Best wishes, David Iberri | Talk 22:59, Dec 27, 2004 (UTC)

Three more questions

thank you for your answers for my question- they really helped ma a lot. but i have got three more

1. is it correct if i say that one HYPOTHESIS, explaining process of learning is LTP...etc. hypothesis is right word even if you said that connection between LTP and memory formation was not proven and it is probable that those two things have nothing in common?

2. that tetanic stimuli is in fact visual/audio stimuli (something we saw/hear) but encoded into sets of signals with frequency about 100Hz? this happens as information pass through eye or ear?

3. you write that hippocampus propagates memory to relevant portions of the cortex-it is responsible for connecting memories of various characters(audio/visual). what is the mechanism of this. i mean if those informations are stored- one in visual and one in auditory cortex the neurons or esp. dendrites and axons have to be few centimeters long. am i right?

nice day and thak you for all--R0SI 18:52, 28 Dec 2004 (UTC)

1. Yes, one hypothesis is that LTP is the underlying physiological mechanism of behavioral learning in animals. In my opinion, it's highly unlikely that LTP and learning are unrelated.
2. That's a good question. LTP experiments in the laboratory typically use tetanic stimulation in the range of 100Hz to induce LTP, but there are few ares of the brain capable of reaching those levels of stimulation in vivo. Recently the hippocampus, if I'm not mistaken, has been observed using these same frequencies, but whether it's for the purpose of LTP is still uknown.
3. Very little work has been done in the area of memory recall (compared to the amount of work done on memory formation), so it's difficult to point to a precise mechanism. What I can tell you is that the hippocampus' main output is the subiculum and entorhinal cortex, which communicate messages from the hippocampus to the rest of the cortex. The length of these axons would be on the order of several centimeters. Dendrites, on the other hand, would tend to be a bit shorter. --David Iberri | Talk 21:00, Dec 28, 2004 (UTC)

One question

What is plasticity-related proteins? --Daxue | Talk 04:20, 8 Jan 2005 (UTC)

A: This link is useful for understanding plasticity or Synaptic plasticity. --Daxue | Talk 09:33, 8 Jan 2005 (UTC)

Greetings

Hi Big Dave. I'm new to WP, but I naturally zeroed in pretty quickly on your LTP article (being a scientist who works in this area). You've done a really nice job with it. I hope you don't mind if I pop in now and again to edit. I've done some already. This is what WP is all about, right? Anyway, just wanted to say "I come in peace ;-)"

Hey there and welcome to Wikipedia! Thanks for the compliments. I'm really looking forward to your edits. It'll be great to have an expert at the helm ;-) --David Iberri | Talk 00:13, Jun 23, 2005 (UTC)

LTP Worldview

Hey Big Dave. Thanks for the help with the wiki defs.

As I said before, I think you've done a really nice job on the LTP article. It is well-written, well-documented, well-organized etc. Yet, as I read it, I find myself frustrated by something I couldn't quite put my finger on. As I corrected a few of what I considered "factual errors" in the piece, I realized that the reason I find myself disagreeing with many of the "facts" is that you've written the piece mostly from a Kandelian point of view. I notice that 10 out of 31 of your notes are to Kandel references. As a researcher in this field, and someone who has a pretty good feeling for its pulse, I can tell you that most of the really reputable labs working on LTP think that when it comes to LTP, that Kandel is........what's the scientific term for this?.........oh yeah!........Full of Crap. Many if not most of his LTP findings are believed by no one but himself and his own people. His papers are rife with fundamental experimental errors, overreaching interpretations, data bent to fit the hypothesis, etc. IMHO, in writing a review of the LTP literature, you could throw out everything produced by Kandel's lab and your review would be better for it. I find myself extremely tempted to rewrite large tracts of your article from a Nicollian, Malenkaian, Malinowian, Collingridgian, Madisonian, Huganirian, Kauerian, Issacian, etc-ian.....ok from a more mainstream and broader point of view. Part of the trouble is that you are writing about things that still have ragged edges. That scientists are still unsure about or are still arguing about. But on the other hand, I find myself reluctant to tear apart what is really a beautiful article (with above noted exception). You obviously put a lot of effort into it. So what I'm going to do instead is this: when I find something I don't like, I'm going to put a link in your article called "point of controversy" that will refer to a separate article where the pro's and con's of a particular issue are disussed.

If you'd enjoy taking a stab at this yourself, I'd suggest that a good goal to shoot for is too achieve a two to one ratio in references by Nicoll vs references by Kandel. Nicoll and his scientific offspring have done more to nail down the mechanisms of LTP with quality reproducible experiments than just about anyone.

I'll just leave you with this thought. Just because it's in the literature doesn't mean it's true. This may particularly be the case for the LTP literature where it sometimes seems that pretty much everything has been shown to be both true and untrue at least once. This may more often be the case when the senior author is an influential and powerful scientist (maybe even a Nobel Prize winner!) who can pretty much bully any paper he wants into a good journal over the objections of his peer reviewers.

Synaptidude

Thanks for your comments. This article will benefit greatly from the experience of an active LTP investigator like yourself -- one who not only has a better understanding of the processes of LTP, but also its culture and controversy. As my only exposure to LTP was in the classroom, I have very little experience with the latter two -- which is no doubt evident from my contributions to the article.

I am aware, however, of some of the strong disagreements between several LTP camps. Interestingly (or not), I studied under one of Kandel's former postdocs, David Glanzman, who (as he tells it) fiercely disagreed with Kandel's idea of exclusively presynaptic mechanisms of LTP induction and expression. So from the get-go I was skeptical of Kandel and his group, becoming more so after reading Hawkins' dismissive reply to Lin and Glanzman's work that suggested a Hebbian contribution to LTP in Aplysia. All that's to say I had a pretty strong bias against Kandel as I began writing this article. My bias wasn't to the point that I'd believe that "no one but himself" agrees with Kandel's research (that's a bit of an overstatement), but I was and am doubtful of some of his work nonetheless.

Yet as you note, Kandel's work is ubiquitous and Kandel is considered an authority on LTP research (you're right -- the Nobel Prize doesn't hurt). That his research winds up in nearly every peer-reviewed journal, textbook, and mainstream publication is a testament to Kandel's pervasiveness. And it's these facts that led me to doubt my own anti-Kandelian tendencies such that I felt a need to pull back on the reigns in an effort to maintain NPOV. Maybe I pulled too far? ;-)

Certainly publication does not imply truthfulness. I think this is merely a cultural misunderstanding coupled with the difficulty in writing an encyclopedia article about a debated topic like LTP. I understand your reluctance to tear apart this article, but that's the spirit of Wikipedia! :-) So rather than branch out discussions of controversial points (e.g. LTP induction), please incorporate those into the article proper. Likewise, if you find a section is particularly riddled with inaccuracies, move it to the talk page (possibly replacing it) for discussion. Also, I agree that we should try to incorporate more work from Nicoll (et al, particularly Malenka and Malinow -- their reviews are especially good).

Incidentally, I wasn't aware that so many of Kandel's experiments aren't reproducible. How (references) have you determined this? This, along with many of the other points you mentioned, is making me think a "Controversies in LTP research" section would be quite appropriate.

Cheers, David Iberri | Talk June 28, 2005 20:47 (UTC)

reply

Hey Big Dave,

You are right about one thing (well more than one thing ;-). My "no one but him" statement was an overstatement. I was thinking about this before I read your reply. I'm trying to think of a way of putting this that would be more accurate, and I suppose it would be: Many if not all of the reputable labs working at the center of the mechanism of NMDA-dependent LTP, believe that Kandel's group is wrong about most of what they've published on LTP.

Here's some backup.

1) On the question of the expression of LTP being pre- or postsynaptic: This debate raged for years without much in the way of conclusive evidence one way or the other. In the last five years, it's pretty much become everyone against Kandel and Stevens (and even Stevens seems to be backing off). I could go through and give you specific references, but there are many and that would be tedious.

But on another LTP related issue, here are some good examples of failure to replicate.

Example 1:

[1]

This is an extraordinary paper. Several years ago, Kandel was pushing the idea that Nitric Oxide mediated LTP through activation of a cGMP-dependent protein kinase (G-kinase). Nobody else who worked on this could reproduce his results. In an attempt to "stop the madness", four labs did something that is practically unheard of. The labs of Nicoll, Malenka, Malinow and Lisman agreed each to try to reproduce Kandel's experiments separately. They met ahead of time to make sure they all agreed on the details of Kandel's experimental protocol, and then they all went to their own labs and did the experiment without communicating with each other. After the experiments were completed, they got back together and compared results. They all four got the same thing. They could not reproduce the Kandel result. I know of this not just from reading the literature, but because they asked me to join in this effort as well. I declined only because I'd already published a failure to replicate this result.

Here is Kandel's original paper:

[2]

Example 2: ..An example not of failure to reproduce, but rather over-reaching interpretation on incomplete experiments.

This paper:

[3]

purports to show that Carbon Moxoxide is necessary for LTP (as a retrograde messenger). The problem is that they used only a single inhibitor of Heme Oxygenase (the enzyme that produced CO in the brain). This just does not conform to accepted practice. You need control experiments. We published a paper where we actually reproduced his finding, but the control experiments clearly indicated that Heme Oxygenase was not the relevant target of his inhibitor, rather those compounds that inhibited Nitric oxide synthase worked, while those that exclusively inhibited Heme oxygenase didn't. This wasn't the worst part. The worst part was that his lab then replicated my findings exactly, but (with a justification I to this day can't understand) interpreted it as supporting his previous conclusion rather than refuting it (as it does). Fortunately (I guess), he never published it. He just presented it at meetings.

Example 3: His lab has published numerous papers showing that cAMP analogues are sufficient to produce LTP. The usual suspects (see #1 above) have tried to reproduce this result and can't

Example 4: an example of failing to do the most basic control experiment. There is a lot of doubt in the LTP community about so-called "late-LTP". Not so much whether it exists or not (several labs have shown this), but rather whether it exists as a independently inducible entity, rather than just an natural follow-on to early LTP. He has based much of his work on his purported ability to induce the two phases separately (thus, being able to show the effects of experimental manipulation on one phase or the other). He induces E-LTP with a weak or brief tetanus, while E-LTP + L-LTP is induced by a strong or longer tetanus. The problem is that the ability to induce LTP has a failure rate. This failure rate goes down with stronger tetani. If you average the failures to get LTP in along with the successes (as he does), you will get what appears to longer-lasting LTP with the stronger tetanus. But what he is seeing is not LTP that lasts longer, but rather LTP that occurs more often. The test for this is simple. Simply average only the successes in inducing LTP. When you do that (as several labs have, you find that weak and strong tetani produce LTP of exactly the same duration (long). He is certainly aware of this critisism, but he can't do the proper analysis, because (IMHO) he's got too much invested in it.

Say Hi to David (I've met him).

Synaptidude 29 June 2005 00:30 (UTC)

Wikipedia????

Hi there, I'm new to Wikipedia and stumbled across this article just the other day. Are regular folks such as myself welcome to edit the main text of articles posted on the Wikipedia website? If that is the case, what is the safegaurd against fraudulent information being posted? As far as citing sources goes: How do we do it? Forgive me if these questions are answered somewhere else on the Wikipedia website, it's just that an open source encyclopedia database is such a revolutionary concept that I am a little taken aback.

Hey there! And welcome to Wikipedia. Yes yes yes!!!!! You are not only welcome, but encouraged to edit the text of articles. Go for it. Be bold The system is self-policing. If someone vandalizes, writes fraudulent stuff, nonesense, factually incorrect statements, bad grammar etc., someone else will edit it.

So have at it! Synaptidude 00:23, 15 July 2005 (UTC)

Oh yeah, if you want an example of this self-policing, go have a look at the talk page for creation science. We've been having a lot of fun over there basically arguing over the inclusion of a single word in the first paragraph ;-). Synaptidude 00:25, 15 July 2005 (UTC)

Early and Late Phase LTP

I have several questions regarding the deliniation of LTP into two distinct phases:

1) I understand that Early Phase LTP is the protein synthesis independent phase of LTP, meaning that the necessary components of synaptic plasticity are already present inside the cell. Are the various receptor subunits already assembled into functional receptors that are awaiting shipment to the dendritic branches, or do they have to undergo some level of processing before shipment?

LTP is expressed via the insertion of AMPA receptors into the extrasynaptic membrane and their subsequent diffusion into the postsynaptic membrane. This happens too fast for synthesis, assembly or transport to be involved. The receptors are preassembled in the ER, consitutively shipped to the dendritic spines where they are contained in recycling endosomes. Upon the induction of LTP, they are simply translocated into the membrane.

2) In reference to the aforementioned receptors/unassembled receptor subunits: Where are these units stored in the cell? What mechanism governs the release/transport of these units? I understand that CaMK II and PKA facilitate the insertion of these new receptors into the membrane. What I would like to know is how these receptors end up at the membrane in the first place?

This topic is still the subject of scientific investigation so the answers are not all known. But what appears to happen is that the receptors are moved from intracellular recycling endosomes into the plasma membrane via a vesicle-dependent exocytosis (kind of like transmitter release, but this time, the cargo is receptors, and they are in the membrane of the vesicle, rather than in the lumen.

3) Late Phase LTP is the protein synthesis dependent phase of LTP. After receiving the necessary signals, transcription occurs and is immeadiately followed by translation and thusly protein synthesis. Is it fair then to say that Late LTP serves mainly or perhaps solely to restock the neuron with the necessary components of plasticity? If so would it be fair to say that Late LTP is really just a homeostatic mechanism, keeping the cell stocked with the materials that it needs to performs it's duties? Do post synaptic changes occur as a direct result of L LTP or are the components of plasticity merely synthesized and left waiting for some later signal?

This is not really known, but it is believed that the protein synthesis is actually building new synaptic structure or new synapses. Some of us think that much of the work ::on L-LTP is BS.

4) How is it possible to activate E LTP but not L LTP? Doesn't the fact that L LTP cannot be induced without first inducing E LTP support the argument that L LTP is a rebuilding process and not a discrete type of LTP?

It is reported that E-LTP is induced by "weak" tetanus and L-LTP by strong. But IMHO it doesn't all add up. I personally believe that E and L are just different aspects of the same inseparable process. However, if you want the party line, read Kandel

Synaptidude 21:35, 18 July 2005 (UTC)

Thank you for your assistance with these questions

coincidence detector?

After further reading, I have some questions about the coincidence detector. How does the removal of the Mg++ ion from the NMDAR channel serve as a coincidence dector? If Ca++ is artificially inserted into the neuron will associative learning not take place unless the magnesium ion is also removed?

The more you can tell about this process the better.

Thank you for your assistance

coincidence detector

In terms of finding a neuronal substrate for learning, you need some sort of coincidence detector. Think Pavlov's dog. The dog learned that a bell meant food was going to be delivered because the bell was presented in coincidence with the food until the dog learned the association. So how could such coincidences be recognized by neural circuity?. One proposal is the the NMDA receptor serves as the coincidence detector. To understand this, you have to understand what TVP Bliss calls "The induction rule for LTP". The induction rule says that you need two things to happen simulataneously for LTP to be induced. Those two things are: 1) Binding of glutamate to the NMDA receptor, and 2) sufficient depolarization of the postsynaptic cell.

The reason that you need these two things to happen is because glutamate opens the gate on the NMDA receptor channel. But current can't flow through the channel because it is magnesium blocked. You also need sufficient depolarization to allow the magnesium to leave the channel. This allows other ions, in particular calcium, to flow through the NMDA channel. The reason that depolarization "works" is because the magnesium binding site inside the NMDAR channel is even with the plane of the plasma membrane and thus, is acted upon by the electric field of the membrane. When the membrane is depolarized, the electric field grows weaker and magnesium is not held as firmly in the channel, allowing it to escape.

The critical factor that induces LTP is the calcium that enters the cell through the NMDAR channel.

So the reason that this provides a coincidence detector is that it means that one synapse can help another synapse to potentiate. If you have a strong LTP inducing stimulus on one synaptic input, it not only meets the induction rule for that input, but it also depolarizes the whole postsynaptic cell. If a "weak" (i.e. not strong enough to induce LTP on it's own) input is activated at the same time, it releases glutamate onto it's own postsynaptic NMDARs. The channel opens, and IF BY COINCIDENCE THE CELL IS ALREADY DEPOLARIZED BY ANOTHER INPUT then the weak input will meet the induction rule, and hence will potentiate.

Back to Pavlov's dog. What if the infomation of "food" (known is classical pavlovian conditioning as the US (unconditioned stimulus)) is carried on the strong input, and the information on the bell (the CS or conditioned stimulus) is carried on the weak input. The bell means nothing to the dog, so the input remains weak. But if the food is consistantly paired with the bell, the strong and weak inputs get activated together, and the neural system is able to detect the coincidence between them because the stong input helps the weak input to potentiate.

See? Synaptidude 00:20, 21 July 2005 (UTC)

Quick question

Plasticity

I know that plasticity refers to the altering of connections between neurons, i just want to make sure that i understand all the different forms that plasticity can assume:

insertion of new receptors phosphorylation or movement of old receptors to increase efficacy increased dendritic spines and branching

did i miss anything??

Plasticity in the larger sense of biology means any change in a system that can be brought about by interaction of that system with it's environment. Synaptic plasticity refers specifically to changes that can occur at synapses in response to the environment, usually in response to synaptic electrical activity.

new receptors, old receptors, spine morphology etc, are aspects of plastic mechanism, but the term "forms" of plasticity, it is usually taken to mean a higher level. Forms of plasticity would be LTP, LTD, PPF, PTP, depotentiation, etc. The things you mention are mechanisms that might underly those forms of plastcity. For LTP, it is thought to be mainly movement of AMPARs into the membrane, while LTD is their removal. Phosphorylation it thought to be a trigger for the receptor trafficking. Spine and branching morphology changes are though perhaps to mediate longer slower forms of synaptic plasticity. Synaptidude 00:41, 15 August 2005 (UTC)

Cooperativity vs. associativity

Hi guys, great effort, great job! Just a remark I came across right now about the confusion of the terms cooperativity and associativity, both claimed to be different properties of LTP. Well, Bruce McNaughton, an electrophysiologist who contributed largely to the field and introduced the term "cooperativity" claims that to be the source of still ongoing confusion as for: there is no difference between cooperativity and associativity. The same occurs in your description of NMDA-receptor-dependent LTP and carries on with this. I think it's worth a fusion to stop muddling. Have a look at the references.

[4]

PMID 12740107

Bye, Tobi.

Associativity and cooperatively with regards to LTP are usually used to denote slightly different aspects of LTP, although both have the same underlying mechanism. Associativity is used to describe how one set of inputs can provide postsynaptic depolarization to another set of input to allow the induction of LTP in the second set. In other words, the association of a strong input and a weak one allows induction in the weak one. Cooperativity is usually taken to mean the number of axons within a given input that must be activated to get LTP at all. A few axons do not sufficiently depolarize the postsynaptic cell to induce LTP. It takes the simultanous activation (the coorperative activation) of many to get LTP.

In both cases, it is sufficient depolarization of the postsynaptic membrane potential to allow the unblocking of the NMDAR (by magnesium) that is the underlying mechanism.

So, they are the same thing, but usually taken to mean different aspects of the same thing. Synaptidude 00:36, 15 August 2005 (UTC)

Thanks

Just wanted synaptidude and all others who contribute to wikipedia for the awesome job you guys are doing explaining some fairly complex topics....thanks

ya'welcome! Synaptidude 18:35, 10 October 2005 (UTC)

More thanks! This article is incrediblely thorough + in-depth.

The role of CREB 1

I have read much of the role of CREB 1, that it is activated by the active subunits of PKA and that it goes on to regulate gene expression, my question is this: Have researchers pinpointed which genes are transcribed or what the end product proteins do inside the neuron? I understand that these genetic products relate to synaptic plasticity but exactly is produced?

The answer, for the most part, is 'no'. Experiments using gene arrays to capture and identify mRNAs produced, increased, or decreased after LTP have been failures. This is probably because brain tissue is so heterogeneous that the number of cells that have actually undergone LTP is very small compared to the number of cells in the tissue (e.g. Glia make mRNA, but are (probably) not involved in LTP.). There are lots of gene products that increase after strong activity (the early-immediate genes), but the role of these is unclear and they don't require LTP, only activity.Synaptidude 18:24, 10 October 2005 (UTC)

merge tag removed

I removed the merge tag (request to consider merging this article with LTP induction). I did this for the following reasons

• It has been there a long time, and there has been no discussion of the potential merge
• LTP induction is a stub that was written by me as a way of starting the process of moving some of the stuff in this article out to subordinate articles. This was done to try to get the main article under the 32K limit. I stoopped this process while waiting for discussion on the potential merging, but since there has been none, its, IMHO, time to think about getting the article under limit.

Definition of Associative LTP

The definition of associative LTP was dead wrong, so I corrected it. January 27, 2006.

For the record, I think you misinterpreted that section as defining "associativity" as it pertains to LTP. Sorry for the confusion. My original intent was to distinguish between the associative LTP that occurs at CA1 synapses and the non-associative LTP that occurs in the mossy fiber pathway. I no longer see that as an important distinction that needs to be on this page, so it's since been removed. --David Iberri (talk) 17:37, 9 April 2007 (UTC)

Don't Like First Sentence

The first sentence sounds like LTP means something to do with physical strength. I think it should be changed but can't think of what to. --Username132 (talk) 19:50, 21 May 2006 (UTC)

I agree, and unfortunately I can't think of a great alternative either. For now, I've changed "long-lasting strengthening of the connection" to "long-lasting enhancement of the synapse". Though now that I think about it more, I don't like this either. You don't enhance a synapse, you enhance the strength of a synapse. Gah. Let's think about this some more... --David Iberri (talk) 14:08, 31 May 2006 (UTC)

How about:

LTP is the long-lasting enhancement in the activity efficacy of a synapse between two neurons.

? --David Iberri (talk) 16:56, 2 June 2006 (UTC)

"Efficacy" to do what? Without using concepts not yet defined. DCDuring 03:29, 6 September 2007 (UTC)

Fair enough. What do you think of the current lead sentence?

In neuroscience, long-term potentiation (LTP) is an increase in the strength of a chemical synapse that lasts from minutes to several days.

--David Iberri (talk) 11:23, 6 September 2007 (UTC)

I agreed with the original objection to the word strength. Without qualification most readers who need a reference to LTP (like me) would probably think of some kind of mechanical concept of stength. DCDuring 11:29, 6 September 2007 (UTC)

I see your point, but I'm still having a difficult time relating the concept without too much verbiage. Do you have any suggestions? (Meanwhile, I've moved your comments about inconsistencies in the lead to the bottom of this page.) --David Iberri (talk) 11:52, 6 September 2007 (UTC)

I think this might strike the right chord. I'm going to replace the lead sentence with this:

In neuroscience, long-term potentiation (LTP) is a persistent increase in the chemical strength of a synapse that lasts from minutes to several days.

Comments welcome. --David Iberri (talk) 20:37, 7 September 2007 (UTC)

Having read some of the linked articles, I am no longer the perfectly naive reader although still very naive. I think I am looking to hook up my layman's model of the brain to what this article says. Is it the

• probability of or
• strength of

a signal being carried across a synapse that is associated with an increase in chemical strength? I think the educated layman believes that signals are transmitted through "nerves" and that that is part of how the brain works. The objective would be to correct or improve this kind of thinking without forcing the fellow to talk a course -- or even to read the whole article. DCDuring 21:31, 7 September 2007 (UTC)

I agree, but I'm not sure all the details belong in this article. For example, there are several ways to increase the strength of a connection between neurons; off the top of my head: 1) put more neurotransmitter vesicles in the presynaptic cell, 2) make more vesicles available in the readily releasable pool, 3) increase the probability of presynaptic vesicle release, 4) increase the number of neurotransmitter receptors on the postsynaptic cell, 5) increase the efficiency with which neurotransmitters trigger electronic or metabolic events in the postsynaptic cell, or 6) increase the number of synapses between the pre- and postsynaptic cell. These (and more) should be listed in our article on synaptic strength, but probably aren't (seeing as that page currently redirects to a section within the chemical synapse article). They probably don't belong on the LTP article because not all of these mechanisms are believed to be used during LTP expression.

As for relating a layman's model of the brain to a physiological one, that's also going to be difficult to accomplish in an article on LTP. It should probably be covered in synaptic transmission, with relevant details included in the articles on chemical synapses and action potentials. Do you think specific points deserve mentioning here though? If so, let's brainstorm. PS: I just finished a major revamp of the section on mechanisms, including some specific mention of ways to improve synaptic strength (eg, increase probability of vesicle release, increase postsynaptic sensitivity to receptor). If you have a spare moment, I'd love to hear your thoughts on it. Cheers, David Iberri (talk) 07:24, 8 September 2007 (UTC)

I think the first sentence of lede has to be very simple, with the lede as a whole suggesting, but not exemplifying the level of complexity of the full article. It also may pay to give the reader some idea why the article is important. If there are numerous candidate mechanisms that might be involved, that would not be the road to simplicity. The road to simplicity might lie in emphasizing the importance of LTP mechanisms for understanding the physiological basis for human memory. As an example of the kind of bridging that I am seeking: The word "Habits", which has a long heritage in both ordinary language and in psychology and pre-psychology was hijacked by behaviorists on WP who insisted that it was all just habituation. The habituation article had very little in it that provided anything of value from the point of view of anyone linking to the term habit. I am trying to create a bridging article that fits my views about readers. You don't need the bridging article, I don't think, but you do need to keep the article from being entirely irrelevant to lay readers. I will carefully examine the other article that you refer to. DCDuring 13:20, 8 September 2007 (UTC)

Great points, all of them. After editing the mechanisms section last night, I realized that the lead was quite a bit more distinct from the rest of the article than it should be. Like you suggest, the lead should give a better overview of LTP rather than carrying out a discussion of it that isn't elaborated upon within the main article text. If you have specific changes in mind, please be WP:BOLD and go for it! --David Iberri (talk) 15:52, 8 September 2007 (UTC)

There is a disconnect between the ordinary notion of memory as soemthing that can last years and lab-demonstrated LTP lasting only days. Explaining why lab LTP differs from naturally occurring LTP would help. Is it believed that the LTP has to be maintained over longer periods by repeated stimulation to give us long-lasting memories ? DCDuring 13:40, 8 September 2007 (UTC)

That question brings up a good point that needs to be explicitly said in the article: LTP decays with time. So yes, repeated stimulation is believed to help support robust LTP. Another consideration is that one memory doesn't equal one synapse undergoing LTP; what we commonly consider a memory is probably the cumulative effect of having LTP at many synapses (thousands, millions?). With time, repeated exposure, training, etc., the weights (ie, strengths) of these synapses get fine-tuned via changes in LTP. This helps explain the disconnect between LTP in the lab (where researchers typically study at most a few dozen or hundred synapses) and memory in living animals. For example, let's take a look at a few of those synapses whose strength was increased when you learned that new memory. Now let's say that that synapse represents a relatively insignificant piece of the memory you formed. Maybe you were taught what a pickup truck looks like, and your initial memory of it made a big deal out of the fact that the wheels were separated by almost exactly one meter. Of course, not all pickup trucks have wheels separated by exactly one meter, so that fact probably needs to get refined in your memory of a pickup truck. The synapses that represent that detail thus need to be weeded out of that memory. With repeated exposure to pickup trucks, and subsequent realization (unconscious or otherwise) that the wheels aren't separated by one meter, those synapses get pruned out. Either LTP decays in these synapses, or they undergo LTP's sister process, long-term depression which decreases synaptic strength. Either way, those unimportant synapses get pruned out of your memory of pickup trucks. I gather that much of these observations are difficult to reproduce in a laboratory setting for several reasons, not the least of which being that memories are so diffusely stored in the brain (eg, what we consider a single memory might be stored in synapses in every area of the cerebral cortex). This is becoming a bit long-winded, I apologize. But hopefully you can begin to see the difficulty in relating some of this in this article. Thanks for your feedback; it's been most helpful. --David Iberri (talk) 15:52, 8 September 2007 (UTC)

Am I right to include Timothy Bliss as a co-discoverer?

I googled that Tim Bliss was a co-discoverer of LTP, and included him in the sentence about the discovery in 1966, but now I have doubts. If I was wrong, -revert my change. --CopperKettle 04:36, 12 August 2006 (UTC)

Most researchers I've spoken with (which admittedly isn't many) seem to regard Bliss as a co-discoverer, but they base that on the 1973 Bliss & Lomo paper that first characterized LTP. But based on Lomo's recent paper on the discovery of LTP, I'm fairly certain that Lomo made the initial discovery by himself in 1966, with Bliss joining the Andersen lab only in 1968. I'll revert your change for now, but feel free to continue the discussion here. --David Iberri (talk) 14:03, 12 August 2006 (UTC)

Behavioral memory

This one's just screaming to be added to the "relationship to behavioral memory" section: Whitlock J, Heynen A, Shuler M, Bear M (2006). "Learning induces long-term potentiation in the hippocampus". Science. 313 (5790): 1093–7. PMID 16931756.{{cite journal}}: CS1 maint: multiple names: authors list (link) --David Iberri (talk) 04:28, 15 January 2007 (UTC)

Done, although it could use some refining. --David Iberri (talk) 19:26, 5 February 2007 (UTC)

Hippocampal anatomy

Jacobp22 (talk · contribs) recently added a brief description of hippocampal connectivity. It's good, but doesn't currently relate enough to LTP, so I've removed it for the time being and copied it here:

Located in the temporal lobe, the hippocampal formation is a specialized region of the limbic cortex. Believed by scholars to be associated with learning and memory, the hippocampal formation receives its primary input from the entorhinal cortex. The axons of these neurons then pass through the perforant pathway and go on to form synapses with granule cells located in the dentate gyrus.

I can see this being used in a section on LTP experimental design, which might describe why the hippocampus was used in early studies, why it is important to memory in general, and why it is important to LTP specifically. The answers to all these questions have to do with hippocampal function, which is so deeply tied to its anatomy that such a section would merit a description of the latter. Cheers, David Iberri (talk) 15:49, 10 April 2007 (UTC)

Comments

• The caption of the first image is too long, it is longer than the introduction!
• Reference 23 is missing.
• Adding more "Main article" links in subsections would help guide readers to other relevant articles.

Hope this helps. TimVickers 04:49, 11 April 2007 (UTC)

Definitely helpful; thanks. I've tightened the first image caption, moving the bulk of the text to Image:LTP exemplar.jpg. I've also fixed the oversight in ref 23 -- thanks for catching it. I agree about the main article links; I'm starting a Retrograde signaling in LTP section as a start (haven't submitted it yet, though). Any additional comments are much appreciated! Cheers, David Iberri (talk) 06:59, 11 April 2007 (UTC)

Duration

hwo long does LTP last?

This is covered in the introduction to the article. --David Iberri (talk) 18:59, 27 May 2007 (UTC)

Late Phase Clarification

Awesome article. Can someone craft the last paragraph in "Late Phase" to be a bit more pedestrian? I'm a bit lost. http://en.wikipedia.org/wiki/Long-term_potentiation#Late_phase —Preceding unsigned comment added by 76.22.5.47 (talk) 02:13, 2 September 2007 (UTC)

I took a look at that section and noticed that there's quite a bit on synaptic transmission mentioned there. But as the basics of transmission are covered in the background section, I'd imagine that's not too far out of the scope of this article. Are there specific points you'd like to be clarified? --David Iberri (talk) 03:45, 3 September 2007 (UTC)

Inconsistencies

Other comments: on the lede. There are a couple of inconsistencies that bother me. Sentence 1 says LTP lasts minutes or hours vs. Later in same paragraph LTP has to do with long-term memory Sentence 1 says minutes or hours vs. paragraph 2 says years. DCDuring 11:37, 6 September 2007 (UTC)

There's a distinction drawn between LTP observed in vitro (done in hippocampal slices, cell cultures, etc.), and LTP in vivo (observed in living animals). In vitro studies have only been able to demonstrate LTP that lasts up to several hours, while in vivo studies have observed LTP lasting for many months. --David Iberri (talk) 11:50, 6 September 2007 (UTC)

Lead rewrite

In light of the recent discussion about overcomplexity of the lead, the lack of continuity between it and the main article body, etc., I've rewritten the lead mostly from scratch. I wound up dropping the paragraph on in vitro and in vivo observations (which unfortunately was the only spot that mentioned a difference in duration between LTP in vitro and in vivo), but this will be added in elsewhere, and shortly. I took some care to be upfront about the mechanisms of enhancing synaptic strength, and what it means to improve the strength of a synapse. Hopefully this is a step in the right direction. Comments welcome. --David Iberri (talk) 02:00, 20 September 2007 (UTC)

PKMz and retrograde amnesia

I've got a question about the assertion that PKMζ inhibitors cause "Retrograde Amnesia" in the paragraph about Late-LTP. Shouldn't this be anterograde amnesia? Otherwise we should be talking about an inability to recall memories with the inhibitors. TDumbarton (talk) 20:55, 18 November 2007 (UTC)

We are talking about an inability to recall memories after adminitering the PKMz inhibitor. That's exactly what the section on L-LTP maintenance is trying to convey. The gist is that PKMz is necessary/sufficient for maintenance of L-LTP and the persistence of long-term memories. PKMz inhibition interferes with the persistence of memories resulting in loss of established long-term memory: retrograde amnesia. It's important to distinguish between the acquisition of memory and the persistence of memory. PKMz appears to have no role in the acquisition of memory, only its persistence. Inhibiting PKMz does not interrupt memory acquisition, so it does not cause anterograde amnesia. See also PMID 16931766. --David Iberri (talk) 16:18, 21 November 2007 (UTC)

GA Review

This article is very well written, well referenced, and meets all of the Good Article criteria. I'm almost at a loss as to make suggestions on how to improve it, though there are a few areas that could use some additional citations (e.g. the last 2.5 paragraphs of 'Synaptic tagging' -- "There is some evidence...", and 'Relationship to behavioral memory'). Other than that, this article is excellent. Good work! Dr. Cash (talk) 17:27, 9 December 2007 (UTC)

Sweet. Thanks so much for reviewing this, Dr. Cash. I'll see about adding some more references for the sections you mentioned. Thanks again, David Iberri (talk) 00:09, 19 December 2007 (UTC)

Recommended addition of synapse-specific LTF topic

Surprised this isn't mentioned or cited: Si, K. et al. (December 26, 2003). A neuronal isoform of CPEB Regulates local protein synthesis and stabilizes synapse-specific long-term facilitation in Aplysia. Cell, Vol. 115, 893-904.

Quick review: http://psychologyreview.blogspot.com/2008/02/neuronal-cpeb-stabilizes-synapse.html

--128.208.35.237 (talk) 21:52, 12 February 2008 (UTC)

Properties of LTP

I don't agree with the first property of Rapid induction. There are no references and the description sounds very much like Posttetanic Potentiation (PTP). PTP is distinctly different from LTP (please refer to Principles of Neural Science, Kandel and Jessel). PTP can last up to about an hour but is not LTP since though it is Ca2+ mediated, it does not involve phoshorylations or Kinases. As is my understanding, firing a high frequency volley through a neuron cannot IMMEDIATELY set up a LTP. It takes a couple of seconds for the cascades to work... at the bare minimum. PTP however is immediate, since it functions on residula Ca2+ from the last stimulation. User: Mubinchoudhury —Preceding unsigned comment added by Mubinchoudhury (talk • contribs) 11:19, 6 May 2008 (UTC)

Sorry... additional note. Furthermore Bliss and Collingridge, 1993 (don't remember the full reference)reported that short trains of high frequency stimulation caused synaptic potentiation. The key concept is that one train was not enough... so a brief tetanus cannot cause LTP. There are possible physiological reasons why more than one stimulation is required, that I shan't go into. User: Mubinchoudhury —Preceding unsigned comment added by Mubinchoudhury (talk • contribs) 11:26, 6 May 2008 (UTC)

Thanks for the input. By all means, be bold and correct the article (providing relevant references, of course) as you see necessary. Much thanks, David Iberri (talk) 23:32, 6 May 2008 (UTC)

Mechanism section poorly worded

I'm doing my PhD on LTP, and even I can't get my head around this section. The intro is supposed to be the easiest to read, but right away it lists 6 different stages of LTP. And it goes into painful detail on the chemical processes. First off you can't expect to describe both Schaffer collateral LTP and mossy fiber LTP in one section. it's too complicated. Secondly, The Early Phase: Induction sections jumps right into the calcium concentration. Perhaps some insight into the NMDA receptor as a coincidence detector should be given. A good way to start would be something like this "There are two types of receptors on the postsynaptic dendrite: AMPA receptors and NMDA receptors. Both are activated by neurotransmitters, but the NMDA receptor is blocked by a magnesium blockade. This can, however, be disloged if the dendrite is sufficiently depolarised (made more positive). The dendrite can be depolarised by this, that and the other thing. Once this, that, and/or the other thing removes the magnesium blockade, the neurotransmitters can can finally activate the NMDA receptor. This will allow a huge amount of calcium in. This calcium influx is what is principally involved in bringing about LTP". Something like that should be in the intro, so the reader has at least the foggiest idea of what is going on :-) Paskari (talk) 18:26, 1 July 2008 (UTC)

Property of NMDA receptors

I have never fully understood this because of the loose wording of most scientists. I can't tell if, once the magnesium blockade is dislodged (due to membrane depolarization because the AMPA receptors allowed sodium in), the NMDA receptor automatically lets calcium in (since it is voltage activated), or if once the magnesium is dislodged, glutamate must still bind to the NMDA receptor to allow it to open and let calcium in. I think it's the latter for two reasons. First of all, what is the purpose of making the NMDA receptor ligand activated, if the threshold which dislodges its magnesium blockade also activates it (is the NMDA receptor actually voltage activated, or is it loosely referred to as such since increases in voltage dislodge its magnesium blockade?). Secondly, for the purposes of LTP, it would conflict with input specificity, since neighbouring regions might depolarize the spine and remove its blockade. If calcium immediately flowed in, then that spine would undergo LTP, even if it received no neurotransmitters. This line from the section on induction of E-LTP is equally as vague "Sufficient depolarization through the summation of EPSPs relieves the magnesium blockade of the NMDAR, allowing calcium influx." Paskari (talk) 18:36, 1 July 2008 (UTC)

I found an answer to my question, and I'm going to go ahead and include it. As it turns out, glutamate must still bind to the NMDA receptor, even after the magnesium has been dislodged. Paskari (talk) 15:29, 6 July 2008 (UTC)

I updated the section, but I didn't cite my sources (because it seems like a headache to do). I am trying to learn how to do it, please don't delete anything yet. Also there is some confusing terminology I used, so for example, in one sentence I said "...AMPA receptors and NMDARs", I personally don't like the terminology NMDARs, but I definitely wasn't going to go back and change all the "AMPA receptors" to "AMPARs", or "NMDARs" to "NMDA receptors". I'll leave that to someone else ;-) Paskari (talk) 15:56, 6 July 2008 (UTC)

Permeability of NMDA Receptor

Another issue scientists are notorious on is downplaying the permeability of AMPA receptors towards calcium, but making a big deal out of NMDA permeability to calcium. For example, on page 130 of Neuroscience (Purves et. al. 2008) it reads all ionotropic glutamate receptors [NMDA, AMPA, and Kainate] are non selective cation channels...allowing the passage of Na and K, and in some cases small amounts of Ca. Then two sentances later it says NMDA receptors have especially interesting properties...[their] ion channels allow the entry of Ca in addition to monovalent cations such as Na and K. It seems as if it's reiterating the previous statement at best, and down playing the role of AMPA and Kainate receptors at worst. Paskari (talk) 18:49, 2 July 2008 (UTC)

Turns out, the AMPA receptors are parmeable to calcium, but their contributions to the calcium concentration are negligable. Paskari (talk) 15:44, 6 July 2008 (UTC)

Silent Synapses

Although silent synapses are not included in this section, they prove my point that NMDA receptors are never properly described. Basically silent synapses are post synaptic spines covered in both NMDA receptors and AMPA receptors, however the AMPA receptors are inactive. Therefore, only the NMDA receptors can let ions in. But NMDA receptors are blocked at rest (by magnesium), so they can't allow ions in, until the AMPA receptors let ions in so as to depolarize the membrane potential (and dislodge the magnesium). See the Catch 22? The only other scenario is if the surrounding synapses contribute to depolarizing the said synapse. This is what I really want to see addressed in this article, an air tight explanation of how LTP works. Thanks Paskari (talk) 14:29, 3 July 2008 (UTC)

I agree that silent synapses may deserve passing mention here, but why do you think they must be invoked to explain LTP? Also, there is no self-respecting, legitimate neuroscientist who could claim to have an "air tight" explanation for LTP's mechanism. --David Iberri (talk) 11:45, 8 July 2008 (UTC)

Yes I suppose any mention of them should be in passing. Also instead of 'air tight' I should have said something to the effect of 'more accurate'. Paskari (talk) 10:40, 9 July 2008 (UTC)

in vivo vs. in vitro

As far as I'm aware LTP has been shown to work both inside an animal and inside a lab, but has it ever been proven to be naturally occuring, or is an artifical tettanus always applied? Basically is LTP a mechanism synapses are capable of exploiting or has it been proven that synapses do in fact exploit LTP in vivo? Paskari (talk) 17:05, 6 July 2008 (UTC)

This is mentioned in the article, specifically under the Relationship to behavioral memory section. --David Iberri (talk) 11:46, 8 July 2008 (UTC)

I'm having some major problems with that section as well. The first paragraph starts off with "In 1986, Richard Morris provided some of the first evidence that LTP was indeed required for the formation of memories in vivo." However, the last sentence in the same paragraph suggests the evidence is not that compelling, "This provided some early evidence that the NMDA receptor — and by extension , LTP — was somehow involved with at least some types of learning and memory." The next paragraph has a similar implication, "providing more support to the notion that LTP is the underlying mechanism of spatial learning." The next paragraph doesn't mention if it's in vivo or in vitro, let alone whether it's naturally occuring, or artificially induced. The last section, offers the closest evidence to concluding that LTP occurs naturally in vivo, it reads "...substantially advance the case for LTP as a neural mechanism for memory." I think it should be stated outright if this occurs naturally or is merely a scientific experiment (even if there is overwhelming evidence to suggest that it might occur naturally). To put it in perspective, facilitation, depression, augmentation and potentiation all do occur naturally. Paskari (talk) 16:32, 8 July 2008 (UTC)

I agree, and have made the following changes: [5]. --David Iberri (talk) 14:54, 3 October 2008 (UTC)

Last edit

Apologies for the rather harsh edit summary I left on my last edit. I apparently haven't learned to avoid the stray Wikipedia edit on sleep deprived nights. --David Iberri (talk) 11:17, 7 July 2008 (UTC)

I don't mind reverting my version, but now it's even more vague than it was before. Let's discuss how we can improve this article, as it truly is one big mess.

• In the section "Types" it should have subsections such as "LTP in the Hippocampus", "LTP in the Aplysia", "LTP in the Amygdala", "LTP in the Visual Cortex"... and each one should have a very brief overview of them. It's, in my oppinion, OK to discuss the steps, just don't scare the casual reader away with too much detail.
  • I think you're right: the level of detail in the "types" section is likely overwhelming to the casual reader. However, I think subsectioning things to this extent has the tendency to detract from the article's readability. Let's compromise and simplify that section, shall we? --David Iberri (talk) 11:39, 9 July 2008 (UTC)
• The "Properties" section is good in that it is brief, but it should be linked with the later explanations. So, for example, it should be explained that input specificity arises from the narrow spine 'neck' which enables the bulbous compartment to act independently by preventing CamKII from spreading to neighbouring dendrites. Furhtermore, it should be explained that NMDAR voltage activated nature leads to associativity and cooperativity. Either this section has to be moved below the 'Mechanism' section, or have one of those '(see here)' anchors at the end of each point.
  • You've brought up the NMDA receptor and its ability to act as coincidence detector several times. In my initial draft of this article several years ago, I went into sufficiently more detail about the NMDA receptor and its very well established properties of magnesium blockade and voltage dependence, and its dependence on binding of glutamate and glycine. Some of this is included in the E-LTP induction section, but perhaps it can be done in a more clear way. I'm not so sure this information needs to be duplicated in the properties section, but I do agree that more should be said about the precise mechanisms for LTP's associativity and cooperativity (really two sides of the same coin). Are you convinced these are NMDA receptor-dependent phenomena? I'm not familiar enough with the extra-CA1 LTP literature to know whether NMDA receptor-independent LTP is capable of associativity and cooperativity, but from what I understand, these are nearly universal truths about LTP regardless of site of expression. If you can shed more light on this very interesting topic, it'd be much appreciated. --David Iberri (talk) 11:39, 9 July 2008 (UTC)
• The 'Mechanism' section should include subsections such as 'Mechanism for CA1', 'Mechanism for DG', 'Mechanism for Aplysia'...In fact it wouldn't be a bad idea to create new articles for all of them.
  • I mostly agree, but had (and still have) three major reservations: 1) I am only vaguely familiar with the mechanisms for LTP outside the CA1 hippocampus; 2) I don't have a list of all the various major types of LTP, so enumerating a list would be artificial and necessarily incomplete; and 3) describing mechanisms is rather complex (ie, lots of text) and shouldn't clutter the main article. But again I mostly agree that we at some point need to have individual articles with all the different descriptions and mechanisms of LTP, linked to from the main LTP article in the way you mentioned. I just don't feel comfortable doing it myself for the above reasons. :-) --David Iberri (talk) 11:39, 9 July 2008 (UTC)
• Ideas seem to be all over the place. I find the best way for me to understand, is to break it down to the monkey stupid. So specifically state that AMPA doesn't allow in calcium (in the CA1 anyway due to the GluR2 sub unit) and that the NMDA receptor does allow Calcium in. Simply saying that, for example, all ionotropic glutamate receptors [NMDA, AMPA, and Kainate] are non selective cation channels...allowing the passage of Na and K, and in some cases small amounts of Ca...NMDA receptors have especially interesting properties...[their] ion channels allow the entry of Ca in addition to monovalent cations such as Na and K. (NEUROSCIENCE, Purves et. al. 2008, p130) and hoping that the reader will understand what in some cases means, is not enough. Also it has to be specifically stated that the NMDA receptor is not voltage activated like the sodium channels. One could depolarise the cell membrane all he or she wants, but that wont open the NMDA channel, it will only dislodge the magnesium. The glutamate must still bind.
• Here is a sentence from the second paragraph of the 'Mechanism' section "In addition to these mediators, there are also modulator molecules, described later, that interact with mediators to finely alter the LTP ultimately generated." That alone should be enough to scare any casual reader into not continuing any furhter.

Lets outline what needs to be changed, then assign sections for different people to update Paskari (talk) 13:28, 7 July 2008 (UTC)

I don't know what the protocol is when the Talk sub sections get nested so recursively, so I'll just post my comments down here

• I agree that having subsections for different types of LTP (LTP in the hippocampus...) might be too complicating, but then how do we approach this? After all, it is misleading to make people believe that CA1 LTP is the only type of LTP. But, alas, this can't be done over night, and we need people who are familiar with the different forms of LTP. I have a text book in front of my, but I am busy for the next week, so I can only glance at LTP in the Aplysia right now. I promise to devote a section to it however.
• I know that the mossy fibers display LTP regardless of what goes on postsynaptically (can never remember if this is homosynaptic or heterosynaptic LTP). So in a sense that completely rules out associativity and cooperativity. Furthermore, LTP in the Aplysia and Cerebellum use somewhat different steps.
• When I mentioned NMDA receptors, I didn't mean to say that I wanted more info on them, only that the info not be misleeding. Even calling the NMDA receptor "votage activated" isn't entirely accurate since it is merely 'voltage regulated'. Paskari (talk) 21:25, 11 July 2008 (UTC)

You've made a couple very important points.

• You're right that discussing all the various types of LTP would be too cumbersome for this article. I tried to think that through in my edits, coming up with what I considered a reasonable solution: introduce the idea that there are multiple forms of LTP, what factors determine the type of LTP exhibited between neurons, and focus the article on the prototypical LTP (NMDAR-dependent LTP in the adult CA1 hippocampus). There are other solutions, of course, but this was the most obvious and sensical to me. I'm not opposed to having better descriptions for the various types of LTP (cerebellar, CA1 hippocampal, CA3 hippocampal, Aplysia, etc.), but as I've said, my lack of experience with them prevents me from giving them fair representation in such a section.
• I may be misreading something about your thoughts on how the NMDAR is covered here. Your comments above made it sound like you felt the article didn't sufficiently address the NMDAR's importance in cooperativity and associativity, which is why I suggested there might be something missing from the article. I think the only time it was mentioned that the NMDAR is "voltage activated" was in your comment above. Is there something I'm missing?

--David Iberri (talk) 02:15, 12 July 2008 (UTC)

Yes you're right, it doesn't appear in this article. I was referring more to the fact that several text books make the mistake of calling NMDA receptors voltage activated instead of voltage regulated. I'm very busy until the 17th, I'll try to read up on LTP after that. Cheers. Paskari (talk) 12:43, 13 July 2008 (UTC)

Fog Index

I just ran the introductory paragraphs into a Fog Index Calculator [6] and I got a Fog index of 21. Even if the calculator is inaccurate and we get something like a Fog Index of 15, that's still slightly higher than the Fog Index of 12 that we need to target the general audience. I have no problem going through the whole document and breaking down sentences which contain more than one idea, for example, here is how I could rephrase the introductory paragraphs (there's an easter egg in there for you, so enjoy)

Long-term potentiation (LTP) is the long-term enhancement in communication between two neurons. This can arise when both neurons are active at the same time. LTP can also arise if only one of the neurons is active. Nevertheless, this activity leads to the strengthening of one or both of their synapses. Therefore, it is believed that communication between these two neurons is carried out via these synapses. Furthermore, it is believed that these same synapses are responsible for storing memories . Therefore, LTP and its opposing process, long-term depression, are widely considered to underlie learning and memory.

LTP improves the ability of a presynaptic neuron to communicate with a postsynaptic neuron. It carries this out by enhancing either of their synapses. Presynaptically, LTP can enhance the axon terminal buttons. Postsynaptically, LTP can enhance the dendritic spines. The precise mechanisms for this enhancement have not been fully established. This is in part because LTP is governed by multiple mechanisms. Furthermore, these mechanisms vary by such things as brain region, animal age, and species.

Of these, the most well understood form of LTP occurs in the hip bone of the sperm whale. Here, LTP is brought on when the presynaptic neuron rapidly fires bursts of action potentials. Upon receiving these bursts, the postsynaptic neuron implements LTP. Therefore, communication, in this case, is enhanced postsynaptically. The postsynaptic neuron begins work to make its spine more sensitive to the presynaptic neuron. It does this by enhancing its receptors. This enhancement, is carried out by inserting new receptors, or making existing receptors more efficient. These receptors now act to receive even more neurotransmitters from the presynaptic neuron. This, therefore, constitutes an enhancement in communication.

LTP shares many features with long-term memory. This makes it an attractive candidate as a cellular mechanism of learning. For example, LTP and long-term memory are rapidly induced. Each depends upon the synthesis of new proteins. Each has properties of associativity. Finally, each can potentially last for many months. LTP may thus account for many types of learning. This learning can be the relatively simple classical conditioning present in all animals. However, it can range to the more complex, higher-level cognition observed in humans.

LTP was discovered in the rabbit hippocampus by Terje Lømo in 1966. It has still remained a popular subject of research ever since. Most modern LTP studies seek to better understand its basic biology. Some studies aim to draw a causal link between LTP and behavioral learning. Still others try to develop methods, pharmacologic or otherwise, of enhancing LTP. These methods are usually meant to improve learning and memory.

There are, nonetheless, problems with this intro. Like why mention associativity, when no one knows what it is, and there is no real description given. Or whether it leads the reader to think of the synapse as a piece of gum connecting two neurons, when it is a pair of separate growths. Paskari (talk) 16:33, 2 August 2008 (UTC)

I used a more accurate FOG index calculator [7] and it bumped the number down to 18.77, still too high. It's not so much the index that bothers me (large words are unavoidable in neuroscience), it's the average words per sentence, which was around 25 words per sentence. Paskari (talk) 18:58, 2 August 2008 (UTC)

Does no one have any opinion on this whatsoever? Paskari (talk) 07:30, 19 August 2008 (UTC)

I'm not familiar with fog indexes or their reliability, and I've never heard of Wikipedians using them to assess an article's readability. I'm much more interested in readers' subjective experience reading the article. Does it go over their heads? Are the sentences too long and unweildy? Is there too much jargon?

In writing this article, I've targeted the fairly informed technical reader and have tried to churn out some decent prose. IMHO, I think the result is satisfactory, which other reviewers have generally agreed with. But I think we agree that the problem is that this article's readership needs to include more than just technically informed readers.

While targeting a particular fog index might be tempting, I think blindly doing so would detract from the pleasure of reading any Wikipedia article. In achieving greater readability, we need to do more than just swap words with synonyms, break long sentences into separate clauses, etc. Sure, doing these things might make the article readable by high schoolers, and might improve our fog index, but it ruins whatever semblance of elegant prose the article had.

A great article needs to not only be accessible, but also be pleasant to read. I'm totally with you that we need to achieve greater accessibility, but let's try to make the article as pleasant as possible to read too. --David Iberri (talk) 14:39, 24 August 2008 (UTC)

Addiction

Here's a nice article discussing LTP in the context of addiction:

• Kauer JA, Malenka RC (2007). "Synaptic plasticity and addiction". Nature reviews. Neuroscience. 8 (11): 844–58. doi:10.1038/nrn2234. PMID 17948030. {{cite journal}}: Unknown parameter |month= ignored (help)

--David Iberri (talk) 12:37, 7 November 2008 (UTC)

small error?

Resolved

I'm pretty sure that Lomo was not recording from individual neurons, as the article now says, but rather was looking at the "population spike" in the field potential following perforant path stimulation. Since I'm frequently wrong, I thought I'd better bring this up here before attempting a change in the article. Looie496 (talk) 22:28, 7 April 2009 (UTC)

Very good point. I'm embarrassed to say that thought didn't even come to me when I made the most recent edits. I'll have to take a look at the original source to confirm. Thanks for pointing this out. --David Iberri (talk) 12:06, 9 April 2009 (UTC)

Looie, your suspicions are confirmed in

Terje Lømo (2003). "The discovery of long-term potentiation". Philos Trans R Soc Lond B Biol Sci. 358 (1432): 617–20. doi:10.1098/rstb.2002.1226. PMID 12740104.

I've updated the article accordingly: [8]. --David Iberri (talk) 22:44, 10 April 2009 (UTC)

(ab)use by advertisers?

Is this why rapid flashes of light appear in TV ads?

This phenomenon, whereby a high-frequency stimulus could produce a long-lived enhancement in the postsynaptic cells' response to subsequent single-pulse stimuli, was initially called "long-lasting potentiation"

-75.145.87.233 (talk) 02:45, 29 July 2009 (UTC)

See Subliminal message for a possible correlation with your idea. --Ancheta Wis (talk) 03:04, 29 July 2009 (UTC)

As far as I know, that has nothing to do with LTP, though. Looie496 (talk) 15:30, 29 July 2009 (UTC)

Subliminal is a one-shot image. Rapid flashes seems to resemble "high-frequency stimulus". Why would that have "nothing to do with LTP"? - 75.145.87.233 (talk) 04:43, 30 July 2009 (UTC)

The reason for using a high-frequency train of pulses in LTP is that a single pulse, however strong, would only activate synapses one time. To get LTP, it's necessary to have multiple activations within a short time window. A continuous visual stimulus produces as much activation in most brain areas as a flashing stimulus, so flashing wouldn't be associated with greater LTP except in change-sensitive parts of the brain. In short, the resemblance of the two situations is only superficial. Looie496 (talk) 15:37, 31 July 2009 (UTC)

I don't perceive the distinction. Rapid flashes of light would seem to qualify as "multiple activations within a short time window". What is "a continuous visual stimulus"? Must it be moving? A painting may be a continuous visual stimulus. It doesn't move, yet it may provoke an emotional response. But in general, paintings don't have such an effect. PBS often shows old photos by zooming and panning them while making them flicker. Why do that? What are the "change-sensitive parts of the brain"? Wouldn't advertisers want to target them? Isn't that the point of advertising, to change behavior; as permanently as possible? Wouldn't LTP be a helpful tool for advertisers? Might they not achieve it with rapid flashed of light? - 75.145.87.233 (talk) 00:47, 1 August 2009 (UTC)

LTP is only "helpful for advertisers" insofar as it appears to underlie learning, memory, and other cognitive domains. It probably gives rise to all sorts of phenomena, including associative learning, habituation, classical conditioning, etc. One can certainly capitalize on these traits to change human behavior, but this is not done by directly accessing LTP. This may be difficult to grasp, and understandably so. The crucial bit is that LTP is a phenomenon observed at a cellular and synaptic level, not a behavioral one. --David Iberri (talk) 23:51, 2 August 2009 (UTC)

Hebbian vs. anti-Hebbian

I am not happy about the use of "nonHebbian". I would prefer to use the term "Hebbian" for any learning rule that alters strengths as a function of both pre-and post-synaptic activity. Thus your "antiHebbian" would still be Hebbian. One would then use "nonHebbian" for cases where learning depended on just the presynaptic, or just the postsynaptic, signal. ab is the product of a and b, but -ab is not the "antiproduct" of a and b - it's still a product, albeit of a and -b Possibly one could say that the classical Hebbian rule is pre before post strengthens, and if it weakens instead it is still Hebbian, but in a more generalised sense. 129.49.104.32 (talk) 20:43, 25 November 2009 (UTC)

I would also like to comment on the statement above: "it should be explained that input specificity arises from the narrow spine 'neck' which enables the bulbous compartment to act independently by preventing CamKII from spreading to neighbouring dendrites". I thin it would be better to drop the specific reference to CaMKII : the narrow neck does ensure high specificity, but by reducing the spread of ANY second messengers, not just CaMKII. In fact, my guess is that preventing the spread dof Ca is far more important than preventing the spread of CaMKII (which does not mean this does not spread.

More generally, it must be a real headache for the authors of this vital article to be simultaneously clear, succinct and correct, and on the whole they do an admirable job. Different experts will differ on which aspect is really important, and only time will tell. But insofar as details may not be highly relevent, they should be eliminated. 129.49.104.32 (talk) 20:54, 25 November 2009 (UTC)

Regarding your first point, the usage of anti-Hebbian for synapses that weaken is universal in the literature as far as I know, so I don't see how we can depart from it. Regarding your second point, I think that you should feel free to fix the problems you see in the article, if you think you can. Looie496 (talk) 22:49, 25 November 2009 (UTC)

Spike timing-dependent plasticity?

Is it possible for anyone to add the relationship between STDP and LTP to the article? I'm no expert (yet!) but I'd try to help out.. thanks --The.Filsouf (talk) 23:47, 1 January 2010 (UTC)

Contradictory information about PKMz

The article presents contradictory information about the role of protein kinase Mz (PKMz). It states:

1. PKMz underlies the maintenance of E-LTP, even though there is no mention of PKMz being activated in the induction phase of E-LTP
2. PKMz underlies the maintenance of L-LTP and is not required for E-LTP.

It seems most likely that PKMz is active in L-LTP and not in E-LTP, especially since the cited article makes no mention of PKMz. If that is the case, what is the mechanism for maintenance of E-LTP?

140.180.35.44 (talk) 16:37, 10 October 2010 (UTC)

According to this review, you are right that it is involved in late but not early LTP. I'm not sure that early LTP needs to be maintained, as by definition it decays if not consolidated into late LTP. Looie496 (talk) 17:25, 10 October 2010 (UTC)

Inclusion of Ramon y Cajal

Is the inclusion of Cajal necessary in this article? There are several inaccuracies in the characterization of his contribution. Notably he did not predict or theorize about information flow or memory storage at synapses, indeed the term synapse had not even been coined at the time of his Croonian lecture. Cajal did muse about the elaboration of the dendritic tree with time. This should be taken in context of Forel's and Cajal's work in developing animal where connections are comparatively sparse relative to the adult. Cajal was not speculating about learning in the adult animal, he was commenting on the increasing complexity of dendritic trees during development. Cajal also never recognized the spines as sites of contact between neurons. —Preceding unsigned comment added by 99.240.130.52 (talk) 04:25, 17 October 2010 (UTC)

Please feel free to make improvements in the article, which is not being very actively maintained. Looie496 (talk) 05:24, 17 October 2010 (UTC)