Talk:Place cell

From Wikipedia, the free encyclopedia

With respect to theta phase precession, the relationship between firing rate and theta phase is highly disputed. More research is required before providing a clear answer. Therefore, the uncoupling of firing rate versus theta phase was altered to include research articles that conflict with the findings of Huxter et al.

USER BIAS:

Again, PLEASE do not use this web-site for self-promotion. Users should provide an unbiased report of the field as it stands today. Remember, this is an encyclopedia. Not a blog. —Preceding unsigned comment added by ThetaMonkey (talk • contribs)

Contents

[edit] Article needs expansion

(A wiki error plus an internet explorer problem just ate my long post here, so I must re-write it and will necessarily make it briefer than before. Drat!)

I think the current article focuses too much attention on hippocampal place cells. I propose the following:

1. Moving the phase precession section to a new article because many neurons show phase precession and some do not have spatial correlates (e.g. medial prefrontal cortex, Hyman et al. 2005, Jones and Wilson 2005).

2. Including discussion of spatial firing in other brain regions. Place cells are not as simple as "principal neurons found in the hippocampus that fire strongly when an animal, typically a rat, is in a specific location in an environment". There is much evidence that shows that there are non-spatial correlates of hippocampal place cell firing as well(Prospective/retrospective effects: Wood et al. 2000, Ferbineanu and Shapiro 2003, Bower et al. 2005. Non-trivial response to environmental changes: Lee et al. 2004, Leutgeb et al. 2005, Wills et al. 2005). More importantly, neurons with spatial firing correlates occur in other brain regions (dentate gyrus, subiculum, parasubiculum, entorhinal cortex, postrhinal cortex, lateral septum... refs available on request) and occur with their own set of interesting properties (grid-like firing fields, head direction correlates, context dependency, path-equivalence, etc).

If no one objects, I will go ahead and make these changes in a few days (keeping as much as the current content and wording as possible while doing so). Digfarenough 22:53, 13 July 2006 (UTC)

Just a thought... is it fair to say that "many neurons show phase precession and some do not have spatial correlates". Hyman et al. 2005 never showed precession (or atleast failed to mention the word 'precession' in their manuscript) and Jones and Wilson, 2005 do not show the best of examples. I am guilty of not reading closely and would appreciate some insight- how many of Jones and Wilson's prefrontal neurons precessed and how many did not? Could they achieved these results by chance? Did any neurons PRO-cess? To my knowledge, virtually all principal neurons in the hippocampus exhibit theta phase precession. So I guess what I am getting at is that it appears that precession is the "rule" in the hippocampus rather than the "exception" seen in the PFC. ThetaMonkey 25 July 2006
I "deboxed" your question. You're right, I forgot that we didn't include any precession stuff in the Hyman et al. paper, but we did look for it and found some hints of it, but not enough to include. You're also right about Jones and Wilson not having the best of precession, but they do sort of show it. The only reference to number of cells on the Jones and Wilson paper that I can find is n=39 mPFC cells that showed phase shifts over space. I can't seem to find where they say the total number of cells recorded, though. Basically I agree with everything you say and I may have been a bit overzealous in making the claim of precessing cells without spatial correlates (I should at least have said clear spatial correlates, because some mPFC do show spatial firing to some degree). However, not all principal cells in the hippocampus show phase precession (look at the big two page figure in Skaggs et al. 1996 and you'll see a fair number that only fire at one particular phase and don't precess). It's also true that the Moser lab has reported that ECII cells but not ECIII cells show phase precession (two abstracts from SfN 2005, which they reference in their recent paper in Nature Reviews Neuroscience--I've only seen the abstracts though, not the posters so I don't know how good the data looks). My personal view (based on a new model I've been working on) is that precession is generated in ECII and HC just inherits it, but that is clearly speculative so I wouldn't actually add it to an article. So I basically agree with your last sentence there, but with the above slight caveats. Digfarenough 17:39, 25 July 2006 (UTC)

[edit] references and jargon

The Place cell article needs to improve citation of sources; it needs to use a standard citation format linked to a reference list. Also, care needs to be taken to explain jargon. --JWSchmidt 16:05, 18 July 2006 (UTC)

Please define: "a late phase" --JWSchmidt 16:52, 18 July 2006 (UTC)
Taken care of. I'll take a look at the references and see where else I can add some quick citations, if anywhere. Digfarenough 17:31, 18 July 2006 (UTC)

[edit] produces vs. appears to produce

Jhuxter, you seem to disagree with the wording I used on an earlier edit, changing my wording of "the place cell appears to produce a rhythmic discharge of a slightly higher frequency..." to "the place cell produces...". I made that change because saying that it produces a rhythmic discharge makes it sound like phase precession is caused by the higher frequency spiking instead of the higher frequency spiking merely reflecting the precession. My interpretation of Zugaro et al. (2005) and Dragoi and Buzsaki (2006) is that precession isn't caused by an intrinsic mechanism in the hippocampus but rather by external input, which is why I prefer a wording that suggests that hippocampal place cells appear to produce higher-than-theta frequency spiking. Probably the simplest way to take care of this is to include a paragraph or two on putative mechanisms for the generation of phase precession, but I tend to lean away from including models from neuroscience in general. In summary: while it is true that mechanisms of phase precession were not being discussed there, I think it best to avoid wording that might lead to conclusions that are not supported by all the data, even if said data isn't actually discussed in the article. Digfarenough 18:38, 18 July 2006 (UTC)

[edit] produces vs. appears to produce - response

Saying that the cells fire at a higher rhythmic frequency than the ongoing theta field oscillation is a simple statement of fact - it is not a comment on mechanism, but a necessary condition for phase precession to occur, and there's no avoiding it. This particular wording gives an alternative way of viewing the phenomenon, and to clarify for people who are less familiar with the concept of circular-linear correlations. That's why I considered your edit a bit misleading - nobody would dispute what the place cells are doing for it to be seen as phase precession - it's not controversial. I think the current wording of the last paragraph now makes it clear to the reader that the mechanism of phase precession is still under debate. Of course, we have not yet given adequate credit to any of the other models of precession - only the depolarisation model. If you would like to correct that particular imbalance...

J. Huxter 19/07, 2006

[edit] Uncoupling phase and rate

I am content with the current wording of the section, although it should be pointed out that the actual data presented by Harris et al. and Mehta et al (2002) are not in conflict with Huxter et al. 2003 - only the interpretation of the 2002 results. For example, the latter replicated the findings of Harris and colleagues with regards to instantaneous firing rate, and simply expanded on them to demonstrate that dissociation is possible both with and without experimental manipulation. That is why I edited the section referring to Huxter et al. as "a conflicting report". To date, there is no published data refuting the dissociation of firing rate and firing phase. That said, I'm perfectly happy to leave the mechanism of phase precession described as "unresolved".

J. Huxter, 19/07, 2006

[edit] Future Additions

I'd like to add a list of things that should be added to the article- if you feel that something is missing, go ahead and add it. I'll try to take a shot at writing a paragraph for these topics in the next few weeks, but if you would like to go ahead and do it (which would be greatly appreciated), feel free!

  • Center of Mass Shift/ Place Field Assymetry
  • CA3 versus CA1 (similarities and differences)
  • A short discussion on the subiculum (as well as pre- & para-subiculum)?
  • Prospective / retrospective coding
  • Different types of Remapping (rate versus position) —Preceding unsigned comment added by ThetaMonkey (talk • contribs)
Related to remapping, Gothard et al. (1996)'s reference frame remapping (the linear track example is pretty easily explainable) showed some interesting results. One thing that I think is often underemphasized is that while place cells have stable place fields, the cells appear to be somewhat stochastic in that it is not uncommon for a cell to hardly fire at all on some passes through the field. I think perhaps the beginning of the article should be slightly changed to reflect that (as it is it sounds like a place cell always fires strongly whenever the animal is in the cell's place field). I'll be out of town for a week, which might actually give me a chance to make a few of these changes to the article. digfarenough (talk) 21:37, 7 August 2006 (UTC)

[edit] Be cautious

Concerning the suggested changes under "Article Needs Expansion":

  • 1. moving phase procession. I agree.
  • 2. Including discussion of spatial firing in other brain regions. Be cautious. Over the years I've seen far too many references to spatial correlates as spatial signals. A correlation is not a code. I suspect that if you recorded in any region of the brain you'd get a weak but reliable spatial correlate. I don't recall ever hearing a formal definition of a "place cell", but, at a minimum, the cell must have a spatial correlate that is not secondary to other correlates. Since it is extremely difficult to show that a correlate is not secondary to other correlates (e.g., spatial view, locomotion, theta, arousal, fear, brightness) weak correlates should be rejected as candidates. Please don't call all cells with spatial firing correlates "place cells". I agree that the assertion that place cells are pyramidal cells is wrong. I can't recall anyone has ever saying that in the way this article does.
  • I'm trying to avoid the "self promotion" directive above, but I think the citations do a disservice to the place cell work that comes from Brooklyn. There is not a single reference to Muller, Ranck, Kubie (me), Fox, Fenton, Quirk, etc. Many of the assertions in the article are solid, and, I feel, should use citations from the Brooklyn group. Moreover, many of the citations are weak. I don't have time right now to work on citations. I'll try. But I'd like to see them edited.
  • Another comment. Overall, I don't think this article would help someone with no background. The choice of figures is poor. There should be at least one firing rate map. The phase procession figure, while pretty, is very hard to figure out. The Golgi image of a hippocampal pyramid, while pretty, doesn't do much. The complexities of extra-cellular recording (where you can't see what you are recording! ) are not clear.
  • The article is extremely weak on theory. In the first few sentences there is reference to the "cognitive map" theory. In the phase procession paragraph there is a reference to how phase procession could help with localization. Other than that, nothing. I think more should be said about how the set of hippocampal neurons might work in aggregate, both in terms of computation of location and, potentially, other functions. The article is extremely dry without theoretical frameworks or suggestions of significance.


John Kubie

Chachaq 14:43, 17 September 2007 (UTC)

Hi John. I've moved this comment to the bottom of the page, where new sections are traditionally added. It's true that place cells historically are characterized as having primarily spatial correlates (and usually other requirements such as a certain level of spatial information or sparsity, a threshold number of contiguous spatial bins with elevated firing rate level above the background rate, etc.). It is hard to decide whether any cell has a primarily spatial correlate, except in particular controlled experiments, and it certainly seems reasonable to me that a given cell might have spatial correlates in one experiment and nonspatial correlates in another and hence would be sometimes a place cell and sometimes not. Consider the results over the past few years on context-dependent firing of CA1 units (Wood et al. 2000, Bower et al. 2005, to name just two). Is a cell that doesn't always fire in its place field a place cell?
Regarding your other point, it's certainly true that no one from that group is cited, and that's rather surprising. You should certainly feel free to add references for any work that you like. If anyone feels that you're overstating things, they'll likely change it or call you on it. Also, if you have a rate map image that you'd like to release into the public domain or license under a free license, certainly feel free to add it. At some less busy time in my future I'll try to work on this page some, but there are others that are in worse condition, so it isn't a large priority for me. digfarenough (talk) 20:30, 17 September 2007 (UTC)