Pre-Publication Policy at the Journal of Neurophysiology

Readers will recall Editor David Linden [ blog ] laying down an assertive editorial on the nature of the scientific publication process when he took up the reins at Journal of Neurophysiology. Now he is pushing to alter the rules of his Journal such that manuscripts that have been deposited in a pre-print archive such as arXiv or Nature Precedings are not excluded from consideration.
Editor Linden has requested our assistance in soliciting feedback on this proposed change which he intends to discuss at an upcoming meeting of the Publications committee of the American Physiological Society. If you have an opinion on this move, please respond to his mini-survey at the bottom of the letter which appears after the jump.


Dear Colleagues,
I am writing today to solicit your opinion on an important policy at the Journal of Neurophysiology. Presently, the journal will not review manuscripts that it considers to have undergone "prior publication." For most submissions this is not a problem: brief abstracts presented at meetings like Society for Neuroscience or Keystone conferences or Gordon conferences are not considered prior publication. However, in recent years it has become more popular for authors, particularly in the systems and computational neuroscience communities, to post full-length draft manuscripts on preprint servers like arXiv (www.arxiv.org) or Nature Precedings (precedings.nature.com). This is considered "prior publication" by the policy of the American Physiological Society. Here's the relevant text from the APS author guidelines:
"Material published by the author before submission in the following categories is considered prior publication: 1) articles published in any publication, even online-only, non-peer reviewed publications, such as Nature Precedings or the physics arXiv; 2) articles, book chapters, and long abstracts containing original data in figures and tables, especially in proceeding publications; 3) widely circulated, copyrighted, or archival reports, such as the technical reports of IBM, the preliminary reports of MIT, the institute reports of the US Army, or the internal reports of NASA."
These preprint servers have become a standard initial mode of scientific communication in the physics, astronomy and chemistry communities. They are permanent archives that are moderated (so they do not fill up with spam or political rants) but are not peer-reviewed. Authors submitting manuscripts to preprint servers retain the copyright to their work, which can then be transferred to the publisher when a later version of the work is accepted at a peer-reviewed journal.
Scientists who use preprint servers value the comments and feedback that help them to revise and improve their experiments, models and writing. In this way, preprint servers function similarly to scientific meetings. In many ways, they are better than feedback from a meeting talk because the audience has the chance to carefully pore over the details of the presentation and construct a more useful critique. In terms of citations and establishment of priority, these unreviewed preprints hold the same status as meeting abstracts--they are citable but do not carry the weight of a traditional peer-reviewed publication.
There is little reason to believe that a change of policy to allow preprint server manuscripts at Journal of Neurophysiology will have a negative impact on either the editorial function or the business model of the journal. Institutions are not going to cancel their JN subscriptions in favor of free preprint access. Journal of Neurophysiology should benefit from getting manuscripts that are better for having received more feedback prior to submission and from receiving submissions from authors who otherwise would avoid JN due to the prior publication policy.
Other publishers are rapidly adopting policies that allow preprint sever manuscripts. These include both for-profit publishers like the Nature Publishing Group (Nature, Nature Neuroscience and the other Nature titles), Elsevier (Brain Research, Neuroscience, NeuroImage, Journal of Neurobiology, Hearing Research, Neurobiology of Learning and Memory, Molecular and Cellular Neuroscience) and Springer (Journal of Computational Neuroscience, Neuroinformatics, Brian Structure and Function) as well as scientific societies such as Society for Neuroscience (Journal of Neuroscience), the National Academy of Sciences (PNAS) and open-access journals (PLOS Biology; Frontiers in Neuroscience, etc.) At present, Journal of Neurophysiology is one of a few remaining neuroscience journals not to adopt this policy.
It is my view that changing the guidelines at Journal of Neurophysiology to allow for submission of preprint server manuscripts can only improve scientific communication and benefit the Journal. It is important to stress that preprint server submission would be voluntary and that the authors' decision to make use of this process would have no bearing on editorial decisions.
The Publications Committee of the American Physiological Society will meet on March 18, 2009 and I have placed a motion to change the guidelines to allow preprint server manuscripts to be reviewed at JN on the agenda. When this issue was considered at the 2008 meeting of the committee, it was rejected. A key factor in the committee's decision will be a measure of where the neuroscience community stands on this issue. I'm asking you to take a moment to register your opinion. You do not have to be a member of APS to vote in this poll. You do not have to be faculty either: students, postdocs and staff are all welcome. You are encouraged to distribute this message to your colleagues. You can vote by filling in the info below and emailing to jlane12@jhmi.edu
Thanks for your time and best wishes,
David
David Linden, Ph.D.
Chief Editor, Journal of Neurophysiology
Your name:
Your institution:
------ Yes, I support amending the guidelines to allow preprint server manuscripts to undergo review at Journal of Neurophysiology.
------ No, I do not support amending the guidelines to allow preprint server manuscripts to undergo review at Journal of Neurophysiology.
Your comments (optional):

29 responses so far

  • Good news, but let's keep some perspective. The policy was draconian and stupid. This journal should absolutely follow this new policy, but this is not much of a move forward. It is a small but important step backwards from a questionable (though widespread, which does not make it right) practice.

  • Allowing full length manuscripts to be copublished in an APS journal or JN? Seriously doods? Does this mean that PLOS articles can be copublished or is it only articles without a "proper" DOI assignment?

  • Dave says:

    In days of yore, pretty much nothing appeared in print without someone explicitly claiming and enforcing copyright. Copyright enforcement was essential because someone had to bear the burden of generating and distributing the documents in question, and no one wanted to bear that burden without some protection. Remember that document creation and distribution was a substantial burden not too long ago (mid 1990's), before PDFs and photocopiers and Word processing and the web. Anyway, ubiquitous copyright enforcement can make reprinting stuff a hassle.
    (An anecdotal aside: I got my PhD in 1997 -- not too long ago, but a different era when it comes to documents. I remember when graduating I had to sign off copyright to my dissertation so that the university could publish & distribute it. Signing this paper was a dopey bureaucratic requirement for graduation. But I had to tell them I couldn't sign the paper, because I had already signed off copyright to most of the stuff in my dissertation to some journal or another. Of course I was not the first one to publish during grad school there, and when the issue went all the way to the dean he said "I recognize that it's technically a copyright violation to do so, and thank you for bringing this issue to our attention, but everyone else just signs the paper. You should do so too." So I did, and graduated. Don't tell the lawyers.)
    Anyway, now that the burden of 'publishing' stuff is so minimal, there is an increased tendency for authors to maintain copyright (granted to authors by default under U.S. law, even if not explicitly claimed). This is definitely the case for preprint servers. Which means it is easy for most authors to eventually sign over copyright to J. Neurophysiology or wherever.
    So really, the issue is not whether the document has been 'pre published', but rather whether the author(s) maintains copyright. The old policy made sense 10-15 years ago. The new policy makes sense now.

  • HI says:

    In terms of citations and establishment of priority, these unreviewed preprints hold the same status as meeting abstracts--they are citable but do not carry the weight of a traditional peer-reviewed publication.

    While I understand that these preprints do not carry the weight of publications on traditional peer-reviewed journals, I don't understand why submitting a preprint should not be considered as an establishment of priority. The date of submitting the manuscript is recorded. Scientists who read the preprint can judge whether the quality of the data and conclusions are sound. Even in the case of a peer-reviewed paper, approval by the reviewers does not guarantee the quality of a paper. There are some crappy papers that published in Nature, Science, and Cell. Ultimately, a value of a paper is decided by opinions of readers in the scientific community.
    If a submission to a preprint server is not given appropriate credit, as is done in physics and other fields, it will invite people to steal ideas. Also, when submitting a manuscript to a peer-reviewed journal, there is some chance of your publication substantially delayed, in some case by reviewers who are competitors. I think submitting a manuscript to a prepreint server can be a good way to counteract it, but only if there is enough credit given to a preprint.

  • Pro.
    Incidentally, smart letter:
    1) Here's the problem
    2) Here's our rule
    3) Anticipated effects
    4) The other cool journals are doing it.
    5) Democracy with a shove.
    nicely thought out. (just came from drdrA's post on paragraph structure; perhaps this is uppermost in my mind)

  • Anonymous says:

    ... unreviewed preprints hold the same status as meeting abstracts--they are citable ...
    Meeting abstracts are citable?
    I don't understand why submitting a preprint should not be considered as an establishment of priority. The date of submitting the manuscript is recorded.
    Notice that the date of formally submitting a peer-reviewed manuscript is also recorded, but doesn't count: only the actual date of publication counts. This was one of the big incentives for "immediate online publication" following successful peer-review. Even that date is sometimes (often?) ignored in favor of the actual date of in-print publication.
    I don't really understand the upside of pre-print servers. Do people really have time to browse these pre-prints in search of something interesting/relevant?

  • HI says:

    I don't really understand the upside of pre-print servers. Do people really have time to browse these pre-prints in search of something interesting/relevant?

    1. In physics they do. What's the matter with biologists?
    2. I often find myself searching pubmed to see if information that would help my research is already known. If I could get that information before publication, that would speed up my research.
    3. Do you remember the old days when we couldn't read a paper until we got hold of the printed journal? But getting a paper online didn't seem like such a big advantage. Things change.

  • 1. In physics they do. What's the matter with biologists?

    One major issue is that there are eleventeen fucktillion times more bioscientists--and thus manuscripts--than there are physicists.

  • Alex says:

    In some fields of physics they largely skip publication and just send stuff to ArXiv. In other fields, not so much. Even then, putting it there before publication is a way to establish priority, get ideas circulating, get feedback, etc. Publication in a peer reviewed journal gets more recognition for promotion, of course, because, for all the things that may be said about peer review, at least there was some sort of quality check (as compared to the preprint server) but pre-publication gets ideas out there. So different repositories for different purposes.
    I see abstracts cited in papers from time to time, but it's rare. For some reason, half the people who cite something in physics called the Purcell Effect (has to do with the lifetime of an excited state) cite an old conference abstract rather than a paper or book.

  • qaz says:

    Abstracts are ABSOLUTELY citeable. They are in print, dated, and permanent. Thus they are an appropriate mechanism to provide first report of data. Obviously, one needs to be aware of which ones have been replicated and which not, but the fact is that in many cases, abstracts are the only thing that is available to cite. While this certainly changes from field to field, in my experience, neuroscience abstracts (particularly SFN abstracts) are a very important part of the literature. There are cases where where the first person to report an effect was not the first person to publish the first full-length paper. In a specific case I know of, we all cite both the abstract (from 1984) and the paper (from 1990) that first describe the effect. In other cases, I have had follow-up work on an abstract that comes out before the full-length paper. There are lots of reasons why a paper might not make it out (ranging from asinine reviewers to graduating students to death of the PI). In my opinion, not citing an abstract one has built from is theft and plagiarism.
    I think it's a great idea to put stuff on Arxiv. (I don't because of current publication rules.) It's not that we would keep our eye on it, but that it would give our colleagues something to cite before the actual publication. In my field, we ALL know what each other is working on, we build on it before publication. Citing abstracts (or Arxiv) is critical.

  • Dave says:

    You can cite anything you want. Obviously, some cited sources are more credible than others.
    My personal favorite most absurd citation comes from Dean et al (2003) "Neurexin mediates the assembly of presynaptic terminals", Nature Neuroscience 6(7), pg 713:
    "...for further analysis and detailed methods, See Baksh, M., C.D., E.I. and Groves, J.T., unpublished data."

  • In my opinion, not citing an abstract one has built from is theft and plagiarism.

    Many journals in the biosciences explicitly prohibit citing abstracts or conference proceedings.

  • JD says:

    "Many journals in the biosciences explicitly prohibit citing abstracts or conference proceedings."
    Wow. I've certainly cited a published abstract a couple of times without issue. It's a useful thing to do when you see somebody do something that's pretty close to your idea at a conference. You wait for a bit. You email them and see if they have any plans to publish. Your paper begins to date but you want to acknowledge that you saw something similar elsewhere and it is unlikely that it did not influence your thinking.
    Thus citing an abstract makes sense. Encouraging the competing paper to come out is, however, much preferred.

  • DrugMonkey says:

    And would these be the exact same journals that encourage / permit statements followed by "data not shown" or "personal communication" or "unpublished data" by any chance?

  • qaz says:

    Comrade PP #12 - yes, what you say is true. That doesn't make those journals' policy correct or right. In a case I know of, a senior professor was able to get her paper published before mine (because she is senior enough to get past asinine reviewers). In this case, she wrote to me, told me her paper was in press, asked if I had a published paper she could cite and apologized profusely when all I had was an abstract. Needless to say, I'm mad at the journal, not at the senior professor, who did her best to cite correctly.
    JD #13 - Most journals encourage appropriate citation of abstracts. It's just a couple of arrogant journals with very strange and strict citation rules. (Note - this tends also to be the journals with citation limits, saying you can only cite so many other papers, no matter how many are actually important to the results.)

  • HI says:

    Comrade PhysioProf:

    One major issue is that there are eleventeen fucktillion times more bioscientists--and thus manuscripts--than there are physicists.

    I don't think that should be such a serious problem. You can always search preprints that are relevant to your research. Do you think that physicists don't screen preprints to read? If you are not interested in the newest information, you don't have to read them, either. I really think it's a matter of culture. PLOS didn't become the arXiv because there was so much resistance from biologists. In order for it to work, enough people must embrace the idea.
    Being in the field myself, I would say that there is certain intellectual laziness in life sciences (more than physics) of judging a paper by the journal where it was published. Also, there is an issue of equality of access to information, a point mentioned by Paul Ginsparg, who founded the arXiv, in his article in physics world. People in big labs have advantage because they can get information from manuscripts that are reviewed by their PIs or from personal communications between PIs.

  • DrugMonkey says:

    this tends also to be the journals with citation limits, saying you can only cite so many other papers, no matter how many are actually important to the results.
    Actively encouraging authors to engage in deceptive citation practices seems to be a very bad thing for science to me. Now where are our ethics-warriors to insist that the GlamourMagz are unethical and fraudulent and all that jazz for having these policies?

  • Dave says:

    HI and qaz, you are forgetting the purpose of citations. Citations are not for weiner-stroking or establishing precedence. Citations are pointers to essential background or supplementary information. They provide a service to the reader. Thus, citations must always be to information that is stable and widely accessible. If they are not, the citation is useless.
    It does not matter what someone did or when; he/she should not expect a citation unless the information is widely accessible (and thus useful to the reader). A private database, a crumpled piece of paper stuffed in an underwear drawer, an obscure journal only accessible to seventeen subscribers, a vague short abstract without any detailed information content, a personal communication -- all are generally useless or unverifiable to the reader and therefore not appropriate for citation, in my opinion.
    That said, I have broken my own guidelines above and made a few questionable citations. On each occasion I did so because: 1) I wanted to blame someone else for something I wanted to write but was too lazy to explain or gather evidence for, or 2) I wanted to suck up to someone (e.g. stroke weiner). I urge you to aspire higher than I have achieved.

  • DrugMonkey says:

    Citations are pointers to essential background or supplementary information. They provide a service to the reader.
    exactly. They form the tree upon which the current paper sits. It is best for scholarship if one can follow the threads of citation backward to accumulate the bulk of knowledge into which the present paper fits.
    Thus, citations must always be to information that is stable and widely accessible. If they are not, the citation is useless.
    No. Accessible? Yes. "Widely"? Not necessary. I've had to do some library requesting to find obscure, old or otherwise inaccessible papers many times in the past and I anticipate again in the future. The library system from which I get my electronic access is pretty big but still misses subscribing to the occasional odd journal which regularly publishes stuff of interest to me. Of course I can get everything on Nature Precedings at a couple of mouseclicks.
    Accessibility has nothing to do with whether you should or should not cite something. To gate on ease of accessibility is lazy, sloppy scholarship that will over time seriously distort appropriate academic crediting.

  • HI says:

    Dave:

    HI and qaz, you are forgetting the purpose of citations. Citations are not for weiner-stroking or establishing precedence. Citations are pointers to essential background or supplementary information. They provide a service to the reader. Thus, citations must always be to information that is stable and widely accessible. If they are not, the citation is useless.

    If I was not clear enough, I was thinking about a system like the arXiv. As far as I understand, stability and accessibility are the strength of a system like the arXiv. Preprints on the arXiv are MORE widely accessible than publications on most traditional journals.

  • Dave says:

    "Accessibility has nothing to do with whether you should or should not cite something. To gate on ease of accessibility is lazy, sloppy scholarship that will over time seriously distort appropriate academic crediting."
    When I said accessibility was important, I was thinking more extreme: Nature = OK; A lab meeting presentation = Not OK. Obviously there are grey zones: Non-English Journals, Meeting abstracts, Limited edition out-of-print books.
    Anyway, I think we agree, except your phrase "appropriate academic crediting" is a euphemism for weiner-stroking. As I said above, weiner-stroking is not really appropriate use of citations, although it might be professionally wise.

  • DrugMonkey says:

    I disagree that it is a euphemism. Or perhaps we have a different view on the meaning of 'weiner-stroking'.
    It is an essential part of scholarship to place your manuscript within a body of relevant work. Doing this involves citing relevant papers. Is it an endless debate over how many and which citations should be included for each point? Sure. Primary data articles are not the same as review articles. Nevertheless, there should be context. My rule of thumb is "about three" for major points. These should definitely include the first report and/or the best report previously published for the point at hand. Sometimes it includes a little mini-debate in the literature. Sometimes it requires a decades-hopping overview of the lit. etc. Focus is on the scholarly thread which supports the point you are making.
    My calculus on citation does not include citing some big wig just for politics, which is what I interpret you to mean by "weiner-stroking".

  • Can you assholes at least spell "wiener" correctly, for fuck's sake?

  • qaz says:

    Dave #18, there is a huge difference between appropriate citation and wiener-stroking. The academic tree from which a result derives is extremely important to placing it in the correct context. One function of citations is to give appropriate credit to the person who discovered something, NOT just to tell the reader where to read a recent review. Otherwise, why is the number of citations a paper has of any meaning at all? Citations are a way of saying "my work builds on this person", "I've discovered X, which depended on Y, but I'm not trying to claim Y." Dave, wiener-stroking is about buttering up some big-shot who may or may not deserve it. Citation is about credit.
    When I switched fields to neuroscience (from another field where journal articles are five years behind the current research and citations are usually to university tech-reports), I was taught to do very much like DM - one citation to give appropriate priority to the origin of the tree and one to a recent, thorough review. In cases where there are several critical steps that have not yet been reviewed cleanly, one should cite those steps. In cases in which there's a controversy, make sure you cite both sides of that controversy. In a world in which credit is judged by number of citations, to do anything else seems criminal to me.
    There are several reasons why one should cite abstracts. Perhaps the author of the abstract was not the first to publish the full paper, but the paper built on that abstract. Perhaps the paper has not yet come out. Citations need to access something stable, but most abstracts are stable. If you would like, you can still find copies of SFN abstracts from the 1980s in many libraries. If yours doesn't have one, then you can get it from interlibrary loan.

  • I don't think citing abstracts is that important. They haven't been peer-reviewed, and you know what? You can put anything you damn well please on your poster, but that doesn't mean it will stand any scrutiny at a reputable journal. I only cite abstracts when (a) the journal allows it, and (b) absolutely no one else has ever published this useful tidbit of information that bolsters my argument. I wouldn't bother to cite an abstract that I thought was an intellectual forbear to my paper (nor am I angry that the C/N/S paper that semi-scooped me recently failed to cite my poster....even though the authors came and hung out by it and took my ideas. I'm angry I didn't publish faster, which is different.)

  • Also, Dave? Citations are not for... establishing precedence.
    WTF? Citations are precisely for establishing precedence. Yes, they're also a tool to the reader, but that's the byproduct, not the purpose. The reason it's unethical not to cite relevant preceding work is not because it fleeces your poor reader, but because it shows intellectual dishonesty in failing to acknowledge idea/data primacy.

  • pumpkinesque says:

    I think bioligists should whole-hartedly embrace preprints like the arXiV. That the number of biologists > number of physicists should not matter; the physics arXiV is streamed into groups - classical and quantum gravity - condensed matter ect. If you have a paper relevant to two fields then you can cross-list it. You can submit your work to the arXiV, wait about a week for feedback (incase there will be any) then submit to journal (not favoured by all but I think exposing your work to more people only increases the quality and amount of feedback). But primarily, if someone has done a nice piece work that they then want to submit to Nature / PRL ect, then you don't have to wait for the months (or even year) to actually read it yourself, you can learn about what's been done before its gone through the peer review process and make your own judgements (and maybe pick up some new ideas / techniques a year or so earlier than if you only read it once it was published in big name journal).

  • Dave says:

    1) Thanks CPP for correcting my spelling. I before E and all that.
    2) I think our talk of citing abstracts has gotten too abstract. I agree with DJ&MH(#25), in that abstracts are generally unaccessible and unreliable foundations upon which to build one's argument. If the best support for your point is to say 'Well, I saw it on a poster of stuff that never got published...', then you are on mighty thin ice.
    My writing style is basically to know what the heck I am talking about, then write it all down in a furious vomit of information. Then I go back and add citations for anything I wrote that did not totally originate in my own mind during that vomit session. I don't worry about who was first, or the latest review, or anything like that. I simply try to provide my reader with the most relevant and useful pointer to the information. You could have been first to mention farkety-fark, but if you didn't say anything useful about it then there is no point in citing you. Unless of course I wrote something like 'farkety-fark was first conceived...' There is also no point citing your brilliant treatise in the non-English Hungarian Archive of Institutional Studies, especially if someone summarized it later in a review, because pointing readers to your paper will do them no good.

  • DrugMonkey says:

    hmm, posted this in the wrong thread originally...

    As it happens, there is an excellent post on the role of academic citation up at Green Gabbro today.
    http://scienceblogs.com/greengabbro/2009/01/academic_citation.php

Leave a Reply