Friday, October 13, 2017

Is Journal of Neuroscience discriminating against behavioral research?

Journal of Neuroscience is the flagship scientific publication for the Society for Neuroscience, a journal that was established in 1981 “to publish research in the field of neuroscience and to serve the membership of the Society for Neuroscience”. Unfortunately, available data suggest that the journal’s impact has been declining, at least as evidenced by the Journal Impact Factor (JIF). Perhaps as a reaction to these trends, in the past two years policy statements from the journal’s editorial board, and anecdotal evidence from the review process, suggest that the journal has instituted a policy of discrimination against a specific kind of neuroscience research: non-invasive, psychophysical experiments are rejected without peer review.

I am alarmed by this change. I believe that the new policy is flawed for a number of reasons.
  • The available data suggest a poor relationship between rejection rates and journal impact factor, making it unlikely that increasing rejection rates will result in improved impact factor.
  • The focus on rejecting a specific body of research is discriminatory and runs counter to the principle of representing research of all members of the Society for Neuroscience.
  • The current policy, as applied to the journal’s own data would eliminate some of the highest cited papers in the journal’s history. That is, the new policy may worsen the plight of the journal.

The declining impact factor and the journal’s new policies
The Society for Neuroscience is an organization that represents scientists who study the field of neuroscience at all levels, including molecular, cellular, systems, behavioral, and computational. Unfortunately, over the past 20 years the flagship journal of this society has seen a general decline in its impact factor.





The reasons for this decline are unclear, though I speculate that one factor is the rapid rise of new, competing journals. Perhaps in response to these trends, in January 2016 the journal published an article describing changes to the review process. The new policy aimed to increase the rate of rejection of papers at the editorial stage, before they are submitted to peer review. The data that were included in a follow up article showed that from January 2014 to March 2016, the total percent of papers rejected increased modestly from around 70% to nearly 80%, while the percent of papers rejected before peer-review increased by a factor of five: from 5% to about 25%.

The new policy described the kind of research that the editors rejected without review. They wrote: “… purely biophysical or behavioral studies should provide novel insights into, and make specific predictions about, neural mechanisms or neural representations.” A typical rejection letter from the senior editor gave the following reason for rejection without review of a study that used psychophysical, non-invasive recording techniques to measure behavior: “we felt that The Journal of Neuroscience is not the right venue because the findings do not inform us about or implicate a specific neural mechanism.”

It appears to me that the journal’s new policy is targeting a specific group of neuroscientists: those who primarily rely on behavioral data to infer function of the brain. Given that this is the official journal of the Society, the discrimination would divide the membership into classes that can and cannot publish in the journal.

What to do
To help address these issues, we need data that quantifies how the policy is affecting the ability of the community to publish in the journal. To help with that, I ask that you to take a few moments to complete this survey.

The objective of the survey is to collect data regarding the publishing experience of scientists who may have been affected by this policy. We hope to be able to answer a simple question: what has been the impact of papers published in the journal that have relied primarily on psychophysical tools to measure behavior? We hope to present the data to the journal’s editors, clarifying the impact of their new policy.
Thank you for your help. 

13 comments:

  1. Hi Reza, thanks for this insight! Indeed, we recently had our psychophysics paper also rejected without review (editor: "We felt that this psychophysical study on sound localization in the mid-sagittal plane, while well-executed, is mostly of interest to a specialized audience, and there is no clear link to specific brain structures or encoding mechanisms. Although the discussion makes note of interesting cross-species differences and dissimilarities between the visual and auditory systems, it does not draw mechanistic insight from the behavioral results.") I just couldn't believe my eyes! It's even demonstrably wrong!
    I have published more than 10 JN papers with a similar scope (and we always made clear links to neural and neurocomputational principles...), which have received lots of citations. However, according to current standards of this journal these papers would never have passed the scrutiny of the editorial board. I am flabbergasted. They are clearly misguided
    (but note that I had more strange encounters with this journal: e.g., for using my own figures in my book JNeurosci wanted to charge me almost US$ 3000! And mind you, we already pay them substantially for publishing our own work!; I obviously refused to go along with this ludicrous quotation).

    ReplyDelete
  2. I have never had good luck with the Journal of Neuroscience, even with EMG recordings, which reflect the activity of motor neurons. My last attempt was in 2013 and it took over a year in review and we ended up being scooped by a matter of days. I don't think this extended time in review is a useful education for my students or a good use of my time. I think our papers speak for themselves and so I would prefer to go to a journal where I'm not having to fight the premise that I am even doing neuroscience research. We could have sent our recent paper on muscle spindles to Journal of Neuroscience (including actual spike trains!) but I was not convinced that studying a peripheral sensory organs that forms the essential input for the most basic sensorimotor circuit would be sufficiently "neural" for the journal. The Journal of Neurophysiology has been thanking us for submitting there and also actively promoting our research through podcasts, press releases, and social media. Even if we were to get published in the Journal of Neuroscience, I don't feel the work would be as appreciated.

    ReplyDelete
  3. Now that you got me thinking about it, the loss of behavioral research is dangerous to neuroscience. The idea that we will understand the brain without a nuanced understanding of behavior, i.e. the only neural output that ultimately matters, is misguided. I would submit that any paper on neuroscience research should be required to discuss behavioral correlates.

    ReplyDelete
  4. A journal that represents the Society should not behave in a way that makes some members feel as second class citizens. How is alienating fantastic scientists like you going to help reverse the trends? Thank you for sharing your comments.

    ReplyDelete
  5. Hi Reza, this is a very timely and important issue, but what do you expect when the very editorial board concerning neuromotor control issues is so negative and borderline rude when it comes to serious math. It is as though just because they do not have the training and assume their theories are right, they ought to reject everything else that seems different. I have seen many people in the field suffer from this, but the rejections are not from someone doing vision or auditory related work. They come from the very people who should be helping the growth of their field. Motor control is dead as far as I am concerned. Computational models are not addressing contemporary issues in clinical domains and with the shifts taking place in the NIH/NSF policies, I worry about the future of behavioral analyses. It has been reduced to observation and video coding, polluted with heuristics and very little self-emerging phenomena. Good luck with this. You have always been a pioneer in my eyes, so perhaps this time around you will start a new wave to revamp the field. I agree with your points but recommend enriching your board of editors and reviewers as we do in other journals these days (e.g. Frontiers in Neuroscience rapidly raising its impact factor). Also, you may want to change the review process to a more interactive, open system that prevents people from being borderline disrespectful to their own colleagues.

    ReplyDelete
  6. This has been an on-going problem, one I first encountered in 2008. My colleagues and I had submitted a psychophysics paper to JN, which rejected it on the basis of one review——a badly flawed review. I wrote the General Editor, who explained that the journal's failure to find a second reviewer was itself an indication that the topic of the paper didn't match the focus of the journal (despite the fact that I had published on a psychophysical paper on the same topic in the journal two years earlier). And what was this focus? Studies providing evidence dealing "directly with underlying neurobiological mechanisms." In other words, for JN psychophysics falls outside the primary focus of neuroscience.

    Two and a half years later we sent another paper we sent to JN and got a repeat of the same nonsense. I got the impression the journal was sacrificing science for its sense of exclusivity. Since then, I don't even consider submitting to JN.

    ReplyDelete
  7. hi Reza et al

    i’m on the JNeurosci editorial board and (speaking for myself) found this blog to be very interesting and constructive. i will continue to monitor your thoughtful comments as this topic will be discussed at the journal’s board meeting next month.

    you raise several important issues, but the specific concern is that JNeurosci has acted to exclude an area of neuroscience research (i.e., behavioral data from which one can infer brain function).

    for that purpose, it would be valuable to hear your opinions about the characteristics of research that belong in JNeurosci. for example, are there behavioral studies from which one cannot infer brain function? more generally: (1) should JNeurosci have any submission criteria and, if so, (2) what should these criteria be?

    thanks for sharing your concerns. feel free to contact me direclty (dhs1@nyu.edu) if you would like to schedule a real-time conversation.

    Dan Sanes

    ReplyDelete
  8. The stated editorial policy regarding desk-rejections, and the survey results regarding actual letters authors have received, give me the impression that there may be a bias against a specific kind of research: behavioral neuroscience.

    In principle, desk-rejections are a powerful tool. The journal does not want to give the impression that a few are able to impose their personal views on the entire community. Rather, the impression should be that the journal is a welcoming place that will give a fair shot to all members of the society, and the power of desk-rejection is exercised with transparency and accountability.

    For example, I do not have the confidence to state what kind of research should or should not be published in JNeurosci. However, I can give you my opinion on a specific paper submitted for consideration.

    Thank you for reading these comments and contributing to them.

    ReplyDelete
  9. Hi Reza

    A change in policy seems evident from scanning the emailed table of contents. For the last 18 months or so, there is little or no behavioral neuroscience research (even in the 'Behavioral/Cognitive' section), in marked comparison to previous years.

    ReplyDelete
  10. It seems two different points are being conflated in this discussion:
    1. The value of behavioural research; 2) the proper scope of your specific science (neuroscience)

    It seems that the editors of JN are not denying the value of behavioural research, but are trying to keep the scope of the journal to the journal’s definition of neuroscience, in part to relieve rising pressure on their review process.

    A standard way of defining a science is in terms of the methods it employs, so the editors’ position seems reasonable (to me, an outsider). However, Prof. Shadmehr makes a good point that the scope of the journal might also be determined by membership of the society.

    Whether this scope is to be defined by methods or society membership, this definition is the prerogative of the society as a whole, not just the editors of its journal. I suggest that you direct your activism to making sure that the society resolves this issue regarding the scope of neuroscience and the journal at the society’s next general meeting in a way that involves its entire membership (e.g. by open debate, followed by a vote).

    Although I say I am an outsider, as a psychologist, I belong to one of several allied sciences concerning behaviour that have embraced multidisciplinary collaboration/communication, but are still struggling to transition from a superficial openness to principled integration. How you resolve this problem matters to all of us. Set a good precedent!

    Robert Ian Bowers
    Centro de Investigação em Psicologia
    Universidade do Minho

    ReplyDelete
  11. If anyone would like to discuss this issue with me at the SfN meeting, I will be in the JNeurosci booth on Sunday (Nov-12) from 1000-1100, and Monday (Nov-13) from 1100-1200.
    best
    Dan Sanes

    ReplyDelete