Data Citation and the Author Byline: Who’s Line Is it Anyway?
An interpretation that would exclude ‘consortium’ authorship, however, results in an impasse for large-scale science that has adopted such practices. Some consortia, such as Alzheimer’s Disease Neuroimaging Initiative (ADNI),4 require, as part of their data use agreement, institutional authorship. Any journal that would prohibit such authorship, therefore, would cut off submission of manuscripts derived from these sources. This would probably not have an adverse effect on the publication of papers as these manuscripts would likely be submitted to journals that do not enforce such an authorship limitation/definition.
When a group name for a specific consortium, committee, study group, or the like appears in an article byline, the personal names of the members of that group may be published in the article text. Such names are entered as collaborator names for the MEDLINE citation. Collaborator names are entered for a MEDLINE citation only when a group (corporate) author name is present for the citation.
It is important to look at why institutional authorship is required by some organizations. The reason seems clear; it is the only currently available way to get an easily accessible, precise counting of publications of the use of the ‘product’ of a consortium. If you want to know how may publications ADNI data is associated with a simple query at PubMed7 “Alzheimer’s Disease Neuroimaging Initiative[cn]” yields 420 citations (on 5/1/2013). This precise number would be vary labor intensive to obtain through the counting of times that ADNI is cited in an ‘Acknowledgements’ section (someone would have to review every paper published to monitor for such occurrences) or through counting of citations to a specific set of data papers (papers can be cited even when the data is not used).
Given the amount of money invested into these various consortia and initiatives, it is clearly important to be able to generate accurate, easy to obtain quantitative metrics for the success of the effort. The question is if this ‘need for counting’ is best served by also usurping the traditional meaning of manuscript authorship? It is unlikely that any consortium investigator includes the 420 citations on their Curriculum Vitae, or that anyone reading the vitae would know what to make of that number if they did.
The problem is exacerbated in that there are many groups/consortia that would like a good counting system to support the assessment of the utility of their efforts. Not only data, but software, websites, equipment, etc. can all be the ‘product’ of a system that would value a precise counting of the utilization of their efforts. Using a ‘corporate author’ procedure for this counting just opens the floodgates for using authorship to count use of everything. Why shouldn’t the software development teams of the FMRIB Software Library (FSL),8 FreeSurfer,9 AFNI,10 Statistical Parametric Mapping (SPM),11 etc. all use corporate authorship to track their software utilization?
Therefore, a different solution to this problem is to make acknowledgement and attributions of resources used in manuscripts countable and trackable (in a similar way that current authorship and literature citations are currently). It was in someone’s interest when the tracking for traditional authorship in publications was first introduced, followed by the tracking of cited literature in articles. Both, in the end, driven by the ‘economics’ for metrics such as the authors ‘h-index’12 and a journals ‘impact factor’.13 It is now in the ‘best interest’ of the neuroscience community to better track the large-consortia investment in numerous ‘big science’ efforts in order to effectively document utilization. In principle, this tracking does not seem that it should be difficult to implement. The MEDLINE citation has already proved quite adaptable to the evolution of citation description needs. As currently constructed, a ‘Corporate Author’, unless it really did perform ICJME authorship activities and this is duly annotated in the manuscript, is essentially a ‘Corporate/Consortia Attribution’. Instead of being transcribed from the ‘author byline’, it should be extracted from the Acknowledgement section of the manuscript. Consortia development and credit for virtually any type of data, software or other resource can be attributed in this fashion, as long as the MEDLINE indexing is updated to follow such conventions. The neuroscience community (its scientists, publications, authors, editors and reviewers) needs to unite to insist on a course of action that preserves the meaning of individual authorship.
One might be tempted to ask “Is it too late already?” and suggest that the community should just embrace and expand their use of the current Corporate Author designation to its fullest. A paper, using the ADNI data, and software from FSL and FreeSurfer, should include ‘corporate’ authorship for each of these development teams. Yet this seems to diminish the scientific value of the ‘author’ concept in favor of the economic value of ‘count-ability’, and hopefully the answer is ‘no’. In support of this position and in response to the scientific community reactions, the journal Neuroinformatics plans to phase out the acceptance of manuscripts with Corporate Authorship, in favor of the adoption of Corporate/Consortia Attribution indexing, by January 2015.
Rohlfing T, Poline JB. Why shared data should not be acknowledged on the author byline. Neuroimage. 2012 Feb 15;59(4):4189–95.
Dai, Y., Wang, Y., Wang, L., Wu, G., Shi, F., & Shen, D. (2013). Alzheimer’s Disease Neuroimaging Initiative. aBEAT: a toolbox for consistent analysis of longitudinal adult brain MRI. PLoS One, 8(4), e60344.