A Bibliometric Analysis of the Top 100 Most Cited Chronotype Research Papers

Bibliometric indices are a widely used measure of research impact. The aim of the current study was to identify and characterise the top one hundred most-cited research articles in the topic of chronotype research. A search of the Thomson Reuters Web of Science database returned 974 eligible articles (published between 1990 and 2016). Citations for the 100 most-cited articles ranged between 438 and 29. The most represented journal was Chronobiology International (n = 30). Nearly 50% of articles originated in Germany and the U.S. The bibliometrics reported identify key publications and provide insight into trends within the topic of chronotype research.


Introduction
Circadian rhythms drive our preferences for wakefulness, periods of activity and sleep. The distribution of the timing of circadian rhythms, or chronotypes, ranges from those individuals that prefer to wake and retire to bed early to those who prefer to go to bed late and rise late, and it is becoming increasingly clear that chronotype impacts on many aspects of life including general health, mental health, cognition and addictive behaviours [1].
Bibliometric analysis measures the number of times an article is cited in the literature and in which particular journal. A citation is received when one article uses another as a reference and the number of citations an article receives can be used as an indicator of impact (an article considered to have greater importance by the research community is more likely to be cited). In addition, citation count also forms the basis for computation of the journal impact factor (IF), which is a widely used measure of journal quality. A review of the mostcited articles within a research field can, therefore, reveal topics of current interest, novel techniques and research trends. The aim of the current study was to identify and analyse the one hundred most-cited publications in the topic of chronotype research.

Article identification
The most-cited chronotype research articles were identified using the Institute for Scientific Information (ISI) Web of science search tool (Thomson Reuters, New York, NY). The ISI web of science is a multidisciplinary database providing a range of bibliometric information indexing over 8500 journals from a wide range of disciplines. Using the topic search term "chronotype" and title search term "*chronotype*" combined with the Boolean operator "or" (i.e. returns the value true if either or both of the operands is true) a total of 976 journal articles published between 1990 and 2016 were identified. Searching by the topic term alone returned 842 items, whereas searching by the second term alone returned 321 items. The 100 most cited chronotype papers were identified from this list. The author reviewed the title and abstract for each article to ensure its relevance (one article was excluded in this process and replaced by the next highest article in terms of citation count).

Article analysis
The following information for each of the top 100 mostcited articles was extracted: 1) authorship, 2) article title, 3) keywords, 4) journal title, 5) year of publication, 6) total number of citations, 7) annual citation count (total number of citations/time since publication), 8) impact factor (IF) , 9) source normalised impact per paper (SNIP), 10) impact per publication (IPP) , 11) the SCImago journal rank (SJR), 12) country of origin (based on the affiliation of the first author at time of publication) and 13) article type (original research, meta-analysis or review). of Biological Rhythms in 2003. Ranking articles by annual citation count, as compared to total citation count, resulted in a change in position range of -58 to +66 (where a negative number reflects a higher ranking based on annual citation count). The article with the most annual citations (rank change -4) was "Social jetlag and obesity" by T. Roenneberg, published in Current Biology in 2012. The journal with the most publications in the top one-hundred was Chronobiology International (n = 30), followed by the Journal of Sleep Research (n = 8) and Sleep and Journal of Biological Rhythms (both, n = 5). The median/range IF, SNIP, IPP and SJR was, respectively, 3.11/0.53-13.31, 1.24/0.43-3.10, 3.27/0.64-10.58, 1.37/0.25-5.93 (see Table 2). Figure 1 shows the distribution of publications by year. The year with the most publications from the top onehundred most-cited articles was 2007 (n = 14) (Figure 1). The most frequent first author was Randler, C (n = 9) ( Table 3), and the majority of articles originated in Germany (n = 26) (Figure 2).
The most frequent article type was original research (n = 89) and of these 54 (61%) were entirely questionnaire based. Ten review articles and one meta-analysis were also included in the top 100 most-cited articles.

Discussion
Bibliometrics are frequently used in the library and information sciences to provide quantitative analysis of academic literature. This approach can be used to explore the impact of a field or research discipline, the impact of an individual/group of researchers or the impact of a particular paper.
The top one-hundred most cited articles in the field of chronotype were cited between 438 and 29 times. Annual citation count was also considered to evaluate current impact of an article. Although there were a number of substantial shifts in ranking, the articles with the top 5 annual citation counts were also in the top seven for total citation count thereby suggesting both historical importance and current relevance.
Perhaps not surprisingly, nearly 50% of the most-cited papers were published in discipline-related journals (Chronobiology International, Journal of Sleep Research, Sleep and Journal of Biological Rhythms, Biological Rhythm Research). The remaining publications encompassed a broad range of disciplines (e.g. psychology, psychiatry, nutrition and addiction) suggesting a wide audience and broad appeal of chronotype research.      With regard to year of publication, an interesting pattern of results was revealed. Seventy-six percent of the most cited articles were published post 2006 (Figure 1). The large increase in most-cited publications from 2007 onwards was mirrored (with a lag) by an equally sharp increase in total publications per year. The broad range of journal disciplines represented in the top one hundred most cited articles and the surge in publication rate suggest expanding interest in chronotype generally and increasing awareness of the impact of chronotype on a broad range of factors including cognitive function, general health and psychopathology.
Over a quarter of the top one hundred most cited articles originated from Germany. It seems unlikely that this can be attributed to the size of the scientific community or underpinning resources as this number outstripped the U.S. Rather, this figure appears to reflect the work of two principal investigators, Professor Till Roenneberg (Institute of Medical Psychology, Ludwig-Maximilian University, Munich) and Christoph Randler (University of Tübingen).
Lastly, the majority of most-cited articles represented original research, of which 61% were entirely questionnaire based. There appears to be scope to draw from other research disciplines and include a broader range of techniques and methodologies (e.g. neuroimaging, endocrinology and genetics).
Blibliometric analyses can provide insight into the development of a research field and highlight patterns, strengths and research trends over time. The approach is not without its limitations, however, and these should be taken into consideration when interpreting the current data. For example, total citation count may preferentially bias older articles. To account for this, a time adjusted metric (annual citations) was also reported. Second, the articles identified for inclusion are dependent on the search parameters selected. In the current study only publications from 1990 onwards were included and the search term was restricted to chronotype. The choice of search parameters may, therefore, exclude influential papers (e.g. Horne, J. A., & Östberg, O. (1976). A selfassessment questionnaire to determine morningnesseveningness in human circadian rhythms. International Journal of Chronobiology, 4(2), 97-110). Further considerations are obliteration by incorporation (influential research subsumed into common knowledge and therefore not cited), omission-bias (ignoring contradictory results) and self-citation. Finally, a direct comparison of the top citation counts across fields or disciplines is challenging and may detract from the importance of individual contributions to a small, specialised field of research. For example, the most cited genetics article had 22,085 citations. A search using the term sleep revealed a top-cited article with 6,365 citations. Nevertheless, the data presented in this report suggest that interest in chronotype research is increasing. Future chronotype citation reports may benefit from the inclusion of altmetrics that allows a broader analysis of the use of scholarly articles and includes not only citations but views, downloads etc.
In conclusion, the importance of assessing impact (on both the research field and the broader general public) is becoming increasingly important, and many researchfunding bodies use impact as a means to justify their spending to governments and donors. The bibliometic analysis reported here included the top one hundred most cited papers in the topic of chronotype research. The data provide insight into the impact of chronotype research, the relative importance of a journal within the field (IF), the impact of a particular article (as indexed by citation count) and the contributions by specific researchers or groups.

Disclosure statement
The author does not have any financial interest or benefit arising from the direct application of this research.

Ethics statement
This study was a bibliometric analysis that did not include procedures involving animal or human participants and was, therefore, exempt from the requirement for ethical review and approval.