Edited by: Henrik Walter, Charité – Universitätsmedizin, Germany
Reviewed by: Markus Christen, University of Zurich, Switzerland; Ralf J. Jox, University of Munich, Germany
*Correspondence: Jon Leefmann
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
In bioethics, the first decade of the twenty-first century was characterized by the emergence of interest in the ethical, legal, and social aspects of neuroscience research. At the same time an ongoing extension of the topics and phenomena addressed by neuroscientists was observed alongside its rise as one of the leading disciplines in the biomedical science. One of these phenomena addressed by neuroscientists and moral psychologists was the neural processes involved in moral decision-making. Today both strands of research are often addressed under the label of neuroethics. To understand this development we recalled literature from 1995 to 2012 stored in the Mainz Neuroethics Database (i) to investigate the quantitative development of scientific publications in neuroethics; (ii) to explore changes in the topics of neuroethics research within the defined time interval; (iii) to illustrate the interdependence of different research topics within the neuroethics literature; (iv) to show the development of the distribution of neuroethics research on peer-reviewed journals; and (v) to display the academic background and affiliations of neuroethics researchers. Our analysis exposes that there has been a demonstrative increase of neuroethics research while the issues addressed under this label had mostly been present before the establishment of the field. We show that the research on the ethical, legal and social aspects of neuroscience research is hardly related to neuroscience research on moral decision-making and that the academic backgrounds and affiliations of many neuroethics researchers speak for a very close entanglement of neuroscience and neuroethics. As our article suggests that after more than one decade there still is no dominant agenda for the future of neuroethics research, it calls for more reflection about the theoretical underpinnings and prospects to establish neuroethics as a marked-off research field distinct from neuroscience and the diverse branches of bioethics.
Ethical questions related to the treatment of those, who suffer from neurological and mental illness, have been at issue in one or the other way since the beginning of medicine (Berrios,
As philosophers have learned from longstanding debates in the philosophy of science, the demarcation of a scientific research field is partly a prescriptive and stipulative issue and partly a descriptive and explanatory issue (Laudan,
The variety of conceptualizations of neuroethics, which have been proposed since 2002, suggests that defining neuroethics has been primarily an issue of adequately marking off the field from other branches of bioethics on the one hand and from neuroscience on the other hand. As a first mapping of such definitional approaches, Eric Racine has suggested that we distinguish between technology-driven, healthcare-driven and knowledge-driven definitions of neuroethics (Racine,
With this study, we aim to complement the ongoing process of conceptualizing neuroethics as a research field with new empirical data from the Mainz Neuroethics Database (MND). Drawing on a large corpus of the scientific literature from the last quarter century, we are interested in two different aspects of neuroethics: First, its development into an institutionalized research field distinct from neuroscience and the diverse branches of biomedical ethics, and second the relations between the seemingly heterogeneous research questions within the field. To understand the current structure and development of neuroethics, we focus on the years between 1995 and 2012 when neuroethics slowly begins to become visible at the margins of the established disciplines of neuroscience, bioethics, medical ethics and the philosophy of mind, and starts to develop into a widely observed research field represented by particular scientific journals and research facilities. Using bibliometric analysis, we (i) investigate the quantitative development of scientific publications in neuroethics, (ii) explore changes in the topics and guiding themes of neuroethics research, (iii) illustrate the interdependence of different research topics within the neuroethics literature, (iv) demonstrate the development of the distribution of neuroethics research in peer-reviewed journals, and (iv) display the academic background and affiliations of neuroethics researchers. These investigations will help to highlight the main differences and similarities of neuroethics and adjoining scientific disciplines and contribute to a better understanding of the current state of neuroethics. Previous bibliometric work on this field has drawn on much smaller subsets of literature (Racine,
Our empirical analysis draws on the scientific literature stored in the Mainz Neuroethics Database (MND) (available at
To allow for bibliometric analysis of the versatile material stored in the MND, we first performed a series of heuristics and manual checks to disambiguate author names present in several orthographical variations (for example, Baumeister, Roy and Baumeister, Roy F are merged into a single author name). This ensured a consistent identification and count of authors across the database and lowered the number of relevant publications. Second, we restricted the data corpus to articles that appeared between 1995 and 2012—a time interval that centers around the year of the founding conferences of current neuroethics and which therefore appeared most relevant for monitoring the development of the field. This left us with a corpus of 2296 relevant references. Third, publications without an English abstract and keywords were sorted out, as they did not provide enough information for our procedure of automatic thematic categorization of publications. This step excluded the bulk of editorials and opinion papers as well as most of the non-peer-reviewed material from the dataset, 371 publications (16.16%). Forth, each of the remaining 1925 publications was categorized according to a set of topical subject-categories (cf. Section Subject-Category Development and Supplements
Subject-categories were created using the three subsequent steps of coding, identification of relevant keywords defining the subject-categories, and checks for the reliability of keyword matches.
Based on the headlines and abstracts, 400 randomly retrieved publications from the dataset were coded according to the topics they addressed. The coding was done independently using a close reading approach by EH and JL, who are both experts in the field of neuroethics. The coding was reviewed for reliability by another research assistant. By careful evaluation of abstracts, headlines and keywords of the bibliographical references, this procedure allowed for development of a system of 44 subject-categories. Candidates for subject-categories were general topics addressed in these text components. For example, “Enhancement” and “neurodegenerative diseases” were suitable candidates for such general topics, because both occurred in the literature with regard to several specific contexts, such as “pharmacological cognitive enhancement” and “brain-stimulation” or “Alzheimer's disease” and “aging” respectively. As the number of publications per subject-category strongly varied after coding, similar subject-categories, which contained less than five publications were merged into a new, more general subject-category. This slightly reduced the number of subject-categories from 44 to 38.
The second step was conducted together with CL, an expert in both neuro-social sciences and the methodology of computational linguistics, and aimed at defining a list of keywords for each subject-category to allow for an automated matching of publications with subject-categories. This step comprised itself five operations (c.f. Supplement
Psychiatric and neurodegenerative diseases and disorders | 49 | 69.39 | 30.61 | 342 | 64.33 | 35.67 |
Neuroimaging | 47 | 74.47 | 25.53 | 279 | 70.61 | 29.39 |
Moral theory | 30 | 56.67 | 43.33 | 181 | 43.09 | 56.91 |
Philosophy of mind and consciousness | 28 | 42.86 | 57.14 | 155 | 38.71 | 61.29 |
Medical research and medicine | 24 | 50 | 50 | 137 | 37.23 | 62.77 |
Neurosurgery | 22 | 59.09 | 40.91 | 115 | 40.87 | 59.13 |
Legal studies | 19 | 36.84 | 63.16 | 108 | 27.78 | 72.22 |
Neuroscience and society | 14 | 35.71 | 64.29 | 81 | 34.57 | 65.43 |
Psychopharmacology | 18 | 66.67 | 33.33 | 88 | 50 | 50 |
Social and economic neuroscience | 18 | 72.22 | 27.78 | 76 | 75 | 25 |
Brain death/severe disorders of consciousness | 14 | 57.14 | 42.86 | 82 | 46.34 | 53.66 |
Brain stimulation | 13 | 61.54 | 38.46 | 65 | 53.85 | 46.15 |
Enhancement | 11 | 45.45 | 54.55 | 59 | 32.20 | 67.80 |
Molecular neurobiology and genetics | 2 | 50.00 | 50.00 | 6 | 50.00 | 50.00 |
Addiction | 4 | 25.00 | 75.00 | 19 | 15.79 | 84.21 |
To check whether the keywords defining the 15 subject-categories were able to cover all the relevant topics of the publications in the database and to control for the adequacy of computer-based assignments of subject-categories, we manually checked the computer-based category assignments of 100 randomly selected publications. Our control suggested an inadequate assignment of categories in only 12% of the articles, a measure, which we consider still sufficiently accurate. To verify that the content of the bibliographical references harvested in each topic was indeed in line with the topic, we additionally analyzed the terms used in the abstracts, titles and keywords of the publication, which were assigned to a subject-category and evaluated wether they matched with the topic of the subject-category. This was achieved by using a “bag of words”-model, which implies counting all terms included in the bibliographical reference categorized as matching a subject-category and using this count as a quantifier for the relevance of a term relative to the subject-category. As the most common terms for all categories would usually be common expressions like “this study shows” or “we conclude,” in a first step of word-bag processing we removed general stopwords (such as “a,” “the,” “have”) and stopwords specific to the scientific domain (such as “we find,” “the hypothesis tested,” “we conclude” etc.). Then, we calculated the term frequency-inverse document frequency (tf-idf) for each term by discounting the frequency of a given term in a subject-category from its frequency in all bibliographical references across subject-categories. This takes into account that a term is all the more descriptive of a subject-category, in which it is specifically used. For example, a term used many times in the bibliographical references classified in the category “Neuroimaging,” which is never used in bibliographical references assigned to other categories, has a high tf-idf score in the category “Neuroimaging.” The results of this reliability check are presented in Supplement
To estimate the development of neuroethics as a discipline we first looked at the quantitative development of published work in the field. Looking at the distribution of publication dates, we found a general increase in publications within the selected time interval. While only eight publications with references to neuroethics have appeared in 1995 (0.35% of all publications) in 2009, there were already 590 publications (25.7%) in the database. Within the examined dataset, the increase in the number of publications was continuous except for the interval between 2009 and 2012. While in this interval the number of publications per year remained on a comparably high level, instead of increasing any further the number of publications per year fluctuated between 145 (6.32%) in 2010 and 348 (15.16%) in 2011. This striking fluctuation does not represent a feature of the development of neuroethics, but is very likely an artifact created through the maintenance of the MND. A closer analysis of the publications added to the database in the years 2009 and 2010 revealed discrepancies in the number of journals that contributed paper to the database (274 in 2009 and 83 in 2010) and the number of monographs and book chapters included. As our data show, this variety concerns all subject-categories and can only partly be explained by the founding of the journal “Neuroethics” and “AJOB Neuroscience” in 2008 and 2010 respectively. It is more likely that it relies on inconsistent in- and exclusions of publications in the years 2009 and 2010.
The general trend of the temporal development of publication numbers in the Mainz database is consistent with the development of the number of publications in other databases such as Web of Science (WoS). We tried several protocols for searches in the WoS database with only one protocol (neuro*/topic AND ethic*/topic) providing an approximately similar number and distribution of articles per year. In both databases there is a moderate increase in the number of publications between 1995 and 2005 (+8.3% in MND and +6.5% in WoS), a stronger increase between 2005 and 2009 (+101.4% and +32.6% respectively) and an inconsistent development between 2009 and 2012 (−109.7% and +17.7%) (cf. Figure
The procedure of subject-category formation provided a structure, which divided the content of the database into 15 subject-categories (see Supplement
There is a general trend in all subject-categories of an increasing number of publications. With only few exceptions most of the publications within one subject-category appeared after 2005 (for details see below). In spite of some variations (for example in the relative weight of the subject-category “
To further investigate the relation between the different subdomains of neuroethics, we determined how subject-categories relate to each other from a content-based point of view. To assess the strength of the connections between two subject-categories, we determined their pointwise mutual information (
We found three groups of closely interrelated subject-categories. The first group comprises literature associated with topics in clinical medical ethics (
Having identified neuroethics as thematically being an endeavor that falls between very diverse and partly overlapping research fields, we wondered how the rise of neuroethics in the last quarter century maps on the landscape of scientific journals publishing contributions in the field. Scientific journals contribute to delineating and structuring scientific disciplines, through their differentiated editorial selections, or more mundanely through the central role they are made to play in the decisions for career promotions for researchers. For this reason, we found it particularly interesting to determine where contributions in neuroethics are published. Hence, using our expertise from our previous research in neuroscience and neuroethics, we distinguished journals of two broad categories: journals focusing on empirical research in medicine and natural sciences were labeled as “biomedical science journals,” while journals publishing primarily qualitative research in the social sciences, philosophical enquiries or ethical issues, were labeled as “social sciences and humanities journals” (SSH journals for short) We controlled this grouping of journals by drawing on the ISI journal classification scheme (cf. Supplement
As the table shows, the overall size of the subject-category in terms of the total number of articles correlates with the total number of scientific journals within the subject-category. Hence,
The four subject-categories with a biomedical science journal to SHH journal ratio lower than one, however, do not come as a surprise:
This analysis only allowed a glimpse into the larger picture of publications in neuroethics, because we did not identify changes in the publication patterns of different subject categories over time. Nonetheless, neuroethics has also begun to establish itself as a discipline by the foundation of two very central scientific journals, Neuroethics (founded in 2008) and the American Journal of Bioethics: Neurosciences (AJOB Neuroscience for short) (founded in 2010). As the foundation of scientific journals especially dedicated to a certain research field can be expected to change publication patterns within the field, we wondered what impact these two widely acknowledged journals had on the distribution of neuroethics publications in scientific journals. To estimate this influence we compared the distribution of papers within the different subject categories in the interval between 2004 and 2008 and the interval between 2008 and 2012 (cf. Table
1995–2007 | 49/105 | 704 | 0 |
2008–2012 | 86/105 | 1592 | 374 |
Our analysis indicates that the foundation of the two journals significantly contributed to the structuring of the field, but that nevertheless, less than a quarter of all relevant papers are published in these journals (cf. Figure
Besides the study of disciplinary journals, the institutionalization process of a research field can be described by the development of research centers and research groups that focus on questions relevant in the described subject-categories. In order to further survey the institutionalization of neuroethics we therefore examined the current affiliation of the most relevant researchers of each subject-category as well as their educatory background. As a first step, we determined the most relevant researchers in each subject-category by retrieving the names of the 20 researchers with the highest number of publications from each of the subject-categories. When two or more researchers with the same number of publications were ranked as number 20 in the respective subject-category, we preferred those with the higher number of first author publications over those with second or third author publications. As several researchers belonged to the top 20 in more than one subject-category the procedure provided a list of the names of 198 researchers. Queries in Google, WoS, and in the MND allowed us to add the academic discipline(s) in which these researchers were initially trained, their highest academic degree obtained, their current affiliation and the country of their current host institution (cf. Supplement
General findings indicate that a majority of researchers in the field of neuroethics have a background in medicine (40.43%), while another 24.47% have entered Neuroethics after having been trained in psychology (14.36%, including behavioral science, cognitive psychology, social psychology, experimental psychology, cognitive science) or neuroscience (10.11%, including psychobiology, neurological science). These finding support the observation that neuroethics is primarily fueled by research from the empirical medical sciences and the life sciences (including psychology). Scientists with a genuine competence for normative and conceptual questions, such as scholars of philosophy and bioethics seem to be not adequately represented in neuroethics, at least if the normative character of many research questions and debates in the field is considered. Nonetheless, according to our database scholars with an initial or additional training in philosophy or bioethics form a proportion of about 17.02% of the authors in neuroethics.
With regard to disciplinary affiliations neuroethics, researchers display surprisingly little mobility. Even though neuroethics is a highly interdisciplinary field and even though there aren't yet any structured programs for becoming a “neuroethicists” so that one might consider a high disciplinary mobility to be an advantage on the job market, we found that many of the scientists who publish in neuroethics stay affiliated with departments of the discipline in which they were initially educated. Most researchers still have their affiliation with a medical center or a science research institute (67.2%). While most of the identified neuroethics researchers are affiliated with university medical centers (43.26%), another 15.96% work in neuroscience laboratories and another 7.98% in psychology departments at universities or private research centers. Hence, also with regard to affiliation, the social sciences and humanities are underrepresented in neuroethics. Only 7.98% of the researchers in our dataset were affiliated with a bioethics department, and another 4.26% with a department specifically dedicated to neuroethics. The complete spectrum of the social sciences (sociology, political sciences, economics) only made up 3.27% of the departments hosting neuroethics researchers with a high publication output (cf. Figure
Besides the US, researchers in neuroethics with a high publication output primarily work in Canada, the UK, Germany and Australia (cf. Table
Australia | 0.329 | 11 |
Belgium | 0.15 | 3 |
Brazil | 2.82 | 3 |
Canada | 0.49 | 19 |
Denmark | 0.078 | 1 |
Finland | 0.075 | 2 |
France | 0.88 | 1 |
Germany | 1.12 | 15 |
Italy | 0.83 | 3 |
Japan | 1.74 | 5 |
New Zealand | 0.0683 | 2 |
Singapore | 0.076 | 1 |
South Africa | 0.75 | 2 |
Sweden | 0.13 | 1 |
Switzerland | 0.11 | 4 |
Thailand | 0.89 | 1 |
The Netherlands | 0.233 | 3 |
Trinidad and Tobago | 0.018 | 1 |
United Kingdom | 0.89 | 18 |
United States of America | 4.42 | 91 |
Unknown | – | 11 |
Total | 198 |
We started our investigation by discussing different approaches to define neuroethics. To shed new light on these stipulative top-down approaches we aimed at complementing this discussion by an in-depth analysis of the recent development and the current landscape of neuroethics literature. Our empirical, bottom-up analysis revealed not only a constant quantitative increase in the neuroethics literature over the years but also a constant relative size of the different subject categories. Hence, our study confirmed for neuroethics as a whole what has been reported by previous studies with a more limited focus (Seixas and Ayres Basto,
The increase of publications under the label of neuroethics correlates with a general increase of publications in neuroscience (Christen,
Because publications about the neuroscience of decision-making in moral contexts (i.e., the subject-categories of
Our corpus of literature instead developed as an open access multimodal compilation of neuroethics literature and from the beginning, and did not restrict possible entries according to the criterion of appearance in one of a restricted number of journals (see Method Section). Because of this, we consider our description of the field more accurate and more reliable. Our findings, however, do not contradict those of Christen (
A more in-depth quantitative analysis further revealed that some topics in neuroethics that are often considered to be central to the field, are indeed not as important. Even though in general the number of publications in the ethics of neuroscience branch of neuroethics has predominated the number of publications of the neuroscience of moral decision-making branch, within the first, issues such as cognitive enhancement or deep brain stimulation have shown to be quantitatively much less important than questions of clinical ethics with relation to psychiatric and neurodegenerative diseases. Much more than being predominantly a technology-driven research field, in the last 25 years neuroethics has been mainly driven by questions of medical and health-care ethics. Hence, contrary to the view communicated for example in the study of Gooray and Ferguson (
Even though only a middle-sized subject-category, our data shows that
Our analysis of the relation between different research topics addressed in neuroethics supports the view of neuroethics as a multifaceted endeavor in which different research foci do not necessarily interconnect. We found three tightly interconnected groups of thematically strongly interrelated subject-categories that can be seen as mapping the three perspectives on neuroethics suggested by Racine (cf. Figure
Particularly interesting with regard to the distinctions drawn by Racine, is our observation that the more theoretical approaches to neuroethics (i.e., neuroscience of decision-making in moral contexts) that represent the knowledge-driven approaches and that are grouped in the subject-categories of
Whether this gap will be overcome in the future and whether neuroethics will follow the dynamics pointing at more knowledge-driven approaches to the field, cannot be inferred from our data. What we see, however, is that the disciplinary background of current neuroethics researchers speaks in favor of the hypothesis that many trained empirical researchers engage with normative questions of applied ethics. A large amount of publications in neuroethics are produced by researchers with an academic background in neuroscience, psychology and medicine. In comparison, researchers with a background in philosophy, bioethics and the social sciences are marginal in the field. This finding fits with the data communicated by Gooray and Ferguson (
Another interpretation, however, would not be as flattering for neuroethics. It suggests that our findings indicate that a large part of neuroethics is the result of the engagement of neuroscientists, psychologists and medical doctors with normative and conceptual issues, for which they have not sufficiently trained in their academic education. To better assess the plausibility of these hypotheses, it would be necessary to consider the contents of the publications. For one might object that many of the publications written by neuroscientists are not actually on the ethics of neuroscience but about the neuroscience of decision-making and reasoning in moral contexts. Given the number of neuroscientists in the field and the amount of papers in the database that deal with the neuroscience of decision-making and moral reasoning, this is, however, extremely unlikely. Only about 8% of all articles relate to questions of moral theory and neuroscience (i.e., the subject-category moral theory), whereas about 56% of all top researchers in neuroethics have a background in neuroscience, psychology and medicine. Hence, it is much more plausible to assume that many neuroscientists have actively contributed to research on the ethical implications of neurotechnology and neuroscience research. At least to a certain degree this implies that the ethics of neuroscience is profoundly shaped by the knowledge and research agendas of neuroscientists themselves. This need not necessarily be regarded as a problematic dilettantism of the neuroethics community, but may be an acknowledgment of the fact that ethical assessments of issues associated with neurotechnologies and clinical neurology structurally rely on knowledge of specialists in the field. However, seen from a more sociological angle, this finding also indicates that the accusation of a lack of distance between neuroscience and neuroethics cannot be easily rejected. This lack of distance, expressed by an uncritical fascination of neuroethicists with the promises of (future) neuroscience and -technology, might indeed become problematic, if the research agendas of neuroscientists started to define the extent of legitimate ethical objections (De Vries,
The critique of neuroethics that flows from such a purported lack of distance between neuroscience and neuroethics has two main points of reference. The first is the worry that neuroethics, instead of critically assessing and regulating the research and practice of neuroscience and neuro-medicine, might engage in generating unrealistic and overoptimistic expectations about the future of neuroscience research (Quednow,
Besides the development and current structure of the thematic landscape of neuroethics, we were also interested in the institutionalization of the field. Institutionalization is an important point of view to assess the demarcation of a research field from other areas of investigation. Besides defining a discipline by its specific topics and research questions an investigation into its research infrastructure provides additional information. Some important indicators suggests that neuroethics is currently on the way to establish itself as a discipline distinct from bioethics and from the neurosciences respectively. This is mirrored by the fact that a scientific society for the discipline has been founded in 2006 (the Neuroethics Society), which in 2011 aligned with other disciplinary groups and networks (International Neuroethics Network) to form the International Neuroethics Society. It is further reflected by the foundation of two journals (
It was particularly striking that subject-categories could be distinguished according to the biomedical science journal to SSH journal ratio of the published articles. Besides the subject-category
Despite the growing organization of neuroethics by scientific journals and research organization, the need for neuroethics research has not yet translated into a corresponding global research infrastructure. Our data indicate that neuroethics was formed primarily inside the established institutional structures of the biomedical sciences and bioethics and only to a much smaller extent in the traditional humanities and the social sciences. Regarding the fact that we found no significant mobility of researchers trained in one discipline toward positions in departments distinct from that discipline (except six philosophers, who worked for bioethics or neuroethics institutes instead of philosophy departments) researchers of neuroethics seem to be as conservative as the academic system with its marked-off disciplines with regard to their affiliations. Eventually the ongoing establishment of interdisciplinary research centers dedicated specifically to neuroethics—especially in the USA and Canada - might shift this disciplinary bias in the future toward a more active involvement of the social sciences and the humanities. This hypothesis is underpinned by our finding that researchers affiliated with research centers and departments of neuroethics mainly work in the US and Canada. However, our data also revealed that only a very small number of the neuroethics researchers with high publication output are currently affiliated with a special neuroethics research center (NRC). This may of course be due to the fact that most of the NRCs are interdisciplinary institutions themselves, so that many researchers do not have their primary affiliation with an NRC but with another institution (i.e., Neuroscience Laboratory, Clinic or Philosophy Department). This finding is supported by the fact that in our data-set there were three NRCs and two larger Research Units that hosted at least one of the top neuroethics researchers (
In the introductory section, we emphasized that every account of neuroethics is partly stipulative and partly descriptive. By using bibliometrical data, we have investigated the structure of the phenomena any account of neuroethics should refer to. This descriptive approach to neuroethics could not decide whether a technology-driven, a health care-driven or a knowledge-driven approach to neuroethics is most adequate. However, by providing an empirical description of the development of neuroethics research in the past quarter century and about the mutual interrelations of topics sailing under the flag of neuroethics, it revealed the diversity and vast extent of the research field. On the other hand, any stipulative top-down approach to neuroethics can neither help drawing boundaries within the research field nor help to declare some issues as more central to neuroethics than others. Certain evaluations and interests, hence, guide any conceptualization of neuroethics. With the diversity of research topics in neuroethics in mind, we showed that many of the current conceptualizations actually relate to what in the scientific literature is understood as neuroethics. Nonetheless, there are conceptualizations of neuroethics that are more inclusive than others. Hence, from our empirical point of view, a concept of neuroethics with a single focus on the neuroscience of decision-making processes in moral contexts would be overly narrow, because quantitatively speaking the neuroethics literature on this area of research is comparatively small. But also a concept of neuroethics, which in the best tradition of a technically induced bioethics would focus mainly on the ethical, legal and social consequences of neurotechnological innovations, would still draw the boundaries of the field much too tight. Even though the quantity of neuroethics research addressing the conditions and consequences of technology implementation is much larger than that of the literature of neuroscience of moral decision-making, it would be somewhat arbitrary to ignore the close connections between the ethical implications of the use of neurotechnologies and questions of clinical neuroethics. Many technical innovations in neuroscience do not only have a direct impact in research contexts but also in clinical settings and in everyday life. Furthermore, such technologies are often developed as medical devices and, hence, it would be difficult to completely ignore this context, when technology assessment is at stake.
Drawing on our data about the temporal development of different research topics, we must conclude that none of the mentioned approaches to neuroethics has yet become dominant. Instead they are not mutually exclusive and sometimes even complement each other. It is in fact true that our data analysis revealed a significant gap between empirical work in neuroethics that addresses the neural basis of decision-making in morally relevant contexts and the questions addressed by applied neuroethics. However, it remains an open question whether neuroethics should work on overcoming this gap and if this is possible at all. It has been argued for both alternatives in the past, and it is likely that this dispute will continue, because our analysis does not underpin the hypothesis that neuroethics will enter a state of transition in the coming years. Rather neuroethics currently is a notably stable field of research that has not significantly changed in the number of publications per year since 2009 and which displays a very robust continuity of research questions and relevant topics. The fact that almost all topics discussed within neuroethics today have been present in debates in medical ethics, pharmacology or philosophy of mind shows this particularly clear. Currently neuroethics as a discipline amounts to bringing together these primarily distinct topics under one label, but not necessarily under one unified research agenda.
From this point of view, it is difficult to assess the potential of neuroethics to establish itself as a discipline of its own distinct form the neighboring research fields in neuroscience, bioethics and philosophy of mind. Even though neuroethics has taken some important steps in this direction on the institutional level (scientific journals, scientific communities, a few research centers), the broad distribution of neuroethics research over a very large corpus of journals and the rather homogenous academic background of successful neuroethics investigators in biomedicine and neuroscience speak against this hypothesis. Suiting this our data indicate that the entanglement of neuroethics with neuroscience is still rather strong. Whether neuroethics will become more independent in the future should be monitored by further studies addressing literature from a much more recent timespan. A regular monitoring of the field would not only help to understand its development but also to gain adequate and actual descriptions of the field that are sufficiently informed by the factual phenomenon of ongoing neuroethics research.
It could be argued that our study is limited in two important regards. First, the quality of the database limits the validity of our results, and second, one might object that our analysis of topics and networks in neuroethics crucially depends on the reliability of our system of subject-categories.
We believe, we can meet the first objection with reference to the fact that neuroethicists themselves know best, what their research is about. The MND was built as a resource for neuro-ethcists by neuroethicists. Hence, the versatile collection of material making up the MND should be more representative of neuroethics than the content of other, alternative databases. Second, this open, participatory structure prohibited that the regular scans of new publications turn out all too subjective.
Concerning the second objection, we are well aware that the differing number of defining keywords per subject-category might raise doubts about the system's reliability. Prima facie the probability of a paper to be tagged in a subject-category defined by more than 20 key-words can be expected to be much higher than the probability of a paper to be tagged in a subject-category defined by much fewer keywords. Hence, the differing length of keyword lists in our approach might undermine the reliability of the applied subject-categories. This objection, however, can be overruled by the observation that imposing a balance in the lists of keywords leads to a less specific matching of papers and subject-categories than in the lists we used. Choosing and approach allowing (a) for a constant number of 16 keywords per subject-category, (b) for tagging of a paper only when at least two terms within their headlines, abstracts and keywords matched with the keywords defining a category, and (c) for automatic labeling of papers to belong to a category only, if terms corresponding to the topic's label were present in the text, produces a much less reliable categorization. This alternative procedure not only leads to an increase in papers categorized in topics that were defined by more keywords than in the original procedure (for example the category “Addiction,” which was defined by nine keywords and relates to 55 paper in our approach, relates to 92 papers in the alternative approach, for further details c.f. Supplement
JL and CL developed the system of subject-categories with help of EH's expertise. CL organized the dataset according to the defined subject categories. JL and EH did the descriptive statistics on the data and drafted the manuscript. JL wrote the final version. All authors carefully read the manuscript and approved of its publication.
The authors were supported by an Open Research Area grant for the NESSHI Project, with funding from the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) (CL) and the Deutsche Forschungsgemeinschaft (DFG) (JL and EH, project code: HI 1328/2-1).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors are much obliged to Kelly Laas (Chicago) for help with language editing of the manuscript and to Hannes Boelsen (Mainz) for his proofreading of parts of the data analysis. Moreover, the authors would like to thank their co-participants in the NESSHI project for stimulating discussions on the relations between neuro- and social sciences.
The Supplementary Material for this article can be found online at:
1