香京julia种子在线播放

In the last decade, the need to ensure reproducibility of research results and to justify public investment in research has led to increased sharing of research data and the imperative for open sharing. Open data platforms supported by research projects have increasingly become the center piece for facilitating open sharing of research data. In neuroscience, a number of these open platforms exist and share big, multitype and multifunctional data from diverse species of organisms for both research and innovation. As Poldrack and Gorgolewski (2014) pointed out, these platforms not only encourage data re-use and increase statistical power to stimulate translational knowledge but also expand the reach and impact of neuroscience research. A critical implication of this data re-use expansion is that data increasingly interacts with different jurisdictions with different regulatory requirements. While most of these datasets are from human participants, many are generated from animals.

Unlike human data that is nationally or regionally regulated (e.g., EU General Data Protection Regulation¹), there are no established legal frameworks that govern the sharing and re-use of animal data nationally or internationally. One reason for this is that the sharing or use of animal data does not raise the traditional data use concerns associated with human data (Stahl et al., 2019) such as; privacy, fairness, human rights and security. The ethical and legal issues around animal data are usually raised during data creation. The scientific, ethical and legal validity of animal research data are mostly determined by the nature of the research procedure/experiment. The moral and legal questions of animal experiments do not always revolve around the implications of animal data usage but on the ethical and legal permissibility of its scientific generation. Crucially, the regulatory and ethical principles that shape animal data generation in many countries and regions are different while the generated data are shared beyond boundaries without any governance mechanism. This means that animal data generated in less restrictive places are openly shared in countries with very restrictive requirements mostly through open data platforms. Thus, this paper asks the question: is it time to consider animal data governance?

The paper shows conceptually and empirically that there is a need for animal data governance that is informed by ethical concerns. Animal data raises different ethical concerns from human data and thus needs to be treated differently. There is a plurality of ethical perspectives on the use of animals in research which any data governance regime will need to take into account. This article therefore arrives at the conclusion that a procedural approach to data governance is called for that does not prescribe a particular ethical position but allows for a quick understanding of ethical concerns and debate about how different positions differ to facilitate cross-cultural and international collaboration.

We organize this paper as follows. We explore the current international data governance ecosystem for responsible biomedical research and innovation; the continued use of animals for neuroscience research, especially non-human primates. We then provide a thematic analysis of interviews conducted with international neuroscientists who conduct animal experiments. On this basis we arrive at the conclusion that a procedural approach to international data governance is called for.

Despite the increasing requirements to implement the 3Rs (replacement, refinement, and reduction) (Russell and Burch, 1960; Guhad, 2005), the use of animals in research continues in many parts of the world, especially in neuroscience research (perhaps more than in any other field of biomedical research) (Jones, 2021). Although public interest in the use of animals for research has significantly reduced the number of animal experiments in Europe and North America (Lankau et al., 2014; European Commission, 2019), animals continue to be central to scientific research in other parts of the world. The rationale for this is varied. Neuroscience research often involves invasive and non-invasive methods that cannot be conducted with humans because of associated risks and ethical concerns. Thus, neuroscientists turn to other animals such as; rodents, ferrets, dogs, pigs, zebra fish and monkeys whose usage in scientific research comes with considerably fewer ethical concerns. The argument for this, borders on the lack of safe and non-invasive approaches to studying the human brain (Preuss, 2010). Rodents are widely used because of their short gestation periods, their cost-effective production and have proved unquestionably effective (Neuhaus, 2018). Fruit flies (Zweier et al., 2009) and zebrafish (Haesemeyer and Schier, 2015) have also proved to be important research resources for neuroscience research.

The use of animals in invasive and intrusive neuroscience research is generally informed by the legal and ethical restrictions of such research with human brains. However, many neuroscientists suggest that there are anatomical and genetic limitations/differences that hamper the scope of their use (Garner, 2014; Windisch, 2014). These animal subjects or models present challenges Kaiser and Feng (2015) referred to as the lack of face validity and predictive validity. The delays in the development of new interventions for brain diseases are associated with the differences in pathophysiological mechanisms between rodents and humans (Ting and Feng, 2013). Differences in brain functions and cognitive behavior, difficulties in scalability of dosage regimens, differences in recovery times and differences in the ratio of white to gray matter in the brain are some of the issues that inform this lack of transnationality (Varki, 2000; Weatheall, 2015). For better predictive validity, some neuroscientists assert that non-human primates (NHPs) are particularly better subjects.

Data governance is defined as “the principles, procedures, frameworks, and policies that ensure acceptable and responsible processing of data at each stage of the data life cycle, from collection, storage, processing, curation, sharing, and use to deletion” (Eke D. O. et al., 2022). The emphasis on the data life cycle demonstrates that data governance is not only required for a specific stage of the life-cycle. It is a robust framework that starts before data collection and continues to the deletion stage. Fothergill et al. (2019) described it as the overall management of the availability, usability, integrity, quality, and security of data in order to ensure that the potential of the data is maximized while regulatory and ethical compliance is achieved within a specific organizational context. This definition introduces ethical compliance as an important aspect of data governance. Data governance is therefore more than legal compliance (Eke D. et al., 2022). It includes adherence to available ethical principles.

Furthermore, whereas the interpretations of data governance in organizations, disciplines and projects are different (Stahl et al., 2018), its goals and objectives are rooted in available laws and ethics. The question of whose laws and ethical values is determined by the context. Available regulations and ethical values are still jurisdictionally constrained while data continues to cross borders and socio-cultural contexts. For instance, data protection laws are established for specific jurisdictions (e.g., EU GDPR, USA HIPAA², Canada’s PIPEDA³ etc.). Ethical values and principles that shape data governance also emerge from specific socio-cultural backgrounds. This means that the meaning or interpretations of data ethics principles such as trust, autonomy, privacy and consent are different in different cultures and societies. These inform relative interpretations of data governance in different cultures.

It is also important to note that established data related regulations and ethical narratives focus mainly on human data. The literature and practice of data ethics and data protection exists to address issues that affect humans in the data processing pipelines. To the best of our knowledge, there is no existing regulation established to address ethical, legal and societal issues related to animal data. This is true for both research and non-research settings. In their systematic literature review of ethical principles that shape data governance discourse in neuroscience, (Ochang et al., 2022) identified a number of ethical principles that often shape data governance discourse in brain research. None of the principles identified touched on animal data concerns. That means that data governance mechanisms often exclude animal data concerns as it relates to ethics and the law. Aspects of technical elements of animal data governance are, however, often included in discussions on findable, accessible, interoperable, reusable (FAIR) data principles. These include aspects of data standardization, integration and interoperability. This falls short of addressing ethical and legal concerns related to the different stages of animal data lifecycle while animal experiments continue to be critical parts of biomedical research.

For a number of reasons, including socio-cultural, ethical and legal differences that inform what is considered permissible use of animals in research, there is a growing interest in the governance of animal data (Eke D. O. et al., 2022). Perceptions on ethical concerns about the use of animals, particularly NHPs, in biomedical research are fundamentally shaped by diverse socio-cultural norms, ethical principles and regulatory requirements. This is evident in the fact that while the use of NHPs in research has decreased significantly in Europe and America due to increased social animal welfare activism, NHPs are still the central focus of big neuroscience projects in Asia (Okano et al., 2016; Poo et al., 2016).

Advancements in genome-editing technologies such as the CRISPR/Cas system and artificial intelligence amplifies the huge possibilities in generating genetically modified NHPs. Although this emerging field (genetically modified NHP) has the “potential to transform the study of higher brain function and dramatically facilitate the development of effective treatment for human brain disorders” (Feng et al., 2020), it raises significant ethical concerns. Responding to the huge potential of genetically modified NHPs, Rommelfanger et al. (2018) raise the concern that researchers may be able to introduce cognitive capabilities that can contribute to blurring the boundaries of personhood and ultimately alter traditional perceptions of animal ethics. In 2019, a group of researchers from the Kunming Institute of Zoology in China claimed to have created transgenic monkeys with improved cognitive capacity (Shi et al., 2019). These modified monkeys were created with human MCPH1 genes and were not modeling any human diseases. They were simply modified to be phenotypically humanlike. During a series of cognitive tests, the researchers reported that these animals displayed better short-term memory than their counterparts in the wild. Basically, their brain development mirrored human brain development in many respects. The underlying logic behind this research, which is to manipulate monkeys to model humanlike capacities, presents a slippery slope concern. The question is where would the line be drawn in the path to generate human-like animals? To say the least, such research will not be permissible in many socio-cultural, ethical and legal contexts where the moral status of NHPs are hotly debated.

Given these differences in attitudes, values, principles, beliefs, regulations on the use of animals in research, Rommelfanger et al. (2018) further noted that, “sharing of brain data between countries that hold different ethical stances on what is considered appropriate animal experimentation raises additional questions.” One of such questions is; “Should a country accept or use data collected elsewhere in a fashion that is not considered locally ethical?” (Ibid). This is a critical question at the heart of animal data governance consideration. It presents an ethical dilemma many scientists currently face in the neuroscience data sharing ecosystem (Eke D. O. et al., 2022). One project that has stated this as a concern is the PRIMatE Data Exchange (PRIME-DE) Consortium that has highlighted the lack of international standards and regulations as barriers to fostering international NHP data sharing and collaborations (Milham et al., 2020). This increased interest in neuroscience deserves more exploration and hence this paper.

The issue of responsible animal data governance requires multi-stakeholder perspectives and insights (Rommelfanger et al., 2018; Eke D. et al., 2022). It calls for the appreciation of diverse cultural values and beliefs while respecting established ethical frameworks. Thus, a semi-structured interview was selected as the methodological choice. The underlying research philosophy, therefore, is to use social actors to provide in-depth and rich perspectives on the reality of animal data governance. The aim was to provide detailed and reasoned insights rather than objectively generalisable positions.

The target population included researchers around the globe who have conducted or are conducting animal experiments to answer diverse neuroscience questions. Participants were drawn from active research projects under the International Brain Initiative (IBI). The IBI is the umbrella body for all the large-scale brain research initiatives including the EU Human Brain Project, the US Brain Initiative, Japan Brain/MINDS, Australian Brain Alliance, Korean Brain Initiative, Canadian Brain Research Strategy, and China Brain Project. We also drew participants from Africa and Latin America. Initial list of 37 researchers was compiled and interview invitations extended to all of them. A total of 15 people accepted the invitation and 3 later withdrew citing busy schedules. A structured interview guide that aligns well with the principles of qualitative research design (Ragin and Amoroso, 2011) was developed and tested on two colleagues. Following feedback from these initial tests, the interview protocol was improved before the start of the interviews.

Ethics approval was obtained for this research from De Montfort University ethics review committee. Information sheet and an informed consent form were then emailed to participants. The information sheet contained comprehensive data on the research including but not limited to the research objectives, expectations from the participants and responsibilities of researchers, potential risks and benefits, the voluntariness of participation and how research data will be used. All participants returned the consent form before the interviews and further verbal consent was sought at the start of the interview to record the session. The interviews were conducted either in person or virtually via Skype or zoom and took approximately 40 min each to complete. These interviews occurred between January 2020 and November 2021. Overall, 13 interviews with 12 participants (10 males and 2 females) were conducted. One participant was interviewed twice because the first interview was cut short due to technical difficulties. A total of 13 interviews were considered sufficient to achieve saturation because according to Guest et al. (2006), theoretical saturation can be achieved even in 12 interviews and basic elements for metathemes can emerge as early as six interviews.

The interviews were transcribed and stored in NviVo 12 where they were inductively coded for themes by the first author (DE). The inductive coding process involved reading through the transcript and identifying common patterns and themes. Thematically, the coding tree included high level nodes that show participants’ perspectives on why it is time to consider animal data governance or otherwise (such as cultural differences, legal differences etc.). Specific themes that align with high level themes are coded as sub-nodes. For example, ethics dumping was identified as a sub-node under the high-level node of ethical differences.

One of the key results that emerged from the interviews was that all the participants agreed that due to the increasing transfer of animal data across countries, it is time to consider animal data governance for a number of reasons. The reasons provided by the participants are diverse and include: differences in regulations, socio-cultural and ethical values and beliefs, as well as scientific quality (see Table 1).

The findings from our empirical work are broadly consistent with our insights from the existing literature. The international neuroscientists we talked to agreed that animal neuroscience research raises ethical concerns and that these concerns have consequences for the way the resulting data can and should be used. The very brief answer to the question in the title to this article whether it is time to consider animal data in data governance is thus a “yes.”

Our empirical work highlights that this is not a purely theoretical problem but that questions of animal data governance do arise in practical neuroscience work. The global discussion of data governance in health-related research ostensibly has the purpose of facilitating collaboration and exchange of data with a view to support the creation of new knowledge and the resulting consequences. This logic can be extended to animal data which calls for animal data governance.

There remain, however, fundamental differences between human data and animal data. The very use of human data can raise ethical concerns, for example where patient record privacy is concerned. Animal data does not raise such intrinsic concerns. For animal data the core of most issues is the generation or collection of data. Data use is nevertheless important because a lack of attention to the use of data may facilitate the use of data which is deemed not to be ethically permissible in a particular jurisdiction or cultural context, thereby sidestepping agreed-upon ethical principles.

A key question is therefore which type of animal research is deemed to be permissible and on what grounds can such value judgments be made. Our interviews showed that neuroscientists are aware of differences with regards to these questions. There seems to be a continuum of ethical severity which starts with cell cultures, moves up via invertebrates, vertebrates, rodents, NHPs and may find its current summit in research on transgenic NHPs. The problem is that the evaluation of these different types of research differ between cultures and jurisdictions and there is no universally agreed position on these questions.

This lack of agreement points to the lively exchange of ideas between cultures which is probably a good thing. Ethics is a topic that often finds its expressions in dilemmas and disagreements, so the plurality of views is not surprising. It seems plausible that an ethical free-for-all is not desirable, neither in animal research, nor in research more generally. At the same time, a uniform ethical position would suppress legitimate positions and thus be likely unethical in itself. Ethical plurality is thus welcome and can stimulate academic debate, for example in the field of neuroethics, where questions are continually triggered by new technologies and methods. While we thus welcome ethical plurality on animal research, it raises practical questions with regards to what data can and should be used for which purposes.

This leaves us with the question of how ethical pluralism can be accommodated in data governance. One plausible response to this question could come from procedural approaches to ethics. What this means is that we should not expect to find agreement on the substance of complex ethical questions, but it may be possible to define procedures that support fruitful engagement on these questions. One can argue that most modern western approaches to philosophical ethics follow such a procedural approach. Elsewhere we have suggested that Discourse Ethics may provide a suitable avenue to pursue (Stahl et al., 2019). We believe, without having the space to make this argument in detail, that such a procedural approach could be applied to the problem of animal data governance.

In practice this would mean that data governance should be designed in a way to facilitate constructive debate about ethical issues and be suitable for supporting ethical consensus, where it exists. This implies that data governance should be used to highlight ethically contentious aspects of animal data. This means that the meta-data of neuroscientific animal research data should clearly show those aspects that we know to be ethically contentious. This would include the species, research question, type of intervention, whether transgenic animals are involved etc.

The result of such an approach to animal data would be that it would be easy to understand ethical agreements and disagreements. If, for example, a culture has a consensus view that in vivo experiments with NHPs or the use of transgenic animals is not ethically acceptable, then filtering out such data would be easy. Probably more importantly, a strong metadata schema would allow highlighting which aspects of the research and the resulting data may be contentious and thus foster communication around the reasons for disagreements and possible ways to shape research and data in ways that are more broadly deemed to be acceptable.

This proposal is of course not overwhelmingly novel. Data governance structures already routinely capture some of these items. As our interviews have shown, researchers see data quality as an (ethical) issue and good metadata is recognized as a means to increase transparency of data and to promote the FAIR principles. The novelty of our proposal is that ethics-related aspects of the metadata could be explicitly defined and collected. While many of these will already form part of data governance, a next step would be to more clearly define them in ways that support researchers who generate the scientific data in the first place and ensure that they are aware of relevant metadata requirements.

This proposal is of course no panacea. There are limitations of our research such as the limited number of respondents and a lack of statistical representativeness of our approach. While we believe that our methodological choices have ensured that we received relevant input, we cannot claim to have represented all possible angles and identified all ethical issues. This article should thus be read as an exploratory study that has confirmed that ethically informed animal data governance is called for. The real work of shaping such a data governance approach will have to follow as a large-scale consultative exercise leading to the co-creation of animal data governance that truly captures ethical issues. This future research can include how these findings individually can or do shape animal data governance.

It is furthermore worth underlining that the existence of ethically sensitive data governance will not make the underlying ethical issues go away. Many of these issues touch on deep convictions of who we are and what we as humans can or should do. Such convictions do not change quickly and different views will remain. But, as indicated earlier, ethical pluralism does not need to be seen as a problem but can be celebrated as an expression of human diversity. What counts is that we find productive ways of dealing with issues. Ethically informed animal data governance can be one mechanism that allows us as researchers, as citizens of different countries, holders of different convictions to come together and have productive conversations about how to understand and deal with our different worldviews. If it achieves this, then this will not only strengthen neuroscience with all its concomitant benefits but also show how science can play an important role in tackling the broader ethical and social questions that our increasingly globalized world faces.

The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Ethics approval was obtained for this research from De Montfort University Ethics Review Committee. Information sheet and an informed consent form were then emailed to participants.

DE: conceptualization, writing, data analysis, original draft/review, and editing. GO: conceptualization, writing, and data collection. WK: conceptualization and writing. BS: conceptualization, writing, and funding. All authors contributed to the article and approved the submitted version.