The University of Groningen (UG) sees the development of open science as one of its priorities for the next 5 years. In this series of interviews, we explore the status of open science at the different UG faculties. This time: the Faculty of Medical Sciences.

Marian Joëls (MJ) is professor of Neurobiology of Environmental Factors, Dean of the Faculty of Medical Sciences and member of the UMCG Board of Directors. Her research revolves around stress studies in animals and humans. 

Marie-Jose van Tol (MJvT) is professor of Mood & Cognition and also Chair of The Young Academy. Her research revolves around neurocognitive factors that play a role in the vulnerability for enduring and recurring emotional disturbances.

Why is open science important in the medical sciences?

MJ: Open science is important in different stages of research within the medical sciences, such as open access publishing and data sharing. For instance, in the medical sciences, approximately 90% of all published laboratory animal studies are underpowered because they use too small a sample. This is a serious problem, putting the reproducibility of clinical findings at stake. These preclinical studies often form the basis for clinical trials. Sharing the underlying FAIR (Findable, Accessible, Interoperable, and Reusable) dataset would thus benefit the research reproducibility in the medical sciences tremendously, as it allows researchers to check each other’s data analysis and reuse data (e.g., in a meta-analysis). Citizen Science is also an important aspect of open science that could be utilised much more because getting input from patients can really improve medical research.

MJvT: I agree that, for any science - and specifically medical sciences - open access publishing and sharing FAIR research data is very important. I would like to mention, however, that we need to keep an eye on accessibility of publishing in open access journals, and prevent that this is only an option for researchers that can afford this. It’s fantastic that with open access publishing everyone can read scientific studies, but we need to prevent a bias in what gets published. Therefore, I like to mention the option of diamond open access publishing in this context as well. Diamond open access journals are free open access journals, which is important especially in low-income countries where they have less availability of money but should still be able to benefit from research that affects the entire world population. We saw this importance with COVID studies, for example. With regard to data sharing, in my field of expertise, Neuroscience, sharing data on activation maps of functional Magnetic Resonance Imaging (fMRI) is very common. This is important because it allows others to develop new methods and ways of analyses, as well as to see how robust the outcomes are. Pre-registration of a priori hypotheses also plays an important role, as it helps to prevent publication bias towards positive results in scientific publications. I think it is very important for researchers to become aware of this in the early stages of one’s career.

What is the current state of affairs in terms of open science at the Faculty of Medical Sciences (FMS)?

MJ: For approximately five years now, all our PhD candidates have taken a mandatory course in the very basics of data stewardship because we find it very important that the datasets that PhD candidates collect are findable and well annotated. We also established a helpdesk ‘Data Steward’ for all researchers, to ask questions and get assistance. Unique to the FMS is that we decided to extend the implementation of Taverne from only applying to articles by UMCG corresponding authors to all articles by UMCG authors. We now have a support service whereby researchers can forward their article acceptance email to the Central Medical Library (CMB) (actonacceptance umcg.nl) and receive tailored advice on the open access procedures to follow This has boosted our OA percentage to almost 95%.

MJvT: It is unfortunate that researchers are often afraid of doing something wrong when making data and publications open access and, as a result, do not do it at all. In addition, opening up your research often also comes with an increased sense of vulnerability. That is, many Principal Investigators (PIs) just don't know where to start with open science, or what to do. The FMS investment in the open access help desk and assistance from the CMB is thus extremely valuable.

Which objectives do you want to achieve in the upcoming two years with regard to the implementation (or further implementation) of open science?

MJ: One of the key success factors in implementing open science is changing the culture. I believe that, in order to change the research culture, researchers have to be convinced that the additional time investment in open science truly improves the quality of their work. Because opening up requires extra effort. We need to offer researchers examples of good open science practices and tools that help them to get there. The quantity of published papers should also be less important than their quality. We thus have to make it as easy as possible for researchers to open up their workflow and focus on changing their intrinsic motivation, so that researchers see open science no longer as a ‘memo from the Board that they need to comply with’ but rather as an investment in the quality of their own research.

MJvT: We also need to look at researchers’ availability of time, and incentives for doing research conform the open science principles. PhD students have limited time for their research and towards the end of their PhD they often have other priorities than making their datasets FAIR. Ideally, it should be a fundamental task of the department heads to be responsible for creating the right circumstances and for rewarding open science practices. In addition, if we want to change the research culture, I think creating more awareness of open science practices other than open access publishing is also very important. Currently, the Strategy Evaluation Protocol (SEP) - the protocol all faculties are evaluated on each year - only mentions open access publishing as an open science practice but I think we can define open science in a much broader way.

What obstacles and challenges do you expect in achieving these objectives, and how will you overcome them?
Prof. Joëls, since you will be retiring soon, what advice would you like to give your successor* in order to help overcome these challenges?

MJ: Culture change is an effortful process. For 20 years we instructed our researchers to publish as many publications in as high impact journals as they could, so it will be quite a challenge to change the current research culture towards a more open one. Just issuing a decree or memo will not change anything, we really need to approach this through the eyes of the researcher -and I consider myself as one of them- and try to see how they experience this (prospective) change. I suggest we go back to the content: focusing on quality instead of quantity in research, and make it fun. Researchers should be part of the new standard. A few good examples should help. When we as researchers are convinced that open science leads to improvement, we can bring that culture change about. I am convinced that my successor thinks likewise because he has been heavily involved in Health-RI, a national organisation that advocates these changes and enables FAIR data principles. We couldn’t have a more suitable new dean, from that perspective.

MJvT: An increased focus on craftsmanship can also help shifting the focus from quantity to quality. How someone conducts their research should be made important to a person's standing in science. Department heads enabling this should also be put in the spotlight, so that within units ‘prestige’ depends more on the rigour of your scientific output than the quantity. This way, we could also strengthen society's faith and trust in science.
In addition, I also foresee the continuation of ‘cold feet’ in sharing medical data as an obstacle when changing the research culture. I think we should aim for stepping away from the argument that data sharing is not permitted in the medical sciences because of the privacy of patients, though this is a serious issue. This is a typical example of ‘cold feet’, and this argument simply often does not hold. Besides, even in cases where it does hold, we can be as open as possible. Meta data and group averages can be published for multiple purposes and uses.

What progress has been made at UMCG in terms of recognition and rewards of open science practices since we spoke to Prof. Joëls in 2019, and how will the
FMS/UMCG move forward from here?

MJ: About 5 years ago we developed four different career paths in our organisation: one with a focus on (lab)research, the second on educational innovation, the third on societal relevance and the fourth with a focus on clinical innovation. All criteria are R&R proof. We thus have made steps in adjusting towards open science practices, but the open science mindset has not been fully adopted in our organisation yet. In proceeding from here, we as FMS/UMCG need to be careful not to fall back into the classic form of evaluations. That the UG will sign the Agreement on Reforming Research Assessment (RRA) soon is a great step, but I feel that international figureheads should now also speak up - the change should come from within the scientific community itself. If we can take the next step together with our American colleagues, I foresee a positive and necessary change.

MJvT: Indeed, room for different talents in career development is important. However, researchers often see themselves competing in an international arena, so aligning the recognition and rewards system for more differential career paths on an international level will also help tremendously to move forward in this respect.

To what extent do FMS/UMCG students learn about open science research
practices, and how is it ensured that they apply them in their bachelor and/or master thesis?

MJvT: Customs vary across departments within the UMCG, but the Interdisciplinary Center of Psychiatry and Emotion regulation (ICPE) is a frontrunner in applying open science practices. In my course, BCN (Behavioural and Cognitive Neurosciences) Research Master, several topics within the open science framework are mentioned. However, practices such as pre-registration of hypotheses and analysis plans are not yet part of the standard course material. Instead of a pre-registration, students write an elaborate analysis plan. I think it is a really good custom to write a thought-through analysis plan, even if you do not register it in an online repository. More importantly, I think it makes research more enjoyable. It makes the researcher proud of their work. Pre-registration and analysis plans are strong motivators for why researchers do what they do.

MJ: In the new Sector Plan Medical and Health Sciences we, as UMCG, have reserved money to include modules on Data Science and AI in all curricula in which we play a major role. This builds on the first steps we have already taken in that direction, developing courses for e.g., master students in Biomedical Sciences or a MOOC on AI that is very successful. But there is still a lot to do and this is high on our wish list for the coming 1-2 years.

*Per 1 February 2023, professor Wiro Niessen (Professor of Biomedical Image Analysis and Machine Learning at Erasmus Medical Center in Rotterdam, as well as holding a position at the Faculty of Applied Sciences of Delft University of Technology (TU Delft) ) will succeed Marian Joëls, who will say farewell to the UMCG to retire.