By April Cashin-Garbutt

Research culture is changing. From the increased sharing of data and tools, to growing international collaborations in neuroscience. It is clear that solving some of the most fundamental scientific questions of our era will require shifts in research culture. But what are some of the key changes that have taken place so far and what more needs to be done? We asked SWC seminar speakers to share their thoughts – here’s what they had to say.

Move towards reproducibility, sharing data and team science

“My research is on the border of neuroscience, psychology and psychiatry and right now, we’re seeing – coming largely from psychology – a big revolution in reproducibility. This means making sure the claims made in papers are actually meaningful and not just spurious due to statistical artefacts or questionable research practices. I’ve been watching that happen for about ten years. It’s been a lot faster in psychology, but it’s happening in neuroscience more as well. That’s really good – it’s helping the field focus on what results are important and true, and reject things that are flashy but don’t pan out. 

I don’t know why it’s faster in psychology than in neuroscience. I think that psychology had several high profile cases and entire research programmes that were completely dependent on spurious effects. I don’t think psychology has these weaknesses and neuroscience doesn’t. There are similar weaknesses in neuroscience too, but a lot of times it takes more specialists to analyse and understand the data because it’s more complex. It is good that this issue of reproducibility is starting to come into neuroscience and psychiatry as well.” Dr Ben Hayden, Professor of Neuroscience at University of Minnesota

“I think within neuroscience specifically there has been this limitation in people’s ability to share their data and draw direct comparisons between the findings of different labs. Everyone performs experiments in their own setups and saves their data in their own specific format and there’s no real standards in the field for how a behaviour experiment should be done or how neural recording data should be shared. There are some standards but they are not widely adopted.

I think moving towards reproducibility, sharing data and team science will make a big difference in research culture and I hope in the future we see more of an effort to allow more direct comparison of results between labs and for the sharing of findings and publishing of datasets and code that people used to analyse the datasets. 

This is something that is much more widespread in some fields, such as the RNA-seq or computational biology, where there are standard toolboxes that people use and there are practices for releasing your data and repositories for sharing. Those kinds of cultures have lagged behind a bit in neuroscience. I think this is because a lot of the experiments are so technically challenging and the data takes so long to collect. I hope in the future, there will be more of a movement towards making it easier to share data and making it easier to share code and findings in a way that can be compared more directly between labs.

There have been quite a lot of changes in the past few years in terms of how people communicate their results. Most people now post their data as a preprint before publishing and they also share on Twitter. We are going through a period of dramatic change and hopefully that will lead to more transparent science and improve the kinds of studies people can do moving forward.” Dr Ann Kennedy, Assistant Professor, Northwestern University

Invest in making science open

“I’m a strong believer in open science. One of the things that is currently happening, and I would like to see more in the future, is to invest in making our science open. This is actually much harder than people think: to make our data open and for them to be reusable, you have to think about what somebody who doesn’t understand the research whatsoever needs to know to use those data. That information is called metadata. Data about what the data are about, the context of the research, what they refer to, how they were collected etc. All critical information that a future self will need in order to make usage of those data in a way that goes beyond what the original study reported, or even the authors intention. It’s very difficult to think about this imaginary person! Making science open is really a paramount goal that we should embrace; and developing tools for incorporating metadata, and to index them will be an important challenge to meet in the future. 

Secondly, in some sense my dream lab would be a lab where collaborations take place between many different people, with diverse and complementary expertise, in which specialization is valued and there is less of a static head of a lab, and more of a community instead. Each and every person can bring something different. For a given project, someone will become a leader and will be leading that, but do you have to lead all the time?

Unfortunately, this is not at all how science is structured. I am hopeful that people will understand that if we want to make this sustainable, while also maintaining the highest standards of reproducibility, we need to distribute resources in a different way, so we can retain more people, while also allowing specialization, and integration on the backbone of a fair and reproducible science.

There’s also a really cool project that we have recently embarked on that makes use of open science but it goes beyond by pairing it with adversarial collaboration. In this project, we have asked two people with different theories about consciousness to come up with an experiment that would arbitrate between these two theories. It’s all open science – a huge team of people works on these experiments. We had a perspective piece in Science to show how cool it would be to do science this other way as a mechanism to resolve scientific discrepancies, and to enable theory building in the era of cumulative science.” Dr Lucia Melloni, Group Leader at the Max Planck Institute for Empirical Aesthetics in Frankfurt am Main, Germany

“Open science and collaboration is extremely important in making the system such that everybody tries to figure out the truth in understanding the brain in the best way rather than being driven by competition and milestones for one’s career.

Open source software and publications are already being advanced a lot right now. I think the more complex the things that we study, the more we will require collaborations. We need to do more than just put data and software out there, as the complexity means we need more background information from the researchers who collected the data. This feels like a missing link in a lot of open data at the moment. We need to focus on standardisation across labs so everybody who uses open source data and software knows exactly what it is doing and how the data was collected.

I would also like to see closer collaboration between basic scientists, clinicians and industry as at the moment there are huge gaps in between. If the gaps became closer, it would be extremely helpful to streamline the process from basic findings to the development of treatments that are relevant for patients.” Dr Sarah Melzer, Postdoctoral Fellow, Harvard University

Recognise and nurture team science

“There has been a change over the last ten years, particularly in systems neuroscience, as to address complex questions you need lots of different types of expertise. People have started to collaborate more and team up in larger groups to tackle behavioural problems and divide them into smaller pieces that can be solved by different research teams. I think this is the way forward as it means progress can happen at a much quicker scale compared to an individual lab trying to address all the different aspects on its own.” Dr Sarah Ruediger, Postdoctoral Researcher, University of California, San Francisco

“IBL targets a very specific aspect of research culture, which is that most research now is done in a single lab way, and sometimes a single investigator way, where each person in the lab has a separate project they’re working on. So that’s the aspect of research culture that IBL and a few other organisations as well have started to target. The Allen Institute for Brain Sciences was before IBL and set the stage for the idea that team science has a big role in neuroscience. Not that we should eliminate smaller, single lab approaches – we need those too – but there are some questions for which a large scale effort is the right way to go. 

I think IBL, along with the Allen Institute and a small number of team science groups, have really pushed that forward and we’ve learned a lot. Working as a team is really different from working individually, and sometimes challenging, but also really rewarding. It’s been amazing, especially to get to know the students and postdocs that are part of that collaboration. To be talking with people in other countries all the time, working together on common problems, has been really exciting and inspiring. 

One of the biggest challenges is credit assignment. Right now, the field is set up to reward individual efforts and not so much team effort. I think that’s partly in the sense that search committees for jobs might mostly just count the number of first author papers rather than looking more deeply at someone’s contribution to a team effort, but I think it even goes beyond that. 

There’s often a misunderstanding that in science one person has a genius idea and suddenly a light bulb goes on and all the answers are clear, but really this isn’t actually how science happens at all. In a sense team science has always been there, it just hasn’t been recognised. It’s always the case that one person has bit of an idea and then someone else builds on it, and two people have coffee together and they’re like, ‘But is that really right? No, but what about this?’ And it grows through the collective efforts of many people. Maybe the real change that is happening now is that we’re starting to truly recognise that and to nurture that. But I think we are not yet fully rewarding that as a culture, and we have to. And other groups have: there are other large collaborative efforts in other areas of science even within biology that have been successful. I think neuroscience is lagging and needs to start recognising as well as rewarding team science.” Dr Anne Churchland, Professor in Neurobiology at UCLA

Foster collaborations and increase diversity

“There are already some initiatives to improve research culture and I feel that it is constantly improving and has improved over the past few years. One important aspect is collaboration. Compared to when I started research, there are now much more incentives to collaborate and work on big projects together instead of competing with each other. That should continue to be fostered in the future to enable collaboration. 

Another important part is increasing diversity – not only is it a matter of justice, but we also urgently need the most talented people bringing in their diverse perspectives to tackle challenging problems in science. We need to create environments where we can attract and keep all the available talent regardless of their background.” Dr Katharina Schmack, Group Leader at Francis Crick Institute and University College London

“Science thrives when people who have different backgrounds and think differently are included and given a platform. Science slows down and sometimes stagnates when that doesn’t happen.

The more diversity, in every sense, that you bring into science the better, and it is much needed in our field for many reasons.” Dr Carl Schoonover and Dr Andrew Fink, Postdoctoral Fellows in the Axel Lab at Columbia University

Address biases and inequalities

“There are all sorts of inequalities in science –  there are huge biases and problems with the way that we conduct science, some of which are to do with the way we treat people and others to do with the way we treat ideas. One change is the introduction of double-blind reviewing, which I think is really important for getting better science. 

Science is really a web of trust and it is hard to shake this with our knowledge of people. For example, I tend to remember the contents of papers better when I know the face of the people writing the paper. This helps me associate the content and style of thinking with who is doing the research but that leads to all sorts of human biases. Once you become aware of the biases that you have, then you can try to compensate for that to some degree.

Particularly in theoretical science, there is a huge gender imbalance that arises from all sorts of mistreatment of women. This imbalance manifests in a certain culture that lionises certain types of thinkers and we need to be more open to different ways of giving credit and attributing value. I’m really pleased to see some of the ideas that people are proposing to improve this.” Dr Xaq Pitkow, Principal Investigator, Baylor College of Medicine and Rice University

Embrace the collective process

“One of the fundamental things we learned from our own research was that in small group discussions, those people who are socially dominant, or the boss, are often not the people with the greatest competence in the room for that given problem. We often remind ourselves of this in my team when we have discussions. We tend to start by giving everyone a few minutes at the beginning of a meeting to write down their first thoughts about the question or problem we are trying to solve. Then we begin the meeting by asking the more junior people, or people that are new to the team, to share their thoughts. We find this approach is very good for getting the discussion started and it is fascinating to listen to the ideas of different people. We really try to practice what we research, to make it a collective process.” Professor Jens Krause, Humboldt University, Berlin

Provide alternative career paths for researchers

“Research can be very high pressure because, by definition, you are working under a lot of uncertainty trying to comprehend things that we don’t yet understand. On top of this pressure, you add the pyramidal structure of research, as there are many fewer positions at the top in terms of running labs than at the entry point, which all conspires to create a lot of anxiety that has a large negative impact on mental health. 

I think research culture can be improved by more explicit efforts to provide alternative paths for researchers and not to make people feel like they’re a failure if they don’t get a Nature paper or a faculty position. Research is a training in critical thinking, evidence-based thinking and problem solving, which are very translatable skills. As a culture I think we need to start acknowledging that we can add value to society not just by the discoveries we make, or the successful scientists we train, but by endowing people with a way of thinking critically about the world in general and making people feel that is a respected skill in any domain.” Dr Joe Paton, Champalimaud Centre for the Unknown

Make practical changes to structures

“2020 was the year of change in so many aspects of society, science included, and problematic aspects of culture have pushed to the forefront. There are many things that can be changed, but a lot of them are tied to the structure of how science is done. And so in addition to talking about changing culture, I think we also need to think about practical ways we change how science is done in terms of hierarchy, funding, job security for trainees. For example, if we pay our trainees a living wage so they can support a family, this will also change the culture because it will change the people that are able to do science.” Dr Cindy Poo, Postdoctoral Researcher, Champalimaud Research

“There are many people in the neuroscience field now –but limited funding, which means it’s hard to get job security or support as young researchers. I think this problem will be solved with more and more people exiting to industry and other fields, and perhaps, more funding from granting bodies. The Wellcome Trust has a very good initiative now where they are extending funding for early- and mid-career researchers. I think that’s important: extending from a 5-year research programme, or even shorter in some cases, to an 8-year programme gives more time to get things done. I haven’t seen the impact of this yet, but I think it would be greatly beneficial. 

I think that the other problem with more people in this field is that the standards for getting papers published are exceedingly high, above and beyond what I feel is actually reasonable. I think that greatly slows the progress of science. Everyone is trying to get into high-impact journals, there is limited space, and as a result the review process is allowed to be difficult for the sake of being difficult, rather than improving the science. So a study which has very good evidence in support of its conclusions would still need many more months of data collection before it actually gets published. I think that’s a waste of already scarce resources – things like this have to change. 

The eLife model of publication greatly improves this because their peer-review process is more open and collaborative. The reviewers can interact with one another and that greatly limits how unreasonable and demanding any one reviewer can be. The reviewing editor and journal as a whole have an ethos where they try to be reasonable and only ask for revisions that can be completed within a few months. Models like this are very important, especially in an underfunded field. Researchers, and ultimately funding bodies, should not have to spend more money and time trying to publish science.” Dr Zahid Padamsey, University of Edinburgh