“Have you ever considered opening up? No worries, I am talking about opening your scientific activities”, says Vanessa Teckentrup. She is a PhD student at the department of Psychiatry and Psychotherapy at the University Hospital in Tübingen who wholeheartedly enjoys bringing open science to life.
In workshops held for other young scientists and hours of discussions, she invites everyone to reflect on the degree of openness of individual research habits and offers her knowledge about the tools and platforms available. This path eventually led to the successful release of the game app Influenca which constitutes an ambitious but exemplary project in the realm of open science.
Open science has become a popular term to describe the endeavors of making the achievements of research widely accessible and the scientific process transparent. Although these seem to be obvious goals for everyone working on the great mysteries of nature, the scientific process is not open by default. Moreover, as science has been repeatedly exposed to harsh criticism and credibility concerns, it becomes important to reflect on the pros and cons of opening up the process of the scientific work. Illustratively, Brian Nosek and the Open Science Collaboration (2015) published a large-scale replication of 100 studies from the field of social psychology, and could only replicate a third of the reported effects. By now, the term ‘replication crisis’ has become well known in many fields of science, and the community fervently discusses the probable reasons for these disturbing results.
In general, the striking disagreement in published results might partly be rooted in small samples and single observations. Moreover, experiment or analysis scripts are mostly written on demand and usually not published along with the results. Sharing code is one concrete example of open science and it would facilitate the proof of validity of paradigms and analyses by other researchers. Besides that, pooled or re-used data would enhance collaborative work across groups. Especially in clinical research, the work is built on patient samples, sometimes targeting rare disorders and challenging recruiting procedures. Here, data sharing is essential to economically test the reliability of reported effects. This, in turn, could effectively advance the knowledge about the development of certain mental disorders.
Admittedly, being open might not always be the best approach for all aspects of your work. For example, making data publicly available bears the risk of personal identification of a patient. While those concerns should be taken very seriously, there are, however, manifold ways of making an informed decision regarding which aspects of the scientific work can be made open. By these means one can contribute to reliable and valuable research that can be actively acknowledged by the community.
Being asked about possible ways to learn more about open science, Vanessa suggests to follow the vivid discussions on social media (e.g. twitter), or to find inspiration from attempts to open up science in your field on conferences or workshops. “But most important: ask around and discuss ideas with your colleagues”, she says. Simply start with people who work next door.
Follow-up questions to reflect on:
1. What is the scientific process like in your lab?
2. Where do you see potential to open up your work?
3. Where do you anticipate potential risks and what could be possible solutions?
4. How can you protect sensitive personal data?
If you want to learn more about open science or discuss your intentions, drop us an email: firstname.lastname@example.org
The Wikimedia fellow program
In 2017 Vanessa obtained a designated open science fellowship by the Wikimedia foundation, including micro-funding to open her own project work, along with ample opportunities for workshops and mentoring. Together with 19 other young researchers from different scientific backgrounds, she developed her project and shared ideas about making science open. The app Influenca is the result of more than a year of dedication.
Influenca: a case study on open science
Influenca is an online game based on a simple reinforcement learning paradigm (adapted from Behrens et al., 2007). Covered by a gamified story, players take on the role of a scientist, trying to find the most effective medication against deadly viruses that threaten humanity. They repeatedly choose between two options and receive feedback about the success in healing due to their choice. By trial and error, they learn which option is currently the most promising one to contain the epidemic. Furthermore, each of the 30 runs is preceded by questions about current mood and eating behavior. That way we will be able to relate the decisions taken in the game to momentary bodily states.
The app is open, as it was written in Haxe, an open-source programming language designed to compile to a variety of operating systems. The whole source code is uploaded and updated regularly on Vanessa’s GitHub page (check QR-code in the figure). On GitHub, everyone can download and re-use the code. Moreover, with a little Haxe knowledge, researchers can make adaptations to the code and adjust the app’s functions to another research objective. Crucially, only fully pseudonymized data will be made openly available to other researchers. By providing installer files on our lab homepage and the Google Play Store, we facilitate accessibility of Influenca, fostering participation of many people with different habits and personalities. Thereby, we will be able to collect a large number of choices in the game very economically and hence provide a more nuanced picture of decision-making behavior based on reinforcement learning.
“With Influenca I want to give a hands-on example of an outright open science project and facilitate the re-use of all the resources”, Vanessa summarizes the purpose of her work. “If people only start to share and appreciate shared resources for their research, they will eventually acknowledge the potential of collaborative efforts to complete the mosaic of the mind.”
The app will be the first data collection step in a study conducted by the neuroMADLAB and funded by the Else Kröner-Fresenius Stiftung, targeting the association between reward sensitivity and eating behavior, especially in extreme manifestations (“binges”). But we are very curious about where the viruses will attack next.
Behrens, T. E., Woolrich, M. W., Walton, M. E., & Rushworth, M. F. (2007). Learning the value of information in an uncertain world. Nature neuroscience, 10(9), 1214.
Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.
Vanessa Teckentrup (left) and Monja Neuser (right) are both PhD students of the GTC and work in the IZKF junior research group, Neuroscience of Motivation, Action, and Desire Laboratory (neuroMADLAB) headed by Dr. Nils Kroemer