First Norwegian Neuroscience Data Sharing Workshop arranged in Oslo

Blog post by Sophia Pieschnik and Ingvild E. Bjerke, University of Oslo

On January 18-19 2024, the first Norwegian Neuroscience Data Sharing Workshop was held in Oslo. This two-day event was organized as a collaboration between the Norwegian EBRAINS Node (EBRAINS Norway), the Norwegian Neuroinformatics Node (INCF Norway) and the Neural Systems Laboratory, UiO. The workshop brought researchers representing diverse fields and topics together with researchers and data curation scientists from the EBRAINS national node to discuss challenges and opportunities in sharing their research data.

The first day of the workshop focused on sharing of human brain imaging data, with participation from researchers affiliated with the University of Oslo (UiO), the Oslo University Hospital (OUS) and the Norwegian University of Science and Technology (NTNU). They showcased their respective research and experiences with sharing and re-using data.

Torgeir Moberget (UiO / OUS), highlighted the pivotal role of large-scale datasets in facilitating the use of machine learning techniques and enhancing statistical robustness of analyses. Atle Bjørnerud (UIO / OUS) presented the SAILOR project, which comprises longitudinal MRI data from 27 patients with high-grade glioma and was successfully shared via EBRAINS. Bjørnerud emphasized the importance of explicit participant consent forms and the cooperation with hospitals to collect data on a larger scale.

Special guest and Data Protection Officer of EBRAINS, Stephanie Heyman (EBRAINS AISBL), weighed in on topics related to protection of human data. Heyman highlighted that researchers need to get educated in the General Data Protection Regulation (GDPR) and its applicability to sensitive data, such as human brain imaging, to uphold principles of legality, fairness, and transparency. She emphasized consent from the participants as a critical requirement for sharing brain imaging data.

Ample time was set aside for discussions, where there was a clear consensus among the researchers on the significance of sharing human brain imaging data. They further shared experiences with sharing data with collaborators, facilitated by Data Transfer Agreements. However, navigating the intricate landscape of data protection laws and ethical considerations poses challenges, and there is a pressing need for comprehensive guidelines to legal, ethical and practical considerations. Such guidelines would help researchers who want to share their data ensure that they abide by responsible data sharing practices.

The second day of the workshop focused on sharing of advanced imaging data from animal experiments, with participants from the University of Oslo and the Norwegian University of Science and Technology.

Mikkel Lepperød (UiO) shared experiences from an ongoing project to re-use electrophysiology data in a large-scale analysis of place cell center distributions. Having asked many principal investigators to share their data, his experience was that “data available upon request” usually comes with the unspoken condition “if my postdoc can find it”. When data were shared however, painstaking work is often needed to understand the structure of the data and metadata, and Lepperød concluded his talk by urging researchers to “pick a standard, any standard”, when organizing their data for sharing.

Charlotte Boccara (UiO) highlighted the need to have access to raw data behind publications. After having attempted to reproduce findings from another laboratory without success, she had been given access to the raw data. Upon further analysis, it became clear that differences in the recording conditions that were not apparent from the methodological descriptions had been the root of the different results. She further emphasized the need for a “phone book” for data, a known place to look up where data for a given publication can be found.

Simon Ball (NTNU), senior engineer at the Kavli Institute for Systems Neuroscience, gave a perspective from aiding researchers in sharing their data over several years. He explained that while the goal of the institute was to go beyond the bare minimum of data sharing, the researchers are primarily motivated by publications and may not feel particularly compelled to comply with proposed metadata standards.

A common theme across these talks, then, was the lack standards for data and metadata organization and the consequent lack of essential information in publications or difficulty in interpretation of shared data. While they can be re-organized post hoc, it can be problematic when data sharing is treated as an after-though or a matter of box-ticking. Ball raised the interesting challenge that when data are re-organized for sharing, it often causes them to be incompatible with related code. Unless willing to commit substantial time to re-write code, researchers in this situation may have to choose between optimizing their data for re-usability or re-producibility.

A key lesson from the workshop is that many of the challenges experienced by researchers may be mitigated by planning for sharing data early on, and by adopting data and metadata standards where they exist. An important task for the Norwegian node of EBRAINS and INCF going forwards will be to convince researchers that these practices are worthwhile and to assist them in adopting them.