THE SHAREBRAIN WORKFLOW FOR SHARING AND RE-USING COMPLEX FUNCTIONAL DATA: FROM LAB-SPECIFIC STORAGE TO SHARED, STANDARDIZED AND MACHINE-ACTIONABLE DATA

Advanced instruments and specialized technologies are increasingly applied to acquire complex experimental data describing functional properties of the brain. While the FAIR principles are widely adopted, reuse of shared data is limited and apparently hampered by methodological complexity and lack of standardization. A key challenge is that researchers employing sophisticated methods have their own modus operandi and data management systems tailored to their experimental protocols, while adaptation of community standards requires knowledge and additional effort. This lack of standardization makes it difficult to utilize open access functional brain data in automated analyses driven by machine-learning. Addressing this, we have developed the SHAREbrain [1] workflow for organizing complex functional data and extracting standardized metadata. The workflow builds on the NANSEN (Neuro ANalysis Software ENsemble) [2], a MATLAB based toolbox providing a user-friendly graphical user interface, and features custom (laboratory-specific) file-adaptors with automated data converters for creating data organized according to the Neurodata Without Borders (NWB) standard and extracting in-depth openMINDS metadata suitable for sharing and reuse via the EBRAINS research infrastructure. We have validated the SHAREbrain workflow on a range of datasets from different laboratories, including electrophysiological recordings, 2-photon microscopy data, and sleep behaviour analysis. The resulting NWB data are shared as machine-actionable data via EBRAINS. We here demonstrate how the SHAREbrain workflow allows standardization of complex data originating from different laboratories and facilitates automated machine-learning based analysis.

[1] https://github.com/Neural-Systems-at-UIO/SHAREbrain
[2] https://github.com/VervaekeLab/NANSEN
[3] https://ebrains.eu (RRID:SCR_019260)

Funding: EU Horizon Europe #101147319 (EBRAINS2.0), Research Council of Norway #350201 (NORBRAIN4), and UIO (SHAREbrain).