Open neuroinformatics infrastructure ecosystem for federated multisite studies
Wang, M.; Bhagwat, N.; Cremonesi, F.; Dugre, M.; Pfarr, J.-K.; d'Angremont, E.; Dai, A.; Jahanpour, A.; Urchs, S.; Cansiz, S.; Chambon, L.; Dincer, A. T.; Torres, J.; Vesin, M.; Pinilla-Monsalve, G.; Song, Y.; Vriend, C.; Jeanson, F.; Monchi, O.; van der Werf, Y. D.; Lorenzi, M.; Poline, J.-B.
Show abstract
Despite growing understanding of the benefits of having Findable, Accessible, Interoperable, and Reusable (FAIR) data, many datasets still cannot be shared. Federated analysis methods can enable multisite studies that do not require the sharing of participant-level information. However, there are many practical hurdles that prevent the large-scale adoption of federated methods. We discuss challenges related to cross-site data preparation for federated learning, present solutions offered by recent neuroinformatics projects, and showcase an example of tool integration applied to neurodegenerative disease data.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.