Neuroscience data in the cloud spans multiple data types—clinical, genetic, neuroimaging, and real-world—as well as multiple modalities, species, and diseases, and thus requires robust and interoperable platforms, said Maryann Martone. Integrating across data types offers additional power and will also require novel analytical tools, added Silvana Borges. One platform that is working in this space is the BRAIN Commons,1 which according to Lee Lancashire, chief information officer at Cohen Veterans Bioscience, has developed a graph-based data model that allows users to capture multi-modal data at the case level so that as new data are generated, they can be incorporated into the data model. Workshop participants also discussed features of other platforms that are integrating multiple sources and types of data, including the AMP-PD, Vivli (Chapter 7); the Psychiatric Genomics Consortium (Chapter 8); the Collaborative Informatics and Neuroimaging Suite (COINS), CBRAIN, the Longitudinal Online Research and Imaging System (LORIS), the Canadian Open Science Platform (CONP), and the National Data Archive (NDA) at NIMH (all in Chapter 9).
In four breakout sessions, workshop participants discussed other challenges and opportunities specific to different types of neuroscience data. These discussions are summarized in Chapters 7 through 10, which cover clinical trial and research data (Chapter 7); genetic data (Chapter 8); neuroimaging data (Chapter 9); and real-world data (Chapter 10). These discussions were intended to be orthogonal to the discussions organized by