“Unprecedented opportunities for brain research”

Supercomputing and large-scale, multidisciplinary collaboration are creating completely new opportunities for brain research. The European Human Brain Project (HBP) has made these new forms of collaboration possible. It has created a digital platform, the EBRAINS research infrastructure, where scientists from all over the world can find data, analysis tools and access to high-performance computing. This opens up unprecedented opportunities for brain research, medicine and technology development.

In our current position paper, more than 100 authors have outlined a common roadmap for digital neuroscience over the next decade. It identifies eight research priorities. These include high-resolution anatomical models such as the Julich BigBrain and detailed mapping such as the Julich Brain Atlas. Other priorities are applications in medicine and artificial intelligence (AI). The work also points the way for neuroscientific research into further medical applications. The “digital twin” will play a key role here: Computer-aided, mathematical brain models that are continuously compared with real measurement data. Such personalized models will increasingly improve the diagnosis and treatment of brain diseases, for example, through more precise surgical treatment of epilepsy.

However, further research must address the problem of the lack of harmonization and exchange of often fragmented clinical data. Together with its partner, the European Academy of Neurology, EBRAINS is developing a process to enable the harmonization, exchange and sharing of health data. There is also enormous potential in new AI-based methods. AI foundation models in particular are changing research in a way that allows completely new questions to be addressed. EBRAINS can play an important role in this and pave the way for European foundation models of the brain in neurological and psychiatric diseases. It offers a well-developed data infrastructure with increasingly large training data sets of high quality.

It is not only the neurosciences that benefit from this. Knowledge about the brain is also being transferred back to AI, and algorithms developed in the HBP using the brain as a model have already demonstrated remarkable advantages: They have shown high energy efficiency, flexibility and plasticity as well as the ability to learn from sparse data. Access to the enormous computing capacities of the first European exascale computer JUPITER at Forschungszentrum Jülich is also important here. It is of great importance that the potential offered by the new exascale computing capacities for the neurosciences can be used effectively through appropriate research programs. In this way, a unique productive cycle between AI and brain research can be created at the interface to computing, further advancing both fields.

Link to the Paper

EBRAINS