Interview with EBRAINS Chief Science Officer Viktor Jirsa: “There was real engagement at the EBRAINS Roadmap Symposium”

What was your overall impression of the event?

It is the first time a public consultation of this kind has been run in neuroscience, so the outcome was not at all guaranteed. The response to the call was superb: we received more than a hundred submissions, 159 in total. More than 160 people then travelled to the meeting from across Europe, and some from other parts of the world, to take part in and contribute to the debates. What struck me was the atmosphere: there was real engagement, and people clearly wanted to be serious about the science being put forward. I was proud to be part of such an engaging process.

What challenges were raised by the discussions? Were they different across themes?

On the one hand, there was surprising coherence across the themes. Function arising from multiscale architecture came up several times, and so did digital twins, which is not surprising for a digital neuroscience infrastructure. On the other hand, there were genuinely competing approaches, and people took clear positions and perspectives. What I found healthy is that it was perfectly acceptable when some of those positions could not be merged into a compromise; in that case they simply must coexist. One example is the question of whether we should focus on a few topics and take them far for a huge impact, or stay as wide as possible to sample many different directions. That is not something that can be reconciled, and it does not have to be: we need both. And these challenges persisted across the themes rather than being specific to any one of them.

Were there any missing tools or services that the proposals identified?

From the proposals themselves, one topic that came up a great deal was what lies outside the brain: integration with the body, and integration into a wider ecosystem. A large number of proposals wanted that addressed. We had expected a lot of AI, and there was a lot, but even that was thinner than the trajectory of the field would predict. Neuro-robotics, too, was much more lightly represented than we had expected. And there was a strong focus on clinical translation, which was not surprising, to the point where essentially all of the neurotechnology that was proposed was oriented toward clinical applications.

Beyond the topics themselves, the same underlying needs reappeared across very different proposals: harmonised and well-annotated data, federated analysis that respects privacy and governance, shared benchmarks and validation frameworks, and interoperability standards, particularly for anything intended to become a clinical-grade digital twin. In that sense the bottleneck is shifting away from the individual model and toward the shared scaffolding that lets models be compared, trusted, and reused.

How will the discussions inform the writing of the EBRAINS Roadmap?

The discussions provide a sharpening, a critical sharpening, of the topics and the content of the submissions. So the writing will be informed by at least three inputs together: all of the submissions that were made, the questionnaires that were previously circulated in the national nodes, and the debates themselves. The debates play a particularly important role, because they are what allow us to structure and to weigh the written inputs that were provided. The before-and-after voting we ran is part of this: it gives a concrete reading of how the discussion shifted priorities, so that competing directions can be weighed honestly rather than by acclamation. All of this enters as material for the writing team. It is worth saying that the symposium was deliberately the debate phase, kept separate from the writing phase; its task was not to draft text, but to hand the writing team a clear and weighted mandate.

Which topics emerged as priorities for the next ten years?

The symposium debates produced a clear consensus on the priorities that will define the next decade. Chief among them is the powerful combination of (agentic) AI and digital twins. Personalized, clinically-oriented brain twins emerged as a standout priority, seen as a particularly promising path toward more precise and effective healthcare. At the same time, participants stressed the urgent need to connect the brain more deeply with the body and its wider environment — and called for significantly stronger research efforts across these interconnected areas.

What are the next steps? When will the Roadmap be published?

The next step is to establish the writing team. I expect something on the order of twenty people, enough to be representative but not so many that it becomes unworkable. Once the team is in place, we will agree on the best strategy for producing the roadmap and on how to work together. The accepted contributions have already been made publicly available, and, beyond this, we seek open-source outlets, so that the record of the consultation is citable. We expect the roadmap itself to be finished towards the end of this year, 2026.