BAMB! 2025: A School for Models, Minds, and the Messy Art of Behavior

In a classroom right next to the Mediterranean, the fifth edition of the BAMB! summer school came to life. For nine days, 30 early-career researchers stepped away from their usual routines and immersed themselves in the elusive craft of modeling behavior. What they found was a set of techniques and a community of minds committed to asking questions about how animals, ourselves included, make sense of a world full of noise, pattern, and uncertainty.

BAMB! was more than lectures. It was a place for connection and the quiet satisfaction of figuring things out together.

The school, officially titled Barcelona Summer School for Advanced Modeling of Behaviour, has established itself as one of the premier training grounds for computational neuroscientists in Europe. Organised by researchers from CRM, ENS Paris, Sorbonne University, Oxford/DeepMind, and the Paris Brain Institute, and hosted this year at the Parc de Recerca Biomèdica de Barcelona (PRBB), BAMB! 2025 stayed true to its founding principle: rigorous hands-on learning rooted in conceptual clarity.

Each day unfolded around a different modeling tradition. From parameter recovery and model comparison to reinforcement learning, Bayesian inference, latent variable models, and simulation-based techniques, the curriculum moved fast and with purpose.

The BAMB! School also centres around group projects designed to make the newly acquired knowledge resonate with each participant’s own research. On the final day, each team had ten minutes to present their approach. No one expected polished results. What mattered was the thinking behind the model, the clarity of the question, and the honest effort to answer it.

Two keynote talks closed two of the school’s days. The first was by Athena Akrami, from the Sainsbury Wellcome Centre at University College London. Her research asks how animals uncover patterns in the world and how they detect structure in noise, build abstract relationships, and carry those priors forward to guide behaviour. From the fluidity of working memory to the rigidity of perceptual biases, Akrami’s work captures cognition as a living system woven from past experience and statistical expectation.

Later in the week, Peter Dayan, director at the Max Planck Institute for Biological Cybernetics in Tübingen, took the stage. A pioneer at the intersection of artificial intelligence, neuroscience, and psychiatry, Dayan’s career has shaped how we understand learning, reward, and neuromodulation. His research deals with the brain’s decision-making machinery and its failure modes, always anchored in formal models.

A place for community and connection

Beyond the lectures and tutorials, BAMB! was held together by something quieter: conversations over lunch, fragments of code exchanged at coffee breaks, one-to-one meetings where career anxieties found sympathetic ears. Faculty weren’t distant figures; they were co-conspirators. The teaching assistants were tireless, orbiting from laptop to laptop like helpful satellites, keeping the chaos just within bounds.

Huge thanks to the organising team, the brilliant teaching assistants, keynote speakers, and all the participants whose energy, curiosity, and collaboration made BAMB! 2025 such a memorable and inspiring experience. This was BAMB!’s fifth edition, and it hasn’t lost its edge.