Scientific Themes
The brain in context
Brain activity and behavior don’t happen in isolation - they are shaped by the body’s internal state, the physical and social environment, prior experience, learning, development, and aging. Where possible, we emphasize tracking neural and behavioral change across contexts and over time, rather than relying on single recording sessions. Measurements distributed across animals, labs, and settings help us tell real biological change apart from noise, and reveal patterns that generalize.
How the brain builds and uses internal models
This theme asks how the brain constructs models of the world to guide decisions - representing uncertainty, anticipating what comes next, weighing possible actions, and planning sequences of behavior. We combine decoding analyses with computational modelling to connect brain-wide measurements to these internal states.
Rigorous basic science with real-world relevance
We study fundamental questions about brain and behavior in rodents - decision-making, motivation, cognitive flexibility, learning, affect regulation - with an eye to their broader significance. IBL does not conduct human research, but we aim to establish circuit-level mechanisms with enough rigor and generalizability to inform human neuroscience.
Methodological Commitments
AI-accelerated science, with humans in the loop
We are building AI-assisted workflows for systems neuroscience: natural-language access to complex datasets, code generation, automated analysis, interactive visualization, model-based experiment selection, and tighter iteration between experiments and computational models. This work is held to the same rigor as any other science - pre-registration, systematic checks for inconsistencies and hallucinations, and confirmation against held-out data.
Longitudinal and chronic recording
We track neural and behavioral change across sessions, contexts, and significant portions of an animal’s life - through chronic electrophysiology and imaging, cross-session neuron tracking, home-cage and freely-moving behavior, sleep and physiological monitoring, and alignment across timescales. Stable neural identity and shared reference measurements let us distinguish genuine biological change from drift in instrumentation or behavior.
Behavioral video analysis
Quantitative descriptions of movement, posture, facial and orofacial signals, and naturalistic or social behavior - via pose estimation, tracking, multi-camera reconstruction, and data-driven behavioral representations - can be integrated with neural and physiological measurements. Shared formats and reusable analysis tools make it possible to compare results across algorithms, tasks, environments, and sites.
Our Ethos
Building resources that last
Software, analytical methods, and datasets are scientific outputs in their own right, not by-products. We prioritise modularity, reliability, stable interfaces, tested workflows, and clear documentation so that resources remain usable well beyond their original context - and so that past analyses can be revisited as methods improve.
Quality control and reproducibility
We treat quality control and reproducibility as core scientific requirements, not optional extras. This means consistent acquisition standards, essential metadata, validation procedures, quantitative measures of data quality, checks that hold up across sites and modalities, and honest reporting of uncertainty. Benchmarks, evaluation datasets, and automated QA tools make these practices something the wider community can reuse.
Collaboration at scale
Multi-site collaboration is central to how IBL works - it’s what makes our findings generalizable and lets us pool expertise, procedures, and quality standards across labs. It also lets us build tools and infrastructure that serve the broader neuroscience community, not just our own network. We’re growing this collaborative model deliberately: deepening partnerships and building shared expertise in our current priority directions as the foundation for an expanding network in the years ahead.
