Priors and Inference in Superior colliculus during Movement

Description du projet

Anticiper l’avenir pour guider les mouvements

Les animaux ne se contentent pas de réagir à leur environnement: ils prédisent aussi constamment ce qui va se produire, en s’appuyant sur leur expérience passée pour prendre des décisions rapides et flexibles. Comprendre comment le cerveau élabore et actualise ces prédictions au cours de comportements naturels reste un défi fondamental en neurosciences. Avec le soutien du programme Actions Marie Skłodowska-Curie, le projet PRISM développera une nouvelle tâche comportementale dans laquelle des souris poursuivront une cible mobile régie par des règles probabilistes, permettant ainsi une étude rigoureuse des attentes et des prédictions. À l’aide de techniques d’électrophysiologie haute densité et d’outils optogénétiques appliqués à des animaux se déplaçant librement, les chercheurs étudieront la manière dont le colliculus supérieur – un centre essentiel du mésencéphale pour l’intégration sensorimotrice – représente et actualise les prédictions internes. Cette étude apportera un nouvel éclairage sur les bases neuronales de la prise de décision adaptative.

Objectif

To survive, animals must do more than react to immediate sensory input; they must predict what is likely to happen next. This ability to integrate past experience with current information is essential for fast and adaptive decision-making. Yet how the brain forms, updates, and uses such expectations during natural behavior remains poorly understood, particularly in naturalistic settings where sensory cues are ambiguous, and decisions must be made rapidly and flexibly.
In this project, I will develop a novel behavioral task in which mice pursue a physical target moving according to probabilistic rules, bridging the gap between classical conditioning paradigms and naturalistic goal-directed behavior. By changing these rules, I will examine how mice form internal models of their environment and adapt to new patterns. I will test whether animals show predictive behaviours that reflect prior expectations, especially when sensory input becomes uncertain. Their behaviour will be interpreted within a Bayesian inference framework, allowing rigorous comparison to an ideal observer model.
At the neural level, I will focus on the superior colliculus, a midbrain hub involved in integrating sensory information to guide movement. Using high-density electrophysiology in freely moving animals, I will investigate how prior expectations are represented and updated in real time. I will complement this with targeted manipulations of inputs to the superior colliculus using optogenetic and chemogenetic tools to test how these circuits contribute to forming and updating internal models.
While I already have strong expertise in behavioural tracking and neuronal recordings, this project will provide me with essential training in Bayesian modelling and causal circuit perturbation, key tools for next-generation systems neuroscience. Hosted at a world-renowned institute, this fellowship will deepen my interdisciplinary skillset and support my transition toward an independent research career.