Team Julie PERROY
Pathophysiology of synaptic transmission
Project Neural bases of decision-making
PRINCIPAL INVESTIGATOR
IGF staff involved
Coline CHEVALLIER
PhD student CNRS
Julie PERROY
DR1 CNRS
Pierre VINCENT
DR2 CNRS
Enora MOUTIN
CRCN CNRS
Vivien SZABO
CCA INSERM
Nathalie BOUQUIER
IE CNRS
We study how cognitive properties such as learning and decision-making emerge from biological substrates. This requires building mechanistic theories of how network of neurons may learn and decide; and testing model predictions with experiments. We use mathematical models at different levels (from synapse biophysics to reinforcement learning) together with molecular tools (chemo- and opto-genetics) and in-vivo recordings in mice performing refined behavioral tasks to get (a lot of) rewards. We welcome applications from biologists and physicists wishing to ask ambitious questions in a inclusive environment.
Neural bases of flexibility
Animals need to learn fast when the environment is changing, but to decrease their rate of learning when the world is stable or uninformative. We aim to assess how such adaptive learning arises from the trade-off between plasticity and stability of the synapses; from the molecular to the network level. To this aim, we are combining state-of-the-art experimental techniques in behaving mice (recordings, manipulations) with computational modeling at the synaptic and network levels.
Left: mice have to choose between two levers associated with different reward probabilities. The probabilities change over time and mice have to adapt their behavior. Right: we compare recordings and inactivation in key brain areas with biophysical models (network, synapses).
Main publications
• Bousseyrol, E et al. (2023) Cell Reports, 42(5).
Funding
• 2023-2027 FRM équipe (Partenaire)
• 2022-2026 ANR LEARN (Partenaire)
Alumni
• Dalila Kritli (Master 2, 2023)
• Emma Debos (Master 1, 2023)
• Emma Marillat (Master 1, 2022)
Neural bases of decision timing
Deciding when to act is as important as deciding what to do. Neurons activate just before properly timed decisions. How does such precisely timed activity emerge? Is this activity a cause or a consequence of correct timing? We use selective and reversible inactivation of neuronal ensembles (using inducible c-fos technology) during the production of a timed behavior, to compare and select among computational models (recurrent networks) embodying competing hypotheses of mechanisms for timed decisions.
From left to right: Mice are required to learn to press lever twice, at a precise inter-press timing. Mice are able to progressively tune their timing. cFos TetTag technology tags the cortical neurons activated during the production of a given inter-press interval of time, which allows further manipulations (inactivation and recordings). Mathematical models of cortical network of neurons can produce a peak activity at the proper timing, providing competing theoretical predictions to be compared with experiments.
Funding
• 2023-2027 ANR TimeTag
Collaborations
• Stéphanie Trouche (IGF)
• Bruno Delord (Sorbonne U.)
Alumni
• Richmond Crisostomo (Master 2)
