Seminal
The magazine and atlas of EU-funded researchA new publication from the web and AI team at Scott Charnley

Atlas / France

Marie Skłodowska-Curie Action · Running

DnReLU Decision noise in reward-guided learning amidst option unavailability

Coordinator
Ecole Normale Superieure, France
Runs
1 October 2024 to 30 September 2026
Status
Running
EU funding
€196k
Total cost
—
Partners
2 organisations in 2 countries
Type
Marie Skłodowska-Curie Action
Grant agreement
101154160

What the project is doing

In the project's own words, from its summary on CORDIS.

Navigating a new city's dining scene often confronts us with unavailable options, e.g., 'fully booked,' shaping our behavioral variability such as random exploratory choices and learning errors. Existing theories of reward-guided learning typically overlook option unavailability. Hypothesized strategies may vary, from adjusting action stochasticityover-exploration to broadly check alternatives, or under-exploration to quickly latch onto available favoritesto adjusting learning process, e.g., weighing past choices for maximal future rewards, or heuristically inflating the value of busy venues while devaluating others ('busier = better').

A nuanced understanding of these tactics requires a deep dive into information sampling, encompassing not just overt evidence-seeking actions but also covert attentional sampling during value updates in the internal model of the world. These complex sampling patterns, unfolding rapidly over hundreds of milliseconds, have been chiefly studied through eye-tracking that captures overt oculomotor variables, leaving the subtleties of covert sampling unexplored and detailed mechanistic insights elusive.

I aim to address this issue across multiple scalesfrom neurotransmitter and neural dynamics to biophysically-grounded modeling of behavior. Using pharmaco-magnetoencephalography (MEG), I will track covert attentional focus across the entire timespan of every single reward-guided decision during learning, in detailed source-localized brain regions, while also perturbing neurochemical systems to establish causal links. This research promises to illuminate longstanding questions, such as why behavioral variability persists despite evolutionary pressures favoring reward-maximizing decisions.

The training through this MSCA in circuit-level mechanistic models, imparted by world-leading experts, will complement my existing expertise, and establish me as an independent researcher in the burgeoning field of cognitive computational neuroscience.

Topics

Who is involved

OrganisationCityCountryRole
Ecole Normale SuperieureParisFranceCoordinator
Universitaetsklinikum Hamburg-EppendorfHamburgGermanyAssociatedpartner

Source: CORDIS, European Commission. Data of 5 Nov 2026, reused under the Commission's reuse policy. View on CORDIS · Grant DOI