Current Projects


Higher-order interactions in human brain networks supporting causal learning

Funding: A*Midex Aix-Marseille University 2024-2028

Consortium: Andrea Brovelli (PI), Alain Barrat (CPT, Marseille)

*Learning is a brain network phenomenon thought to arise from synergistic interactions between multiple brain regions. Although central, this hypothesis has never been fully tested, yet. Indeed, progress has been limited by the lack of approaches for studying brain interactions beyond pairwise relations, the so-called higher order interactions (HOIs). Our objective is to build a theoretical and data analysis framework to demonstrate the role of HOIs in human brain networks supporting causal learning. *

Brain Health Trajectories (BHT)

Higher-order interactions in human brain networks supporting causal learning

Funding: PEPR Santé Numerique (ANR) 2024-2028

Consortium: Andrea Brovelli & Matthieu Gilson in collaboration with Viktor Jirsa (INS, Marseille), Demian Battaglia (Strasbourg Univ), Bertrand Thirion (Paris Saclay)

Our goal is to develop tools that characterize the trajectories of brain health during aging and deviations in pathological conditions. Our project focuses on the study of fMRI data and directional network interactions using Granger causality and effective connectivity modelling.


Inference and validation of network-level models

Funding: EBRAIS (EU) 2024-2027

Consortium: Andrea Brovelli in collaboration with Viktor Jirsa (INS, Marseille), Daniele Marinazzo (Ghent Univ)

Our goal is to provide the tools for use of brain models in the EBRAINS ecosystem, enabling high performance simulation and parameter sweeps, model validation and inference of computational models for brain network activity.


Whole-brain anatomo-functional modelling of healthy and pathological brain conditions

Funding: Laënnec Institute, NeuroMarseille (AMU) 2023-2025, ARSLA 2023-2025

Consortium: Matthieu Gilson (PI), Wafaa Zaaraoui, Aude-Marie Grapperon (CEMEREM, Marseille), Xenia Kobeleva (Bochum, Germany)

Amyotrophic lateral sclerosis (ALS) is a devastating disease that results in loss of muscle function. This project aims to develop and test prediction tools with a particular focus on the stratification of ALS patients in subtypes with slow versus fast disease progression. We aim to build a multimodal prediction pipeline that combines sodium, quantitative structural and functional magnetic resonance imaging (MRI).


Whole-brain anatomo-functional modelling of healthy and pathological brain conditions

Funding: Amidex Aix-Marseille Université, Oct 2022 - Sept 2025

Matthieu Gilson (PI)

Collaborators: Andrea Brovelli, Bjorg Kilavik, Laurent Perrinet, Emmanuel Daucé (INT, Marseille), Olivier David, Meysam Hashemi (INS, Marseille)

The aim of this project is to develop analysis tools of neurophysiological signals to interpret them in terms of brain communication. The general approach relies on dynamic network models fitted to the data to reproduce their propagation pattern, in particular by fitting directional interactions between neuronal populations or brain regions, also known as effective connectivity (EC). A specific focus consists in constraining EC by anatomical data (e.g. structural connectivity, SC) such that it describes the strength modulation of synaptic connections or cortico-cortical white-matter fibers. In this way, we obtain robust EC-based signatures (or biomarkers) to characterize the neuronal dynamics at the single trial/session level, which allows us to decode cognitive or pathological states and can further be analyzed using network theory to identify e.g. functional communities of brain regions.


Representation of natural sounds in the human brain: Transforming acoustics into the semantics of sound sources in the environment

Funding: Funding: Agence Nationale de la Recherche ANR - PRC, Apr 2022 - Mar 2026

Consortium: Bruno L Giordano (PI), Daniele Schön (INS, Marseille), Thierry Artieres (LIS, Marseille), Elia Formisano (Maastricht University, The Netherlands)

Our brain recognize objects and events in the environment (e.g., glass breaking, baby crying) by transforming what we hear into representations of the semantics of the sound source. This project aims to characterize this transformation by measuring how three key properties of cerebral sound representations (spatio-temporal dynamics - MEG; transfer between brain areas - iEEG; differences across cortical layers - high-field fMRI) are explained by a wide range of computational models (acoustics, semantics, deep neural networks).


Bio-mimetic Agile aerial roBots flying in real-life conditions

Funding : Agence Nationale de la Recherche ANR - PRC, March 2021 - Sept 2024

Consortium: Emmanuel Daucé (Partner), Laurent Perrinet (PI), Ryad Benosman (IDV, Paris), Stéphane Viollet (ISM, Marseille)

In this project, we propose a solution that will integrate bio-inspired rapid visual detection and stabilization dynamics into an end-to-end event based neuromorphic system. The key to this approach will be the optimization of delays through predictive processing. This will allow these robots to fly independently, without any user intervention.

Past Projects


Cognitive architectures of causal learning

Funding: Agence Nationale de la Recherche ANR - PRC, Jan 2019 - Nov 2023

Consortium: Andrea Brovelli (PI), Mateus Joffily (GATE, Lyon), Mehdi Khamassi (ISIR, Paris), Julien Bastin (GIN, Grenoble)

The aim of this project is to study the neural and computational bases of human goal-directed causal learning by combining human neurophysiology (MEG and intracranial SEEG) and neuroimaging (fMRI) techniques with computational models of learning (Reinforcement Learning and Active Inference).


Model-free and model-based inference and validation workflows for causal brain network discovery

Funding: Human Brain Project HBP/EBRAINS, Oct 2020 - Sept 2023

Consortium: Andrea Brovelli (PI), Jean Daunizeau (ICM, Paris), Gustavo Deco (UPF, Barcelona), Stefano Panzeri (IIT, Italy), Petra Ritter (Charité, Berlin), Olivier David (GIN, Grenoble)

The project aims at combining data-driven and model-based approaches for the inference of brain causal connectivity and the validation of whole-brain models. The goal is to integrate workflows and methods into EBRAINS. This projet is part of WP1 of SGA3 of the HBP.


Modelling information dynamics

Funding: 2 yrs post-doc fellowship from the Institut de Convergence ILCB, May 2019 - April 2021

Consortium: Andrea Brovelli (co-PI), Demian Battaglia (co-PI; INS, Marseille; USIAS, Strasbourg)

The project aims at developping an multi-area interacting model based on the ring model. The goal is to study neural computations such as stimulus encoding, information transfer and information modification by means of simulation studies. This projet is part of Institut de Convergence ILCB.


The influence of directional interactions in brain networks in predicting cognitive deficits post-stroke

Funding: FLAG-ERA HBP Partnering project, Jan 2018 - dec 2020

Consortium: Andrea Brovelli (co-PI), Marizio Corbetta (PI; Univ Padova, Italy)

The project aims to develop and test methods for the analysis of directional interactions among brain regions in resting state and task-evoked fMRI data in healthy controls and stroke damaged individuals, and to develop a computational model of an injured brain that replicates both the empirically measured patterns of connectivity abnormalities and behavioral deficits.