EN FR
EN FR
NERV - 2025

2025Activity reportProject-Team​NERV

RNSR: 202324451G
  • Research​‌ center Inria Paris Centre​​ at Sorbonne University
  • In​​​‌ partnership with:CNRS, INSERM,​ Sorbonne Université
  • Team name:​‌ Systems neuroengineering to model​​ and interface brain networks​​​‌
  • In collaboration with:Institut​ du Cerveau et de​‌ la Moelle Epinière

Creation​​ of the Project-Team: 2023​​​‌ October 01

Each year,​ Inria research teams publish​‌ an Activity Report presenting​​ their work and results​​​‌ over the reporting period.​ These reports follow a​‌ common structure, with some​​ optional sections depending on​​​‌ the specific team. They​ typically begin by outlining​‌ the overall objectives and​​ research programme, including the​​​‌ main research themes, goals,​ and methodological approaches. They​‌ also describe the application​​ domains targeted by the​​​‌ team, highlighting the scientific​ or societal contexts in​‌ which their work is​​ situated.

The reports then​​​‌ present the highlights of​ the year, covering major​‌ scientific achievements, software developments,​​ or teaching contributions. When​​​‌ relevant, they include sections​ on software, platforms, and​‌ open data, detailing the​​ tools developed and how​​​‌ they are shared. A​ substantial part is dedicated​‌ to new results, where​​ scientific contributions are described​​​‌ in detail, often with​ subsections specifying participants and​‌ associated keywords.

Finally, the​​ Activity Report addresses funding,​​​‌ contracts, partnerships, and collaborations​ at various levels, from​‌ industrial agreements to international​​ cooperations. It also covers​​​‌ dissemination and teaching activities,​ such as participation in​‌ scientific events, outreach, and​​ supervision. The document concludes​​​‌ with a presentation of​ scientific production, including major​‌ publications and those produced​​ during the year.

Keywords​​​‌

Computer Science and Digital​ Science

  • A5.1.4. Brain-computer interfaces,​‌ physiological computing
  • A5.2. Data​​ visualization
  • A5.9. Signal processing​​​‌
  • A6.1. Methods in mathematical​ modeling
  • A6.4.3. Observability and​‌ Controlability
  • A8.8. Network science​​
  • A9.3. Signal processing

Other​​​‌ Research Topics and Application​ Domains

  • B1.2. Neuroscience and​‌ cognitive science
  • B2.1. Well​​ being
  • B2.2. Physiology and​​​‌ diseases
  • B2.5. Handicap and​ personal assistances
  • B2.6. Biological​‌ and medical imaging
  • B5.10.​​ Biotechnology
  • B9.5. Sciences

1​​​‌ Team members, visitors, external​ collaborators

Research Scientists

  • Fabrizio​‌ de Vico Fallani [​​Team leader, INRIA​​​‌, Senior Researcher,​ HDR]
  • Mario Chavez​‌ [CNRS, Senior​​ Researcher, HDR]​​​‌
  • Marie-Constance Corsi [INRIA​, Researcher]

Faculty​‌ Member

  • Laurent Bougrain [​​UL, Associate Professor​​​‌ Delegation]

Post-Doctoral Fellows​

  • Diego Candia Rivera [​‌ICM]
  • Andrea Civilini​​ [INRIA, Post-Doctoral​​​‌ Fellow, from Jun​ 2025]
  • Tristan Venot​‌ [INRIA, until​​ Sep 2025]

PhD​​​‌ Students

  • Bruno Aristimunha Pinto​ [INRIA, Co-supervised​‌]
  • Camile Bousfiha [​​INRIA, Co-supervised]​​​‌
  • Cassandra Dumas [ICM​, Co-supervised]
  • Baptiste​‌ Fague [ESSILOR,​​ CIFRE, from Oct​​​‌ 2025]
  • Marc Fiammante​ [SORBONNE UNIVERSITE]​‌
  • Jules Gomel [INRIA​​ Saclay, ISAE Supaero,​​​‌ Co-supervised]
  • Martin Guillemaud​ [SORBONNE UNIVERSITE, ENS,​‌ INRIA]
  • Alice Longhena​​ [SORBONNE UNIVERSITE, INRIA​​​‌, until Apr 2025​]
  • Camilla Mannino [​‌SORBONNE UNIVERSITE, INRIA,​​ Co-supervised]
  • Marion Pavaux​​​‌ [THALES, CIFRE​]
  • Adrien Pontlevy [​‌INRIA]
  • Wafa Skhiri​​ [INRIA]
  • Sébastien​​ Velut [UNIV PARIS​​​‌ SACLAY, Co-supervised]‌

Technical Staff

  • Celestine Allombert-Blaise‌​‌ [ICM, Engineer​​, from Sep 2025​​​‌]
  • Marion Couton [‌ICM, Engineer,‌​‌ from Sep 2025]​​
  • Arthur Desbois [INRIA,​​​‌ ICM, Engineer]‌
  • Camille Gabot [INRIA‌​‌, Engineer, from​​ Sep 2025]
  • Laurent​​​‌ Hugueville [CNRS,‌ Engineer, 20%]‌​‌
  • Erin Soller [INRIA​​, Engineer, from​​​‌ Dec 2025]
  • Sergio‌ Talegon De La Fuente‌​‌ [INRIA, Engineer​​, from Sep 2025​​​‌]

Interns and Apprentices‌

  • Diana Baili [ICM‌​‌, Intern, until​​ Jul 2025]
  • Marion​​​‌ Couton [SORBONNE UNIVERSITE‌, Intern, until‌​‌ Jun 2025]
  • Linon​​ Denis [ICM,​​​‌ Intern, from Oct‌ 2025]
  • Francesco Farina‌​‌ [INSERM, Intern​​, from Feb 2025​​​‌ until Jul 2025]‌
  • Camille Gabot [ICM‌​‌, Intern, until​​ Jul 2025]
  • Pierre-Baptiste​​​‌ Mathieu De Carvalho [‌Loria, Intern,‌​‌ from Apr 2025 until​​ Sep 2025]
  • Laura​​​‌ Pitti [INSERM,‌ Intern, from Sep‌​‌ 2025]
  • Mario Roca​​ [INSERM, Intern​​​‌, from Feb 2025‌ until Aug 2025]‌​‌
  • Giovanni Sitti [INSERM​​, Intern, from​​​‌ Sep 2025]
  • Bintou‌ Soumaoro [ICM,‌​‌ Apprentice, until Sep​​ 2025]

Administrative Assistant​​​‌

  • Helene Milome [INRIA‌]

Visiting Scientist

  • Giovanni‌​‌ Messuti [Università degli​​ Studi di Salerno,​​​‌ from Nov 2025,‌ PhD Student]

2‌​‌ Overall objectives

NERV is​​ an Inria project-team joint​​​‌ with the Paris Brain‌ Institute (ICM) at the‌​‌ Pitié-Salpêtrière hospital (AP-HP) in​​ Paris. NERV was created​​​‌ as a project-team in‌ 2023 and later became‌​‌ team at the Paris​​ Brain Institute (ICM) in​​​‌ 2025. NERV has a‌ joint affiliation to Inria,‌​‌ CNRS, Inserm and Sorbonne​​ University.

NERV is thus​​​‌ located both within a‌ leading neuroscience institute and‌​‌ within a large hospital.​​ This unique position has​​​‌ several advantages: direct contact‌ with neuroscientists and clinicians‌​‌ allows us to foresee​​ the emergence of new​​​‌ problems and opportunities for‌ new methodological developments, provides‌​‌ access to unique datasets​​ and experimental platforms, and​​​‌ eases the transfer of‌ our results to clinical‌​‌ research and clinical practice.​​

Our broad goal is​​​‌ to consider brain-behavior problems‌ at the intersection of‌​‌ statistical physics, biomedical engineering,​​ and clinical neurosciences, that​​​‌ can be tackled using‌ complex systems theory. To‌​‌ this end, we propose​​ to create a new​​​‌ team focused on systems‌ neuroengineering, developing new‌​‌ analytical tools and technologies​​ to image, decode, and​​​‌ modulate the brain in‌ order to comprehend its‌​‌ functions and to repair​​ its dysfunction. Specifically, our​​​‌ team will tackle two‌ main scientific thrusts in‌​‌ the next five years:​​

  1. Analyzing, modeling and controlling​​​‌ multiscale brain networks.‌ Our ambition is to‌​‌ better understand the structural​​ and functional organization of​​​‌ the human brain. To‌ this end, we propose‌​‌ new computational frameworks to​​ characterize the spatio-temporal complexity​​​‌ of brain networks from‌ multimodal (e.g. structural or‌​‌ functional) and longitudinal neuroimaging​​​‌ data.
  2. Designing a new​ generation of noninvasive brain-computer​‌ interfaces (BCIs). There is​​ a critical need to​​​‌ improve the usability of​ BCIs. Our original approach​‌ consists in introducing network​​ methods into the BCI​​​‌ pipeline to better decode​ the user mental state,​‌ model the skill acquisition​​ process, and reinforce mental​​​‌ intention-related brain patterns via​ neuromodulation.

The methodological and​‌ technological developments resulting from​​ these goals, will be​​​‌ mainly applied to brain​ diseases, in close collaboration​‌ with neuroscientists and clinicians,​​ in order to: i)​​​‌ provide new insights into​ the associated neural reorganizational​‌ processes ii) identify network-based​​ biomarkers of disease and​​​‌ outcome and, iii) propose​ innovative network-based BCI neurorehabilitation​‌ strategies.

Thanks to the​​ close interaction with ICM,​​​‌ which physically hosts the​ NERV team, we aim​‌ to play a major​​ role in the complex​​​‌ systems and BCI community​ by capitalizing on the​‌ collaboration with other Inria/ICM​​ teams and clinical units,​​​‌ as well as on​ technological transfer of our​‌ scientific expertise to the​​ development of med-tech products.​​​‌

3 Research program

The​ NERV project-team has a​‌ two-fold reserch trhust. On​​ one hand, it develops​​​‌ methods from signal processing,​ complex systems, network science,​‌ to model and analyze​​ the interconnected nature of​​​‌ the nervous system. On​ the other hand, it​‌ developes noninvasive brain-computer interfaces​​ to allow humans interacting​​​‌ with the external world​ using brain activity.

3.1​‌ Analyzing, modeling and controlling​​ multiscale brain networks

3.1.1​​​‌ Multilayer analysis of multimodal​ brain networks

As in​‌ many other real complex​​ systems, the type of​​​‌ interactions between the regions​ of a same brain​‌ might be of different​​ nature, this giving rise​​​‌ to multiple networks between​ the same nodes (eg,​‌ structural, functional, etc). Despite​​ such one-to-many relationship, common​​​‌ network approaches have been​ traditionally conceived to analyze​‌ and model one single​​ type of connectivity. Our​​​‌ project-team develops novel network​ approaches to the case​‌ of multiple interconnected systems.​​ We specifically focus on​​​‌ the development of methods​ based on multilayer network​‌ theory to fully exploit​​ the rich multimodal nature​​​‌ of brain networks.

3.1.2​ Temporal models of dynamic​‌ brain networks

Current approaches​​ in network neuroscience assume​​​‌ static or time-invariant network​ that could not capture​‌ dynamical mechanisms, such as​​ the persistence or formation​​​‌ of specific connectivity patterns,​ which are instead crucial​‌ in time-varying networks. Another​​ crucial limitation is that​​​‌ standard approaches are basically​ data-driven, so that the​‌ obtained network indices lack​​ of confidence intervals thus​​​‌ making difficult the generalization​ of the observed results.​‌ NERV proposes to simultaneously​​ overcome these limitations by​​​‌ introducing novel model-based approaches​ that support statistical inference​‌ on the connection mechanisms​​ underlying the observed time-varying​​​‌ networks.

3.1.3 Theoretic controllability​ of brain networks

Controllability​‌ of networks refers to​​ the possibility of driving​​​‌ the current state of​ a system to a​‌ specific final state by​​ means of external control​​​‌ inputs. While encouraging results​ have been obtained in​‌ brain networks this field​​ remains quite largely unexplored.​​​‌ How expensive is to​ drive brain states? Can​‌ brain networks be steered​​ effectively from few nodes?​​ What type of input​​​‌ signal should be used?‌ Our project-team addresses the‌​‌ above questions and provide​​ a more robust framework​​​‌ that can be used‌ to identify intervention strategies‌​‌ facilitating desired behavior (eg,​​ learning, Stroke recovery) and​​​‌ counteracting clinical conditions (eg‌ Alzheimer’s disease, epilepsy).

3.1.4‌​‌ Latent geometry of brain​​ networks

Brain networks are​​​‌ a special type of‌ networks embedded in a‌​‌ physical space, so that​​ geometric network models, taking​​​‌ into account distance as‌ a costly factor in‌​‌ links formation, appear particularly​​ relevant. However, the existence​​​‌ of a possible latent‌ structure behind the observed‌​‌ brain network properties is​​ still poorly understood. Is​​​‌ there a hidden geometry‌ model that can explain‌​‌ the observed connectivity? Our​​ project-team explores non-Euclidean geometries​​​‌ to represent complex brain‌ networks and unveil hidden‌​‌ structural properties of the​​ system in a complementary​​​‌ and coherent way. We‌ apply these models to‌​‌ understand and unveil network​​ mechanisms in brain diseases​​​‌ across multiple scales.

3.2‌ Designing a new generation‌​‌ of efficient noninvasive BCIs​​

3.2.1 Enriching the features​​​‌ space of BCIs with‌ multimodal brain network metrics‌​‌

The research of alternative​​ features for improving BCI​​​‌ performance has been quite‌ limited and rather crude‌​‌ univariate features, such as​​ frequency band power or​​​‌ time point concatenation of‌ the brain signals, have‌​‌ been typically used. However,​​ the information contained in​​​‌ brain signal interactions across‌ multiple physiologically-relevant frequency bands‌​‌ has been surprisingly neglected.​​ NERV aims to improves​​​‌ BCI performance by enriching‌ the feature space by‌​‌ inlcuding brain connectivity and​​ network metrics. The project-team​​​‌ develops and text experimentally‌ these advances in both‌​‌ healthy and diseased subjects.​​

3.2.2 Informing adaptive BCIs​​​‌ through generative network models‌ of human learning

In‌​‌ a BCI, both the​​ human and the computer​​​‌ are part of the‌ same system. In such‌​‌ co-adaptive environment, it is​​ paramount that the computer​​​‌ could adapt to the‌ physiological nonstationarity of the‌​‌ brain features. However, classification​​ algorithms should adapt in​​​‌ practice when the user‌ is in the loop‌​‌ still need to be​​ clarified. NERV develops new​​​‌ statistical network models to‌ characterize temporally dynamic brain‌​‌ networks. By building these​​ new models, we first​​​‌ identify the dynamic brain‌ network properties that significantly‌​‌ predict with the BCI​​ skill acquisition. Once identified,​​​‌ these features will be‌ used to design innovative‌​‌ BCI architectures that take​​ into account the dynamics​​​‌ associated with the user’s‌ learning in an effort‌​‌ to improve the overall​​ performance.

3.2.3 Boosting BCI​​​‌ performance through targeted brain‌ stimulation

While richer brain‌​‌ features and an enhanced​​ understanding of the process​​​‌ of learning itself may‌ enhance BCI accuracy on‌​‌ average, challenges still remain​​ for single individuals. An​​​‌ alternative approach is to‌ draw on recent advances‌​‌ in noninvasive brain stimulation​​ technology, such as transcranial​​​‌ magnetic stimulation (TMS), which‌ can directly influence the‌​‌ brain plasticity by sending​​ an external signal that​​​‌ interfere locally and alter‌ the internal neural dynamics.‌​‌ NERV aims to leverage​​ theoretical network controllability models​​​‌ to experimentally validate the‌ ability of single brain‌​‌ regions to steer target​​​‌ brain areas towards BCI-related​ spatiotemporal activity. By means​‌ of this approach, we​​ aim to identify which​​​‌ brain areas and what​ type of input signal​‌ is needed to favor​​ BCI-related plasticity and improve​​​‌ performance of individual subjects.​

3.2.4 Towards multimodal and​‌ augmented noninvasive BCIs

While​​ BCI performance has been​​​‌ mainly thought as a​ classification or features extraction​‌ issue, current evidence suggests​​ that many other factors​​​‌ can actually affect the​ accuracy of the system.​‌ Among others, the use​​ of alternative physiological signals​​​‌ (ECG, EMG among others),​ can influence the motivation​‌ of the subjects, their​​ sense of agency and​​​‌ in turn their performance.​ Our project-team develops complementary​‌ approaches based on hybrid​​ signal control and enriched​​​‌ feedbacks, to create immersive​ and ecological BCI setups​‌ in an effort to​​ further ameliorate the BCI​​​‌ performance. This activity also​ explores a new technology​‌ based on optically pumped​​ magnetometers to have more​​​‌ accurate brain signals.

4​ Application domains

Our methodological​‌ and technological development will​​ be mainly applied to​​​‌ solve neuroscience-related problems that​ can be tackled using​‌ systems-level approaches. In line​​ with the Inria challenges​​​‌ plan, our project-team will​ contribute to modeling and​‌ simulation for digital health.​​ Thanks to the strategic​​​‌ position within ICM, we​ will also design and​‌ perform innovative experimental protocols​​ to gather data in​​​‌ humans and validate our​ theoretical outcome and tools.​‌ To this end, we​​ will capitalize on our​​​‌ long-lasting expertise in experimental​ data acquisition and protocol​‌ validation from national and​​ international ethical committees (eg,​​​‌ NIH, CPP, CNRS, Inserm,​ Coerle). In addition, we​‌ can rely on a​​ unique access to large​​​‌ cohorts of patients (eg,​ INSIGHT cohort for Alzheimer’s​‌ patients, stroke and epileptic​​ patients from the clinical​​​‌ units at the hospital)​ which can significantly contribute​‌ to the statistical value​​ of our discoveries.

4.1​​​‌ Network-based biomarkers of brain​ diseases

Accumulating evidence indicates​‌ that the symptoms of​​ a neurological disease are​​​‌ often associated with an​ abnormal organization of the​‌ connections in the brain.​​ Network science and complex​​​‌ systems theory provide therefore​ natural tools to analyze​‌ and model brain diseases​​ as well as to​​​‌ identify pathological reorganizational mechanisms.​ Our project-team will specifically​‌ focus on:

  • Stroke is​​ a medical condition in​​​‌ which parts of the​ brain die due to​‌ blood supply cut-offs, thus​​ leading to motor-cognitive impairments​​​‌ associated with the dead​ zone. Predicting the impact​‌ of a brain damage​​ and the ability of​​​‌ patients to restore their​ lost functions is a​‌ major issue in stroke​​ neuroscience. Because the brain​​​‌ is an interconnected system,​ stroke damages -which are​‌ local- will also have​​ effects on the rest​​​‌ of the network and​ induce global reconfiguration processes.​‌ By introducing original analytical​​ and modeling tools, we​​​‌ aim to better understand​ neural plasticity after stroke​‌ and derive brain network​​ signatures of functional recovery​​​‌ at individual level. To​ this end, we already​‌ collaborate with the Stroke​​ Unit at the Pitie-Salpetriere​​​‌ Hospital (PU-PH C Rosso)​ and with P Bartolomeo​‌ (INSERM DR) who is​​ a renowned expert in​​ clinical neuroscience at ICM.​​​‌

    While the NERV team‌ will focus on stroke,‌​‌ it will continue the​​ ongoing collaborations to terminate​​​‌ projects related to:

  • Epilepsy‌ is a group of‌​‌ neurological disorders characterized by​​ recurrent epileptic seizures, which​​​‌ can range from brief,‌ often imperceptible events to‌​‌ prolonged episodes of intense​​ convulsive activity caused by​​​‌ abnormal electrical discharges in‌ the brain. The detection‌​‌ of epileptic seizures, the​​ prediction of their onset,​​​‌ and the identification of‌ the most effective pharmacological‌​‌ treatment for individual patients​​ represent critical challenges from​​​‌ both fundamental and clinical‌ perspectives. In our research‌​‌ works, we address these​​ challenges by adopting an​​​‌ original framework based on‌ network science, aiming to‌​‌ improve our understanding of​​ the role of underlying​​​‌ brain connectivity in the‌ generation and propagation of‌​‌ seizures and to identify​​ more accurate and reliable​​​‌ biomarkers. To this end,‌ we benefit from established‌​‌ collaborations with the Epilepsy​​ Unit at the Pitie-Salpetriere​​​‌ Hospital (PU-PH V Navarro)‌ and with the Universitary‌​‌ Hospital, in Strasbourg (MCU-PH​​ V Dinkelacker).
  • Neurodegeneration is​​​‌ caused by the progressive‌ loss of structure or‌​‌ function of neurons, which​​ leads to a range​​​‌ of cognitive and motor‌ impairments, from mild to‌​‌ severe, and eventually to​​ death. Most of research​​​‌ currently focuses on predicting‌ as soon as possible‌​‌ those individuals who will​​ develop the disease, so​​​‌ as to adopt the‌ best therapeutics and slow‌​‌ the disease progression. By​​ adopting a network perspective,​​​‌ we aim to understand‌ the abnormal reorganizational connection‌​‌ processes behind the disease​​ and provide alternative biomarkers​​​‌ that can be integrated‌ with existing ones (eg‌​‌ atrophy, behavioral, metabolic) to​​ improve the accuracy prediction.​​​‌ To this end, we‌ already collaborate with the‌​‌ ARAMIS team and with​​ the Experimental Neurosurgery team​​​‌ at ICM (Parkinson, B‌ Lau CRNS and C‌​‌ Karachi MCU-PH) and we​​ have long-lasting collaborations with​​​‌ the Institut de la‌ memoire et de l‌​‌ Alzheimer (IM2A) at the​​ Pitie-Salpetriere hospital (PU-PH Prof​​​‌ B Dubois).

4.2 Improving‌ BCI efficiency for clinical‌​‌ applications

Enhancing the accuracy​​ of the BCI performance​​​‌ not only has a‌ fundamental interest, but it‌​‌ also has practical consequences.​​ Better decoding the user’s​​​‌ mental intent means better‌ understanding the underlying neural‌​‌ process and transform it​​ in more reliable external​​​‌ commands. Our project-team aims‌ to unlock BCI clinical‌​‌ applications by specifically unveiling​​ new network connectivity features​​​‌ of brain functioning. We’ll‌ specifically focus on:

  • Neuromotor‌​‌ rehabilitation aims to aid​​ recovery from a nervous​​​‌ system injury (eg, stroke,‌ Parkinson) and to minimize/compensate‌​‌ for motor alterations resulting​​ from it. Identifying the​​​‌ best rehabilitation strategy for‌ each patient is a‌​‌ major challenge as it​​ significantly affects the quality​​​‌ of recovery. By developing‌ high-performance BCI prototypes we‌​‌ aim to introduce innovative​​ intervention strategies that ease​​​‌ the neuromotor recovery process‌ through noninvasive neurofeedback experimental‌​‌ protocols. To this end,​​ we already collaborate with​​​‌ the Experimental Neurosurgery team‌ at ICM for the‌​‌ application to Parkinson’s subjects​​ (ANR Betapark) and we​​​‌ are in the process‌ of establishing new collaborations,‌​‌ within the ERC BCINET,​​​‌ with important stroke neurorehabilitation​ units at the Pitie-Salpetriere​‌ Hospital (P Pradat, AP-HP)​​ and at the Saint-Maurice​​​‌ Hospital (F Colle, AP-HP).​

    While the NERV team​‌ will focus on Neuromotor​​ rehabilitation, it will continue​​​‌ the ongoing collaboration to​ terminate projects related to:​‌

  • Brain monitoring aims to​​ detect events associated with​​​‌ mental states that emerge​ from background ongoing brain​‌ activity (eg, stress or​​ epileptic dynamics). The accuracy​​​‌ of their detection is​ crucial to decide and​‌ execute the most appropriate​​ action from the computer.​​​‌ Our project-team aims to​ fine-tune and optimize the​‌ innovative BCI prototypes for​​ real-time target applications. To​​​‌ this end, we already​ have collaborated with Air​‌ Liquide Medical Systems and​​ the Pitié-Salpêtrière Hospital (Intensive​​​‌ Care Units, PU-PH T​ Similowski, M Raux), for​‌ the development of an​​ EEG-based BCI which detects​​​‌ respiratory discomforts in ventilated​ patients (patent WO 2013/164462).​‌ Also, we have consolidated​​ collaborations with the start-up​​​‌ Mybraintech (ICM spin-off) to​ develop portable BCIs for​‌ predicting mental stress (CIFRE​​ partnership), and with IBM-France​​​‌ and Armand-Troussaud hospital (MD-PH​ AI Vermersch) to develop​‌ EEG-based aid diagnostics tool​​ for predicting newborn hypoxia.​​​‌
  • Assistive technology is used​ to increase or maintain​‌ the functional capabilities of​​ disabled people (eg, wheelchairs,​​​‌ prostheses). Assistive BCIs represent​ therefore a promising tool​‌ for allowing users to​​ control external devices directly​​​‌ with their brain. Although​ our project-team is more​‌ focused on rehabilitative BCIs,​​ the development of high-performance​​​‌ BCIs can unlock assistive​ BCI applications, too. To​‌ this end, we have​​ recently started a collaboration​​​‌ with the ISIR lab​ (LIP6) for the design​‌ of a multimodal BCI​​ prototype that controls a​​​‌ robotic arm and grab​ objects in a 3D​‌ space (PhD T Venot).​​ We next aim to​​​‌ integrate augmented reality and​ enriched feedbacks (virtual hands)​‌ to render more ecological​​ environments and improve the​​​‌ sense of agency of​ patients.

5 Highlights of​‌ the year

5.1 Awards​​

  • Marie-Constance Corsi received the​​​‌ Early Career Award by​ the BCI Society
  • Marc​‌ Fiammante received the Prix​​ Félix Innovateur from Central​​​‌ Supélec Alumni

5.2 Startup​

  • Marc Fiammante was accepted​‌ at the “Pepinière Paris​​ Santé” program of the​​​‌ Cochin Hospital (Paris) to​ create his startup, focused​‌ on neonatal brain monitoring.​​
  • Tristan Venot joined the​​​‌ Inria Startup Studio program​ to create the startup​‌ Cirus, focusing on brain-computer​​ interfaces.

6 Latest software​​​‌ developments, platforms, open data​

6.1 Latest software developments​‌

6.1.1 HappyFeat

  • Keywords:
    BCI,​​ Connectivity, Brain-Computer Interface, Classification,​​​‌ GUI (Graphical User Interface),​ Signal processing, Biomedical data​‌
  • Scientific Description:

    Two main​​ use-cases are targeted: -​​​‌ Using MI in a​ clinical setting (e.g. stroke​‌ rehabilitation), by greatly reducing​​ the risks of mistakes​​​‌ during the offline analysis​ and the time needed​‌ to perform this step,​​ quickly bridging the gap​​​‌ between EEG data acquisition​ and online BCI usage.​‌

    - Exploring new, alternative​​ metrics for discriminating between​​​‌ mental states. To this​ aim, prototypes for prospective​‌ methods need to be​​ validated on signal databases,​​​‌ before moving on to​ experimental conditions. HappyFeat helps​‌ bridging this gap, and​​ provides a framework in​​ which such methods can​​​‌ be tested, after implementation.‌

    The targetted audience is‌​‌ therefore on the one​​ hand clinicians, neurophysiologists and​​​‌ neuroscientists who want to‌ use BCI in their‌​‌ programs, but also the​​ research community in BCI,​​​‌ brain networks, and functional‌ connectivity.

  • Functional Description:

    Brain‌​‌ Computer Interfaces (BCI) has​​ a strong potential in​​​‌ clinical applications such as‌ post-stroke rehabilitation. However, in‌​‌ such constrained contexts, obtaining​​ satisfactory performances from a​​​‌ BCI system can be‌ a challenging task, mainly‌​‌ due to the need​​ for fine-tuning the classification​​​‌ algorithm used to distinguish‌ between mental tasks. Training‌​‌ the classification algorithm with​​ adequately selected features obtained​​​‌ from spectral analysis or‌ other alternative metrics is‌​‌ crucial.

    HappyFeat is a​​ software assistant for feature​​​‌ extraction and selection in‌ BCI. It proposes a‌​‌ trial-and-error oriented workflow, where​​ experimenters can extract, visualize​​​‌ and select features of‌ interest for training as‌​‌ many times as needed,​​ in a short time,​​​‌ until a satisfying classification‌ training accuracy is reached.‌​‌

    Every operation from signal​​ loading and feature extraction​​​‌ to classifier training is‌ handled from a unified,‌​‌ dashboard-like GUI, removing the​​ need to use different​​​‌ softwares for data acquisition,‌ feature analysis, classifier training‌​‌ and online classification, and​​ to manage data formatting​​​‌ across the different environments.‌

    Along with the commonly‌​‌ used Power Spectral Density​​ (PSD), HappyFeat enables to​​​‌ work with Functional Connectivity,‌ allowing to use novel‌​‌ network-based approaches based on​​ recent research.

    HappyFeat uses​​​‌ OpenViBE in the background‌ for the extraction and‌​‌ training parts, as a​​ fast and efficient processing​​​‌ engine, taking advantage of‌ its optimized C++ implementation‌​‌ of signal processing methods.​​ The generation and manipulation​​​‌ of use-case scenarios is‌ entirely automated via scripts‌​‌ and templates, removing the​​ inherent risk of mistakes​​​‌ in a time constrained‌ environment.

    HappyFeat puts the‌​‌ emphasis on reproducibility, by​​ keeping track of all​​​‌ manipulations (EEG sessions file‌ lists, signal processing, classification‌​‌ attempts) and allowing to​​ save, load and export​​​‌ previous work.

  • Release Contributions:‌

    - Support for Timeflux,‌​‌ with a 2-class MI​​ protocol based on PSD​​​‌ (with Welch's method) -‌ Tutorial for newcomers, using‌​‌ Timeflux - Dependencies can​​ be installed with conda​​​‌ - Python 3.12.8, dependencies‌ update - Parameters for‌​‌ the AutoFeat mechanism can​​ be set via menus​​​‌ - Most visualization tools‌ use plotly, and figures‌​‌ are saved in the​​ workspace

    - GUI fixes​​​‌ - Templates updated -‌ Fixed various crash sources‌​‌ & stability issues

  • URL:​​
  • Publication:
  • Contact:​​​‌
    Arthur Desbois
  • Participants:
    Arthur‌ Desbois, Marie-Constance Corsi, Fabrizio‌​‌ De Vico Fallani

6.1.2​​ VIZAJ

  • Name:
    A free​​​‌ online interactive software for‌ visualizing spatial networks
  • Keywords:‌​‌
    Complex Systems, Data visualization​​
  • Functional Description:
    In many​​​‌ fields of science and‌ technology we are confronted‌​‌ with complex networks. Making​​ sense of these networks​​​‌ often require the ability‌ to visualize and explore‌​‌ their intermingled structure consisting​​ of nodes and links.​​​‌ To facilitate the identification‌ of significant connectivity patterns,‌​‌ many methods have been​​ developed based on the​​​‌ rearrangement of the nodes‌ so as to avoid‌​‌ link criss-cross. However, real​​​‌ networks are often embedded​ in a geometrical space​‌ and the nodes code​​ for an intrinsic physical​​​‌ feature of the system​ that one might want​‌ to preserve. For these​​ spatial networks, it is​​​‌ therefore crucial to find​ alternative strategies operating on​‌ the links and not​​ on the nodes. Here,​​​‌ we introduce Vizaj a​ javascript web application to​‌ render spatial networks based​​ on optimized geometrical criteria​​​‌ that reshape the link​ profiles. While optimized for​‌ 3D networks, Vizaj can​​ also be used for​​​‌ 2D networks and offers​ the possibility to interactively​‌ customize the visualization via​​ several controlling parameters, including​​​‌ network filtering and the​ effect of internode distance​‌ on the link trajectories.​​ Vizaj is further equipped​​​‌ with additional options allowing​ to improve the final​‌ aesthetics, such as the​​ color/size of both nodes​​​‌ and links, zooming/rotating/translating, and​ superimposing external objects. Vizaj​‌ is an open-source software​​ which can be freely​​​‌ downloaded and updated via​ a github repository. Here,​‌ we provide a detailed​​ description of its main​​​‌ features and algorithms together​ with a guide on​‌ how to use it.​​ Finally, we validate its​​​‌ potential on several synthetic​ and real spatial networks​‌ from infrastructural to biological​​ systems. We hope that​​​‌ Vizaj will help scientists​ and practitioners to make​‌ sense of complex networks​​ and provide aesthetic while​​​‌ informative visualizations.
  • URL:
  • Publication:
  • Contact:
    Fabrizio​‌ De Vico Fallani

6.2​​ New platforms

6.2.1 Noninvasive​​​‌ brain-computer interfaces (BCI)

Participants:​ Marie-Constance Corsi, Arthur​‌ Desbois, Tristan Venot​​, Laurent Bougrain,​​​‌ Laurent Hugueville, Fabrizio​ De Vico Fallani [Correspondant]​‌.

NERV coordinates the​​ research and development activity​​​‌ of the Brain-Computer Interface​ (BCI) platform at the​‌ Centre EEG/MEG of the​​ neuroimaging core facility of​​​‌ the ICM. The R&D​ activity consists in assembling,​‌ testing different hardware, software​​ developments to ensure the​​​‌ highest reliability and performance,​ as well as to​‌ test innovative technological solutions.​​ Several projects, including our​​​‌ NETBCI NIH/ANR, MANET ANR-JCJC,​ and ATTACK Big-brain theory​‌ funded projects, as well​​ as experiments by different​​​‌ researchers of the Institute​ (ANR BETAPARK Project), and​‌ the BCINET ERC Consolidator​​ grant (F De Vico​​​‌ Fallani) are currently being​ run.

The BCI experimental​‌ platform is closely linked​​ to the development of​​​‌ the Happyfeat software by​ our engineer A. Desbois​‌ (see Software section). HappyFeat​​ is part of a​​​‌ larger INRIA BCI software​ suite together with OpenVibe​‌ (HYBRID team) and BCIVizApp​​ (CHRONOS Temporal). HappyFeat allows​​​‌ to easily intergate new​ functionalities based on our​‌ methodological development on brain​​ connectivity networks and integrates​​​‌ efficient graphical user interfaces​ for easy use by​‌ clinicians.

7 New results​​

7.1 Low-dimensional controllability of​​​‌ brain networks

Participants: Remy​ Ben Messaoud, Camile​‌ Bousfiha, Marie-Constance Corsi​​, Mario Chavez,​​​‌ Fabrizio de Vico Fallani​ [Correspondant].

Identifying the​‌ driver nodes of a​​ network has crucial implications​​​‌ in biological systems from​ unveiling causal interactions to​‌ informing effective intervention strategies.​​ Despite recent advances in​​​‌ network control theory, results​ remain inaccurate as the​‌ number of drivers becomes​​ too small compared to​​ the network size, thus​​​‌ limiting the concrete usability‌ in many real-life applications.‌​‌ To overcome this issue,​​ we introduced a framework​​​‌ that integrates principles from‌ spectral graph theory and‌​‌ output controllability to project​​ the network state into​​​‌ a smaller topological space‌ formed by the Laplacian‌​‌ network structure. Through extensive​​ simulations on synthetic and​​​‌ real networks, we showed‌ that a relatively low‌​‌ number of projected components​​ can significantly improve the​​​‌ control accuracy. By introducing‌ a new low-dimensional controllability‌​‌ metric we experimentally validated​​ our method on N​​​‌ = 6134 human connectomes‌ obtained from the UK-biobank‌​‌ cohort. Results revealed previously​​ unappreciated influential brain regions,​​​‌ enabled to draw directed‌ maps between differently specialized‌​‌ cerebral systems, and yielded​​ new insights into hemispheric​​​‌ lateralization. Taken together, our‌ results offered a theoretically‌​‌ grounded solution to deal​​ with network controllability and​​​‌ provided insights into the‌ causal interactions of the‌​‌ human brain.

More details​​ in 5.

7.2​​​‌ Interpretability of Riemannian tools‌ used in brain computer‌​‌ interfaces

Participants: Tristan Venot​​, Marie-Constance Corsi [Correspondant]​​​‌.

Riemannian methods have‌ established themselves as stateof-the-art‌​‌ approaches in Brain-Computer Interfaces​​ (BCI) in terms of​​​‌ performance. However, their adoption‌ by experimenters is often‌​‌ hindered by a lack​​ of interpretability. In this​​​‌ work, we propose a‌ set of tools designed‌​‌ to enhance practitioners' understanding​​ of the decisions made​​​‌ by Riemannian methods. Specifically,‌ we develop techniques to‌​‌ quantify and visualize the​​ influence of the different​​​‌ sensors on classification outcomes.‌ Our approach includes a‌​‌ visualization tool for high-dimensional​​ covariance matrices, a classifieragnostic​​​‌ tool that focuses on‌ the classification process, as‌​‌ well as methods that​​ leverage the data's topology​​​‌ to better characterize the‌ role of each sensor.‌​‌ We demonstrate these tools​​ on a specific dataset​​​‌ and provide Python code‌ to facilitate their use‌​‌ by practitioners, thereby promoting​​ the adoption of Riemannian​​​‌ methods in BCI.

More‌ details in 31.‌​‌

7.3 Automatic Ocular Artifact​​ Correction in Electroencephalography for​​​‌ Neurofeedback

Participants: Cassandra Dumas‌, Marie-Constance Corsi [Correspondant]‌​‌.

Ocular artifacts can​​ significantly impact electroencephalography (EEG)​​​‌ signals, potentially compromising the‌ performance of neurofeedback (NF)‌​‌ and brain-computer interfaces (BCI)​​ based on EEG. This​​​‌ study investigates if the‌ Approximate Joint Diagonalization of‌​‌ Fourier Cospectra (AJDC) method​​ can effectively correct blink-related​​​‌ artifacts and preserve relevant‌ neurophysiological signatures in a‌​‌ pseudo-online context. AJDC is​​ a frequency-domain Blind Source​​​‌ Separation (BSS) technique, which‌ uses cospectral analysis to‌​‌ isolate and attenuate blink​​ artifacts. Using EEG data​​​‌ from 21 participants recorded‌ during a NF motor‌​‌ imagery (MI) task, we​​ compared AJDC with Independent​​​‌ Component Analysis (ICA), a‌ widely used method for‌​‌ EEG denoising. We assessed​​ the quality of blink​​​‌ artifact correction, the preservation‌ of MI-related EEG signatures,‌​‌ and the influence of​​ AJDC correction on the​​​‌ NF performance indicator. We‌ show that AJDC effectively‌​‌ attenuates blink artifacts without​​ distorting MI-related beta band​​​‌ signatures and with preservation‌ of NF performance. AJDC‌​‌ was calibrated once on​​ initial EEG data. We​​​‌ therefore assessed AJDC correction‌ quality over time, showing‌​‌ some decrease. This suggests​​​‌ that periodic recalibration may​ benefit long EEG recording.​‌ This study highlights AJDC​​ as a promising real-time​​​‌ solution for artifact management​ in NF, with the​‌ potential to provide consistent​​ EEG quality and to​​​‌ enhance NF reliability.

More​ details in 37.​‌

7.4 Riemannian fusions of​​ EEG-based features for motor​​​‌ imagery detection under propofol​ sedation

Participants: Camilla Mannino​‌, Marie-Constance Corsi,​​ Laurent Bougrain [Correspondant].​​​‌

The brain is a​ complex system requiring multimodal​‌ approaches to better understand​​ cognitive or motor functions.​​​‌ Thus, different and complementary​ electroencephalographic (EEG) neurophysiological features​‌ are available at various​​ spatial, frequency, and temporal​​​‌ scales, e.g., brain connectivity,​ complexity, or entropy. However,​‌ they are usually not​​ investigated all together. In​​​‌ this study, we combine​ and compare five EEG-based​‌ connectivity features with covariance​​ matrices, defining five Riemannian​​​‌ fusion methods and three​ Euclidean ones as references.​‌ We do so for​​ classifying motor imagery EEG​​​‌ signals, both in awake​ and sedated subjects, with​‌ the future goal of​​ detecting accidental awareness during​​​‌ general anesthesia. Covariance matrices​ alone yielded the best​‌ accuracy, with and without​​ sedation. Phase-based connectivity estimators​​​‌ appear to be the​ most promising fusion with​‌ covariances. No significant differences​​ were found between the​​​‌ best fusion of features​ and that of classifiers.​‌

More details in 30​​.

7.5 Linking heartbeats​​​‌ with the cortical network​ dynamics involved in self-social​‌ touch distinction

Participants: Fabrizio​​ De Vico Fallani,​​​‌ Diego Candia-Rivera [Correspondant].​

Research on interoception has​‌ revealed the role of​​ heartbeats in shaping our​​​‌ perceptual awareness and embodying​ a first-person perspective. These​‌ heartbeat dynamics exhibit distinct​​ responses to various types​​​‌ of touch. We advanced​ that those dynamics are​‌ directly associated to the​​ brain activity that allows​​​‌ self-other distinction. In our​ study encompassing self and​‌ social touch, we employed​​ a method to quantify​​​‌ the distinct couplings of​ temporal patterns in cardiac​‌ sympathetic and parasympathetic activities​​ with brain connectivity. Our​​​‌ findings revealed that social​ touch led to an​‌ increase in the coupling​​ between frontoparietal networks and​​​‌ parasympathetic/vagal activity, particularly in​ alpha and gamma bands.​‌ Conversely, as social touch​​ progressed, we observed a​​​‌ decrease in the coupling​ between brain networks and​‌ sympathetic dynamics across a​​ broad frequency range. These​​​‌ results show how heartbeat​ dynamics are intertwined with​‌ brain organization and provide​​ fresh evidence on the​​​‌ neurophysiological mechanisms of self-social​ touch distinction.

More details​‌ in 13.

7.6​​ Assessment of a learner's​​​‌ mental state: search for​ EEG markers that can​‌ distinguish fluctuations in sustained​​ attention and cognitive engagement​​​‌

Participants: Pierre-Baptiste Mathieu de​ Carvalho, Marie-Constance Corsi​‌, Laurent Bougrain [Correspondant]​​.

We aimed to​​​‌ establish the methodological foundations​ for distinguishing, using EEG,​‌ between sustained attention and​​ cognitive engagement in a​​​‌ learning context. This preliminary​ work based on a​‌ large review made it​​ possible to explore and​​​‌ test the relevance of​ the approach and the​‌ robustness of an experimental​​ protocol approved by the​​​‌ Operational Committee for the​ Evaluation of Legal and​‌ Ethical Risks of Inria​​ (COERLE 2025-66). The main​​ finding of this study,​​​‌ a priori, lies in‌ the heterogeneity of individual‌​‌ profiles: while attention mechanisms​​ seem to follow a​​​‌ common logic, the way‌ in which individuals engage‌​‌ cognitively appears to be​​ a more personal strategy.​​​‌ This variability, which will‌ need to be monitored‌​‌ on a larger sample,​​ offers serious avenues for​​​‌ further research. Indeed, it‌ suggests that the development‌​‌ of personalized approaches could​​ be a promising alternative​​​‌ to the search for‌ universal markers. Thus, by‌​‌ capturing the dynamics specific​​ to each learner, this​​​‌ future work paves the‌ way for the development‌​‌ of neuro-adaptive interaction loops,​​ which will ultimately be​​​‌ able to assist learning‌ more precisely or improve‌​‌ BCI control, as envisaged​​ in the introduction.

More​​​‌ details in 45.‌

7.7 Median nerve stimulation‌​‌ to assess Motor Imagery-BCI​​ performances

Participants: Laurent Bougrain​​​‌ [Correspondant].

Motor Imagery-based‌ Brain-Computer Interfaces (MI-BCIs) enable‌​‌ device control through ElectroEncephaloGraphy​​ (EEG), yet intra- and​​​‌ inter-subject variability remains a‌ critical challenge affecting system‌​‌ reliability. Median Nerve Stimulation​​ (MNS) has emerged as​​​‌ a promising alternative motor‌ task, but its variability‌​‌ characteristics and predictive value​​ require systematic investigation. This​​​‌ study quantifies EEG variability‌ in MNS-induced Event-Related Desynchronization‌​‌ (ERD) compared to MI,​​ and evaluates MNS-ERD as​​​‌ a performance predictor using‌ Linear Discriminant Analysis (LDA)‌​‌ and Least Absolute Shrinkage​​ and Selection Operator (LASSO).​​​‌ Results demonstrate that MI‌ elicits stronger ERD with‌​‌ lower intra-subject variability than​​ MNS, while inter-subject variability​​​‌ remains comparable between tasks.‌ For performance prediction, LDA‌​‌ and LASSO achieved 74%​​ accuracy for two-group classification​​​‌ (low vs. high performers),‌ with hierarchical clustering reaching‌​‌ 83% accuracy. Topographical analyses​​ revealed enhanced motor cortex​​​‌ activation in high performers‌ during both tasks. These‌​‌ findings establish MNS-induced ERD​​ as a reliable, non-invasive​​​‌ predictor for early user‌ stratification while providing quantitative‌​‌ insights into EEG variability​​ patterns essential for personalized​​​‌ BCI design and applications‌ including intraoperative awareness monitoring.‌​‌ Two workshops have been​​ co-organized linked with this​​​‌ topic (see 10.1.1)‌

More details in 25‌​‌, 27, 29​​.

7.8 Connectivity-based prediction​​​‌ of the surgery outcome‌ in temporal lobe epilepsy‌​‌

Participants: Martin Guillemaud,​​ Mario Chavez [Correspondant].​​​‌

Epilepsy surgery is a‌ key treatment for patients‌​‌ with drug-resistant temporal lobe​​ epilepsy (TLE), yet predicting​​​‌ surgical outcomes remains challenging.‌ We introduce a novel‌​‌ connectivity-based biomarker derived from​​ structural brain network changes​​​‌ induced by surgery, analyzed‌ using hyperbolic graph embeddings.‌​‌ Using structural and diffusion​​ MRI data from 51​​​‌ patients, we compared pre-‌ and post-surgical connectivity networks‌​‌ and applied hyperbolic Poincaré​​ disk embeddings to distinguish​​​‌ favorable from poor outcomes.‌ The approach identified connectivity‌​‌ patterns in contralateral brain​​ regions as potential biomarkers​​​‌ of surgical success. Model‌ validation using leave-one-out cross-validation‌​‌ yielded an AUC of​​ 0.86 and a balanced​​​‌ accuracy of 0.81, demonstrating‌ strong predictive performance. These‌​‌ results highlight the potential​​ of non-Euclidean network embeddings​​​‌ to improve personalized outcome‌ prediction in TLE surgery.‌​‌

More details in 21​​.

8 Bilateral contracts​​​‌ and grants with industry‌

8.1 Bilateral contracts with‌​‌ industry

8.1.1 CIFRE PhD​​​‌ - Reliev Technology

Participants:​ Mario Chavez [Correspondant].​‌

  • Partner
    : Startup Reliev​​ Technology (Nantes)
  • Description
    :​​​‌ This project aims at​ developing a non-invasive multimodal​‌ system for predicting the​​ risk of epileptic seizures,​​​‌ based on artificial intelligence,​ which will be integrated​‌ into a continuous monitoring​​ system in patients allowing​​​‌ the acquisition in ambulatory​ mode.
  • Coordinator
    : Mario​‌ Chavez
  • Duration
    : 3​​ years

8.1.2 CIFRE PhD​​​‌ - Thales

Participants: Marion​ Pavaux [Correspondant].

  • Partner​‌
    : Thales (Paris-Saclay)
  • Description​​
    : Brain–Computer Interfaces (BCIs)​​​‌ create a direct link​ between brain activity, often​‌ recorded via EEG, and​​ a computer. They are​​​‌ used for assistance, rehabilitation,​ and remote control, but​‌ their performance remains limited​​ due to reliance on​​​‌ univariate brain measures. Analyzing​ brain connectivity networks shows​‌ promise, yet current methods​​ are too computationally demanding​​​‌ for real-time use. This​ project explores deep learning​‌ architectures designed to produce​​ universal, stable, and efficient​​​‌ EEG representations for real-time​ brain activity decoding.
  • Coordinator​‌
    : Fabrizio De Vico​​ Fallani
  • Duration
    : 3​​​‌ years

8.1.3 CIFRE PhD​ - EssilorLuxottica

Participants: Baptiste​‌ Fague [Correspondant].

  • Partner​​
    : EssilorLuxottica (Paris)
  • Description​​​‌
    : This thesis project​ first aims to better​‌ understand the relationship between​​ brain response and visual​​​‌ perception through cylindrical corrective​ lenses, a common optical​‌ device used to correct​​ astigmatism. Secondly, leveraging advances​​​‌ in EEG technology, we​ aim to develop an​‌ automated method for determining​​ an individual’s refractive error​​​‌ by measuring only their​ brain response.
  • Coordinator
    :​‌ Marie-Constance Corsi
  • Duration
    :​​ 3 years

9 Partnerships​​​‌ and cooperations

9.1 International​ initiatives

9.1.1 Participation in​‌ other International Programs

FACE​​ Foundation - FR-US partnership​​​‌

Participants: Marie-Constance Corsi.​

  • Project title:
    Biophysical modeling​‌ to inform Brain-Computer Interface​​ learning mechanisms
  • Partner:
    University​​​‌ of California, San Francisco​ (UCSF)
  • Date/Duration:
    2 years​‌
  • Amount:
    20keuros
  • Coordinators:
    Parul​​ Verma (formely postdoc at​​​‌ UCSF, now at IIT​ Madras, India), & Marie-Constance​‌ Corsi
  • Summary:
    BCI is​​ a promising tool for​​​‌ patients who suffer from​ neuromuscular pathologies or lesions.​‌ Nevertheless, it fails to​​ detect intents in 30%​​​‌ of the BCI users,​ even after several weeks​‌ of training. To circumvent​​ it, it is crucial​​​‌ to better understand the​ mechanisms underlying the BCI​‌ training. In this project,​​ we aim at using​​​‌ the spectral graph model​ (SGM) developed by the​‌ US project leader’s lab​​ to identify biophysical changes​​​‌ occurring while controlling a​ BCI. SGM captures the​‌ relationship between brain structure​​ and brain function with​​​‌ a reduced number of​ interpretable parameters. We will​‌ apply SGM to a​​ longitudinal BCI dataset collected​​​‌ by the French project​ leader. We will fully​‌ explore the potentiality of​​ this approach to identify​​​‌ biophysical markers that inform​ the neural mechanisms underlying​‌ the BCI training. Such​​ insights could pave the​​​‌ way to tailored BCI​ training programs.

9.2 International​‌ research visitors

9.2.1 Visits​​ of international scientists

Other​​​‌ international visits to the​ team
Giovanni Messuti
  • PhD​‌ student
  • Institution of origin:​​
    Università degli Studi di​​​‌ Salerno
  • Country:
    Italy
  • Dates:​
    from November 2025 to​‌ May 2026
  • Context of​​ the visit:
    The project​​ aims to leverage a​​​‌ multimodal framework integrating both‌ EEG and MEG, which‌​‌ could (i) improve decoding​​ accuracy in an interpretable​​​‌ way, and (ii) shed‌ light on the neural‌​‌ mechanisms underlying the learning​​ process during BCI training.​​​‌ This is the first‌ attempt to use M/EEG‌​‌ data within a latent​​ space framework to improve​​​‌ BCI classification and to‌ study patterns that arise‌​‌ while learning to control​​ a BCI system. This​​​‌ project marks the beginning‌ of our collaboration with‌​‌ Prof. S. Scarpetta (Università​​ degli Studi di Salerno)​​​‌ on the application of‌ biophysics tools in neuroscience.‌​‌
  • Type of mobility:
    Research​​ stay
  • Correspondant:
    Marie-Constance Corsi​​​‌

9.2.2 Visits to international‌ teams

Research stays abroad‌​‌
Laurent Bougrain
  • Visited institution:​​
    National University of Entre​​​‌ Ríos
  • Country:
    Argentina
  • Dates:‌
    18-22 August 2025
  • Context‌​‌ of the visit:
    Collaboration​​ on Brain-computer interfaces for​​​‌ stroke patients and transfer‌ learning for future international‌​‌ calls
  • Mobility program/type of​​ mobility:
    (sabbatical, internship, research​​​‌ stay, lecture…) Invited professor‌ program from the National‌​‌ University of Entre Ríos​​
Laurent Bougrain
  • Visited institution:​​​‌
    Kyushu Institute of Technology‌ (Kyutech)
  • Country:
    Japan
  • Dates:‌​‌
    7-11 April 2025
  • Context​​ of the visit:
    2023-2025​​​‌ Collaborative project Université de‌ Lorraine on "Human/Robot Social‌​‌ Interactions: engagement and affect​​ analysis during gaming tasks​​​‌
  • Mobility program/type of mobility:‌
    Erasmus+ mobility
Tristan Venot‌​‌
  • Visited institution:
    Kyushu Institute​​ of Technology (Kyutech)
  • Country:​​​‌
    Japan
  • Dates:
    3-12 February‌ 2025
  • Context of the‌​‌ visit:
    program for students​​
  • Mobility program/type of mobility:​​​‌
    JST Sakura Science Program‌ (Japan)
Pierre-Baptiste Mathieu de‌​‌ Carvalho
  • Visited institution:
    Kyushu​​ Institute of Technology (Kyutech)​​​‌
  • Country:
    Japan
  • Dates:
    7-11‌ April 2025
  • Context of‌​‌ the visit:
    2023-2025 Collaborative​​ project Université de Lorraine​​​‌ on "Human/Robot Social Interactions:‌ engagement and affect analysis‌​‌ during gaming tasks
  • Mobility​​ program/type of mobility:
    2023-2025​​​‌ Collaborative project Université de‌ Lorraine on "Human/Robot Social‌​‌ Interactions: engagement and affect​​ analysis during gaming tasks​​​‌

9.3 European initiatives

9.3.1‌ Horizon Europe

MSCA NETCORE‌​‌

Participants: Diego Candia-Rivera [Correspondant]​​.

  • Project title:
    Biomarkers​​​‌ of the interplay between‌ brain networks and cardiac‌​‌ dynamics for the evaluation​​ of non-invasive brain-computer interfaces​​​‌
  • Duration:
    2024-2026
  • Amount:
    212k€‌
  • Coordinator:
    DiegoCandia-Rivera
  • Otherpartners:
    ICM‌​‌
  • Summary:
    Brain-computerinterfaces(BCI) hold promise​​ in the restoration of​​​‌ lost sensorimotor abilities after‌ stroke, a leading cause‌​‌ of disability. Yet, their​​ effectiveness varies because BCI​​​‌ typically need to be‌ customized for each patient.‌​‌ Our innovative methodology focuses​​ on the brain-heart interplay​​​‌ and combines network science‌ and biomedical signal processing‌​‌ to estimate interactions between​​ these two systems in​​​‌ the context of motor‌ imagery. We will explore‌​‌ various approaches, such as​​ generative data methods, multi-layer​​​‌ networks, higher-order dependencies, and‌ deducing potential causal interactions‌​‌ from physiologically informed neural​​ models. Our ultimate goal​​​‌ is to pave the‌ way for future biomedical‌​‌ breakthroughs in the emerging​​ field of brain-heart interplay.​​​‌ Through these efforts, NETCORE‌ strives to enhance the‌​‌ potential of BCI in​​ aiding brain-injured patients and​​​‌ showing the potential of‌ studying brain-heart interplay in‌​‌ healthcare and neuroscientific research.​​

9.3.2 H2020 projects

ERC​​​‌ BCINET

Participants: Fabrizio De‌ Vico Fallani [Correspondant].‌​‌

  • Project title:
    Non-invasive decoding​​​‌ of brain communication patterns​ to ease motor restoration​‌ after stroke
  • Duration:
    2020-2027​​
  • Amount:
    2M€
  • Coordinator:
    Fabrizio,​​​‌ De Vico Fallani
  • Otherpartners:​
    ICM
  • Summary:
    Brain–computer interfaces​‌ (BCIs) can bypass the​​ skeletomuscular system, assisting paralysed​​​‌ people in control and​ communication. However, despite their​‌ application in neuromotor rehabilitation,​​ the accuracy of sensory​​​‌ feedback is still highly​ variable, limiting their use​‌ in everyday life. Scientists​​ of the EU-funded BCINET​​​‌ project propose to address​ this issue through a​‌ novel generation of BCIs​​ that do not solely​​​‌ rely on data from​ selected brain regions but​‌ integrate the user’s brain​​ network information. Using a​​​‌ combination of neuroimaging and​ experimental methods within a​‌ modern computational framework, they​​ will study brain dynamics​​​‌ to improve BCI architecture​ and accuracy. Apart from​‌ refining BCIs, the project​​ has the potential to​​​‌ unveil solutions for motor​ restoration after stroke.

9.4​‌ National initiatives

ANR-PRC BETPARK​​

Participants: Mario Chavez,​​​‌ Fabrizio De Vico Fallani​ [Correspondant].

  • Project title:​‌
    Neurofeedback for Parkinson’s disease​​
  • Duration:
    2021 - 2025​​​‌
  • Amount:
    712k€
  • Coordinator:
    Nathalie​ George
  • Other partners:
    CNRS​‌ CCLE; ICM
  • Summary:
    Parkinson’s​​ disease (PD) is a​​​‌ complex neurodegenerative disease caused​ by death of midbrain​‌ dopaminergic neurons. This calls​​ for better understanding the​​​‌ pathophysiology of PD in​ order to pave the​‌ way to new non-pharmacological​​ and non-invasive treatment options​​​‌ for PD. We propose​ to use neurofeedback (NF)​‌ to test whether PD​​ patients can learn to​​​‌ self-regulate their brain activity​ to reduce pathological neural​‌ activity and thereby motor​​ symptoms. We will leverage​​​‌ NF to target regulation​ of pathological beta band​‌ (8-35 Hz) oscillations, and​​ we will characterize training-induced​​​‌ changes in cortical network​ activity and their relationship​‌ with symptom severity. Our​​ goal is to provide​​​‌ direct evidence of the​ functional role of beta​‌ rhythms in the pathophysiology​​ of PD while assessing​​​‌ NF as a new​ non-pharmacological and non-invasive tool​‌ for ameliorating PD motor​​ symptoms.
ANR-PRC MEO

Participants:​​​‌ Laurent Hugueville [Correspondant].​

  • Project title:
    Overcome SQUID​‌ and alkali OPM limitations​​ for Epilepsy: 4He OPMs​​​‌
  • Duration:
    2022 - 2025​
  • Amount:
    639k€
  • Coordinator:
    Francesca​‌ Bonini
  • Other partners:
    MAG4Health,CRNL,INS,APHM,​​ ICM
  • Summary:
    The main​​​‌ objective of this highly​ interdisciplinary and collaborative project​‌ is to demonstrate that​​ innovative optically pumped magnetometers​​​‌ (OPM) using helium-4 (4HeOPM)​ can overcome the limitations​‌ of SQUID and alkali​​ OPM sensors to monitor​​​‌ brain activity in epileptic​ patients including children and​‌ to perform long-term recording​​ of epileptic seizures. Innovative​​​‌ optically pumped magnetometers (OPMs)​ using helium-4 (4HeOPM) have​‌ been developed by Mag4Health,​​ which operate at room​​​‌ temperature. These sensors can​ be placed near the​‌ scalp and have a​​ wider dynamic range and​​​‌ bandwidth more suitable for​ detecting epileptic activities. Our​‌ aim is to demonstrate​​ their capability to overcome​​​‌ the limitations of commercially​ available MEG systems (sMEGs)​‌ as well as prototype​​ alkaline OPM, thus opening​​​‌ new horizons for the​ non-invasive pre-surgical evaluation of​‌ epilepsy including recording of​​ seizure onset.
MinArm-Inria Boucle​​​‌ Dort

Participants: Mario Chavez​ [Correspondant].

  • Project title:​‌
    Neurofeedback for Parkinson’s disease​​
  • Duration:
    2025 - 2027​​
  • Amount:
    312k€
  • Coordinator:
    Mario​​​‌ Chavez
  • Other partners:
    Inria,‌ SSA
  • Summary:
    Severe hemorrhage‌​‌ remains the leading cause​​ of death among wounded​​​‌ military personnel. The objective‌ of the Boucle DORT‌​‌ project is to continue​​ the development of an​​​‌ embedded automated system for‌ the management of critically‌​‌ injured patients and to​​ extend its application beyond​​​‌ hemorrhagic trauma alone. The‌ envisioned uses of the‌​‌ system range from automated​​ resuscitation to the comprehensive​​​‌ management of severe trauma,‌ including hemorrhagic shock, traumatic‌​‌ brain injury, and severe​​ burns. The current project​​​‌ focuses on the development‌ of algorithms and a‌​‌ software interface for to​​ the development of a​​​‌ prototype dedicated interface, together‌ with a software application‌​‌ enabling patient categorization, integration​​ with physiological monitoring systems,​​​‌ and control of an‌ automated device capable of‌​‌ administering selected treatments in​​ closed loop (catecholamines, fluid​​​‌ resuscitation, and sedation).

9.4.1‌ ANR

Grasp-IT, ANR PRCE‌​‌ CES 33 (interaction, robotics)​​
  • Title:
    Design and evaluation​​​‌ of a tangible and‌ haptic brain-computer interface for‌​‌ upper limb rehabilitation after​​ stroke
  • Duration:
    Jan2020-July2024
  • Coordinator:​​​‌
    Laurent Bougrain (LORIA/NeuroRhythms)
  • Partners:‌
    • LORIA (Lorraine Research Laboratory‌​‌ in Computer Science and​​ its Applications)
    • Center for​​​‌ research Inria Rennes -‌ Bretagne Atlantique
    • Center for‌​‌ research Inria Sophia Antipolis​​ - Méditerranée
    • IRR UGECAM-NE​​​‌ centre Lay Saint Christophe‌
    • CHU Rennes / Physical‌​‌ Medicine and Rehabilitation Service​​
    • CHU Toulouse
    • SARL ALCHIMIES​​​‌
  • Loria contact:
    Laurent Bougrain‌
  • Summary:
    This project aims‌​‌ to recover upper limb​​ control improving the kinesthetic​​​‌ motor imagery (KMI) generation‌ of post-stroke patients using‌​‌ a tangible and haptic​​ interface within a gamified​​​‌ Brain-Computer Interface (BCI) training‌ environment. (i) This innovative‌​‌ KMI-based BCI will integrate​​ complementary modalities of interactions​​​‌ such as tangible and‌ haptic interactions in a‌​‌ 3D printable flexible orthosis.​​ We propose to design​​​‌ and test usability (including‌ efficacy towards the stimulation‌​‌ of the motor cortex)​​ and acceptability of this​​​‌ multimodal BCI. (ii) The‌ GRASP-IT project proposes to‌​‌ design and integrate a​​ gamified non-immersive virtual environment​​​‌ to inter- act with.‌ This multimodal solution should‌​‌ provide a more meaningful,​​ engaging and compelling stroke​​​‌ rehabilitation training program based‌ on KMI production. (iii)‌​‌ In the end, the​​ project will integrate and​​​‌ evaluate neurofeedbacks, within the‌ gamified multimodal BCI in‌​‌ an ambitious clin- ical​​ evaluation with 75 hemiplegic​​​‌ patients in 3 different‌ rehabilitation centers in France.‌​‌ The GRASP-IT project represents​​ a challenge for the​​​‌ industrial 3D printing field.‌ The materials of the‌​‌ 3D printable orthosis, allowing​​ the integration of haptic-tangible​​​‌ interfaces, will come from‌ a joint R&D work‌​‌ performed by the companies​​ Alchimies and Open Edge.​​​‌
BCI4IA, ANR PRC CES‌ 19 (Technologies for health)‌​‌
  • Title:
    a New BCI​​ Paradigm To Detect Intraoperative​​​‌ Awareness During General Anesthesia‌
  • Duration:
    Jan2023-Dec2026
  • Coordinator:
    Claude‌​‌ Meistelman (CHRU Nancy)
  • Partners:​​
    • CIC regional university hospital​​​‌ of Nancy
    • LORIA
    • Center‌ for research Inria Bordeaux‌​‌ - Sud-Ouest
    • Anesthesia and​​ intensive care unit/CHU-Brugmann, Belgium​​​‌ (unfunded)
    • Laboratory of Neurophysiology‌ and Movement Biomechanics/Université Libre‌​‌ de Bruxelles, Belgium (unfunded)​​
  • Loria contact:
    Laurent Bougrain​​​‌
  • Summary:
    The BCI4IA project‌ aims to design a‌​‌ brain-computer interface to enable​​​‌ reliable general anesthesia (GA)​ monitoring, in particular to​‌ detect intraoperative awareness. Currently,​​ there is no satisfactory​​​‌ solution to do so​ whereas it causes severe​‌ post-traumatic stress disorder. "I​​ couldn't breathe, I couldn't​​​‌ move or open my​ eyes, or even tell​‌ the doctors I wasn't​​ asleep." This testimony shows​​​‌ that a patient's first​ reaction during an intraoperative​‌ awareness is usually to​​ move to alert the​​​‌ medical staff. Unfortunately, during​ most surgery, the patient​‌ is curarized, which causes​​ neuromuscular block and prevents​​​‌ any movement. To prevent​ intraoperative awareness, we propose​‌ to study motor brain​​ activity under GA using​​​‌ electroencephalography (EEG) to detect​ markers of motor intention​‌ (MI) combined with general​​ brain markers of consciousness.​​​‌ We will analyze a​ combination of MI markers​‌ (relative powers, connectivity) under​​ the propofol anesthetics, with​​​‌ a brain-computer interface based​ on median nerve stimulation​‌ to amplify them. Doing​​ so will also require​​​‌ to design new machine​ learning algorithms based on​‌ one-class (rest class) EEG​​ classification, since no EEG​​​‌ examples of the patient's​ MI under GA are​‌ available to calibrate the​​ BCI. Our preliminary results​​​‌ are very promising to​ bring an original solution​‌ to this problem which​​ causes serious traumas.
ANR-JCJC​​​‌ MANET

Participants: Marie-Constance Corsi​ [Correspondant].

  • Project title:​‌
    Multimodal Approaches based on​​ Neurophysiological markers to Enhance​​​‌ brain-computer inTerfaces (MANET)
  • Duration:​
    2026 - 2029
  • Amount:​‌
    285k€
  • Coordinator:
    Marie-Constance Corsi​​
  • Summary:
    Despite being beneficial​​​‌ for patients, controlling a​ BCI system is a​‌ learned skill that a​​ non-negligible proportion of the​​​‌ users cannot develop even​ after training sessions. This​‌ phenomenon, called "BCI inefficiency"​​ constitutes a strong limitation​​​‌ to the BCI diffusion.​ The “BCI inefficiency” concept​‌ is useful to understand​​ why some users cannot​​​‌ interact well with BCI​ technologies. Nevertheless, it relies​‌ on the assumption that​​ users are expected to​​​‌ reach an accuracy level​ within a finite time;​‌ and it infers that​​ it is inherently on​​​‌ the user. Therefore, instead​ of considering the user​‌ as a cause of​​ the observed variability, it​​​‌ appears essential to develop​ BCI systems that aim​‌ at considering the user's​​ specificity. The underlying challenge​​​‌ here is to extract​ the most relevant information​‌ to discriminate properly the​​ user's mental state, referred​​​‌ as “features”. My goal​ is to develop methods​‌ to assess and to​​ improve BCI performance by​​​‌ considering the subjects' specificity.​ Aware that the BCI​‌ training relies on mutual-learning​​ schemes between the users​​​‌ and the machines, I​ will propose a framework​‌ that tackles the "BCI​​ inefficiency" phenomenon from both​​​‌ the neural decoder and​ the users’ sides. I​‌ first propose to enrich​​ the decoded features by​​​‌ considering combination of multimodal​ and heterogeneous information to​‌ develop innovative classification tools​​ (WP1). Then, I will​​​‌ generalize this framework by​ introducing new features based​‌ on the identification of​​ reliable and subject-specific neurophysiological​​​‌ markers of BCI performance​ (WP 2). Finally, I​‌ will conduct an experimental​​ validation through online BCI​​​‌ experiments based on new​ generation of MEG sensors,​‌ namely the optically-pumped magnetometers​​ (OPMs) (WP 3).
AI​​ Cluster PrAIrie-PSAI

Participants: Fabrizio​​​‌ De Vico Fallani [Correspondant]‌.

  • Project title:
    Hihger-order‌​‌ interactions in brain networks​​
  • Duration:
    Since 2024
  • Amount:​​​‌
    75M€
  • Coordinator:
    Isabelle Ryl‌
  • Other partners:
    PSL,CNRS,Paris Cite,‌​‌ Inria, Pasteur
  • Summary:
    As​​ AI confirms its role​​​‌ as a disruptive technology‌ and its impact on‌​‌ all sectors of society,​​ the PRAIRIE - Paris​​​‌ School of AI project‌ is positioned as a‌​‌ catalyst for innovation and​​ research in AI, with​​​‌ the ambition of becoming‌ the world leader that‌​‌ France needs to remain​​ competitive on the international​​​‌ stage. Winner of the‌ IA Cluster call for‌​‌ 75 million, it brings​​ together the same players​​​‌ who, since 2019, have‌ made the 3IA PRAIRIE‌​‌ Institute a success and​​ established it on the​​​‌ world stage as a‌ leading player in Artificial‌​‌ Intelligence (AI) research and​​ training. Taking full advantage​​​‌ of this momentum, PRAIRIE-PSAI‌ will broaden the positioning‌​‌ of the 3IA PRAIRIE​​ Institute, by federating the​​​‌ interdisciplinary research and training‌ initiatives of its partners.‌​‌ The strength of the​​ consortium is unique in​​​‌ this respect.

10 Dissemination‌

10.1 Promoting scientific activities‌​‌

10.1.1 Scientific events: organisation​​

General chair, scientific chair​​​‌
  • Marie-Constance Corsi co-chaired the‌ "Next Generation Brain-Computer Interface"‌​‌ workshop during the IEEE​​ MetroXRAINE conference in October​​​‌ 2025 (Ancona, Italy)
  • Marie-Constance‌ Corsi co-chaired the "Brain‌​‌ models as a tool​​ for a multimodal integration"​​​‌ symposium during the 1st‌ OHBM Satellite Meeting in‌​‌ September 2025 (Virtual)
  • Marie-Constance​​ Corsi co-chaired the Special​​​‌ Session on "Decoding the‌ brain time series" during‌​‌ the 35th IEEE International​​ Workshop on Machine Learning​​​‌ for Signal Processing (IEEE‌ MLSP 2025) in August‌​‌ 2025 (Istanbul, Turkey)
  • Marie-Constance​​ Corsi co-chaired with Tristan​​​‌ Venot the "Exploring features‌ to improve BCI: challenges‌​‌ and opportunities" workshop during​​ the 11th BCI meeting​​​‌ in June 2025 (Banff,‌ Canada)
  • Laurent Bougrain co-chaired‌​‌ the "Exploring Altered States​​ of Consciousness Through EEG​​​‌ and BCI" workshop at‌ the 47th IEEE IEEE‌​‌ Engineering in Medicine and​​ Biology Society (EMBC 2025)​​​‌ in July 2025 (Copenhaguen,‌ Denmark)
  • Laurent Bougrain co-chaired‌​‌ the "Exploring the Clinical​​ Integration of BCI Technology​​​‌ in General Anesthesia Monitoring"‌ workshop at the BCI‌​‌ Meeting in June 2025​​ (Banh, Canada)
Member of​​​‌ the organizing committees
  • Fabrizio‌ De Vico Fallani co-organized‌​‌ Network Neuroscience 2025 at​​ NetSci Maastricht, Neitherlands
  • Marie-Constance​​​‌ Corsi co-organized PracticalMEEG 2025‌ (sponsored by Inria) in‌​‌ October 2025 (Aix-en-Provence, France)​​
  • Marie-Constance Corsi co-organized with​​​‌ C. Cury and P.‌ Maurel (EPI Empenn) the‌​‌ "Journées Scientifiques Inria" in​​ June 2025 (Paris, France)​​​‌
Data competition
  • "EEG Foundation‌ Challenge: From Cross-Task to‌​‌ Cross-Subject EEG Decoding" (EEG​​ Challenge 2025) - Submision​​​‌ accepted to the NeurIPS‌ 2025 Competition Track. This‌​‌ competition was led by​​ Bruno Aristimunha Pinto together​​​‌ with the other core‌ coordinators. Marie-Constance Corsi was‌​‌ part of the Domain​​ experts. There were 1,183​​​‌ teams/participants and more than‌ 8,000 submissions on the‌​‌ open source platform Codabench.​​ A record for the​​​‌ EEG domain!

10.1.2 Scientific‌ events: selection

Member of‌​‌ the conference program committees​​
  • Fabrizio De Vico Fallani​​​‌ was member of the‌ Complex Systems Society conference‌​‌ 2025
  • Fabrizio De Vico​​​‌ Fallani served as member​ for NetSciX 2025
  • Fabrizio​‌ De Vico Fallani served​​ as member of Complex​​​‌ Networks 2025
  • Fabrizio De​ Vico Fallani served as​‌ member for Complenet 2025​​
Reviewer
  • Fabrizio De Vico​​​‌ Fallani served as reviewer​ for NetSci 2025
  • Fabrizio​‌ De Vico Fallani served​​ as reviewer for NetSciX​​​‌ 2025
  • Fabrizio De Vico​ Fallani served as reviewer​‌ for Complex Networks 2025​​
  • Fabrizio De Vico Fallani​​​‌ served as reviewer for​ Complenet 2025
  • Marie-Constance Corsi​‌ served as reviewer for​​ NeurIPS 2025
  • Marie-Constance Corsi​​​‌ served as reviewer for​ IEEE MetroXRAINE 2025
  • Marie-Constance​‌ Corsi and Laurent Bougrain​​ served as reviewer for​​​‌ CORTICO 2025

10.1.3 Journal​

Member of the editorial​‌ boards
  • Fabrizio De Vico​​ Fallani served as Academic​​​‌ Editor for Brain Topography​
  • Marie-Constance Corsi served as​‌ Academic Editor for PLOS​​ ONE
  • Diego Candia-Rivera served​​​‌ as Academic Editor for​ The Journal of Physiology​‌
Reviewer - reviewing activities​​
  • Fabrizio De Vico Fallani​​​‌ served as reviewer for​ Nature Communications, Network Neuroscience,​‌ Journal of Neural Engineering​​
  • Marie-Constance Corsi served as​​​‌ reviewer for eNeuro, NeuroImage:​ Clinical, Brain Topography, Brain​‌ Connectivity, Journal of Neural​​ Engineering, IEEE Transactions on​​​‌ Biomedical Engineering (TBME), IEEE​ Reviews in Biomedical Engineering,​‌ Scientific Reports, PLOS ONE,​​ Brain‑Computer Interfaces, Epilepsy Open​​​‌
  • Mario Chavez served as​ reviewer for Physical Reviw​‌ E, Brain Cmmunications, Proceeedings​​ of the National Academy​​​‌ of Sciences Nexus, Chaos,​ NPJ Digital Medecine
  • Laurent​‌ Bougrain served as reviewer​​ for Brain Topography, Virtual​​​‌ Reality'25, Cortico'25
  • Diego Candia-Rivera​ served as reviewer for​‌ Computers in Biology and​​ Medicine, Clinical Neurophysiology, Experimental​​​‌ Physiology, Biological Psychiatry, Annals​ of the New York​‌ Academy of Sciences, Communications​​ Biology, Nature Reviews Neurology,​​​‌ Social Cognitive and Affective​ Neuroscience, Physiological Measurement, Nature​‌ Human Behaviour, Progress in​​ Neurobiology, Springer Books
  • Andrea​​​‌ Civilini served as reviewer​ for Communications Physics, Chaos​‌ Solitons and Fractals, Journal​​ of Complex Networks

10.1.4​​​‌ Invited talks

  • Fabrizio De​ Vico Fallani gave an​‌ invited talk at the​​ Bernstein network computational neuroscience​​​‌ conference, Frankfurt 2025
  • Fabrizio​ De Vico Fallani gave​‌ an invited talk at​​ the 11th Brain-computer interface​​​‌ meeting, Banff Canada, 2025​
  • Fabrizio De Vico Fallani​‌ gave an invited talk​​ Inria-Brasil Workshop on digital​​​‌ health, Hybrid 2025
  • Fabrizio​ De Vico Fallani gave​‌ an invited talk NETSCI​​ Workshop on networks in​​​‌ biology and medicine, Maastricht​ 2025
  • Marie-Constance Corsi gave​‌ a plenary talk during​​ the 11th BCI meeting​​​‌ (Early Career Award) in​ June 2025 (Banff, Canada)​‌
  • Marie-Constance Corsi gave a​​ talk during the AI-Data​​​‌ workshop of the 11th​ BCI meeting in June​‌ 2025 (Banff, Canada)
  • Marie-Constance​​ Corsi presented her work​​​‌ during the France-Taiwan STC​ workshop (AI & Health)​‌ in October 2025 (Paris,​​ France)
  • Marie-Constance Corsi presented​​​‌ her work during an​ Essex BCI-NE webinar in​‌ December 2025
  • Marie-Constance Corsi​​ presented her work during​​​‌ a seminar organized by​ the Institut de Neuromodulation​‌ in November 2025 (Paris,​​ France)
  • Marie-Constance Corsi presented​​​‌ her work during a​ seminar organized by the​‌ "Ingénierie Cognitive et Neurosciences​​ appliquées" (ICNA) department of​​​‌ the ONERA in August​ 2025 (Salon de Provence,​‌ France)
  • Marie-Constance Corsi was​​ invited to present during​​ the NxGenBCI workshop during​​​‌ the last IEEE MetroXRAINE‌ in October 2025 (Ancona,‌​‌ Italy)
  • Mario Chavez was​​ invited to give the​​​‌ talk “The intrinsic geometry‌ of complex brain networks‌​‌ as biomarkers in epilepsy”,​​ within the School on​​​‌ Synchronization : from colective‌ motion to brain dynamics,‌​‌ held at the ICTP-Sao​​ Paulo, Brazil, in February​​​‌ 2025
  • Mario Chavez was‌ invited to give the‌​‌ talk “The intrinsic geometry​​ of complex brain networks​​​‌ as biomarkers in epilepsy”,‌ at the UBICS, University‌​‌ of Barcelona, in Mars​​ 2025
  • Mario Chavez was​​​‌ invited to give the‌ talk “Brain Connectivity and‌​‌ Cinical Monitoring”, at the​​ Taller de Sistemas Complejos,​​​‌ Universidad Autonoma del Estado‌ de Morelos, Cuernavaca, Mexico,‌​‌ in May 2025
  • Mario​​ Chavez was invited to​​​‌ give the talk “The‌ intrinsic geometry of complex‌​‌ brain networks as biomarkers​​ in epilepsy”, at the​​​‌ Conference LANET, Punta del‌ Este, Uruguay, in July‌​‌ 2025
  • Mario Chavez was​​ invited to give the​​​‌ talk “Hyperbolic embedding of‌ brain networks for detecting‌​‌ regions disrupted by neurodegeneration”,​​ at the seminar “Redes​​​‌ complejas, Estructura Procesos Dinamicos”,‌ at the Universidad de‌​‌ Buenos Aires, Argentina, in​​ July 2025
  • Laurent Bougrain​​​‌ was invited to present‌ his work during two‌​‌ talks on "Artificial Intelligence​​ for EEG-based Brain-Computer Interfaces"​​​‌ and "Designing non-invasive brain-computer‌ interfaces" at the engineering‌​‌ of the national university​​ of Entre Rios(Parana, Argentina)​​​‌
  • Laurent Bougrain was invited‌ to present his work‌​‌ on "Designing a non-invasive​​ BCI for upperlimb rehabilitation​​​‌ after stroke" Bordeaux university‌ on September 29, 2025,‌​‌ Bordeaux, France
  • Diego Candia-Rivera​​ was invited to present​​​‌ at the University of‌ Glasgow, December 2025 (Glasgow,‌​‌ UK)
  • Diego Candia-Rivera was​​ invited to present during​​​‌ the International Union of‌ Physiological Societies World Congress‌​‌ 2025 (Frankfurt, Germany)
  • Diego​​ Candia-Rivera was invited to​​​‌ present during the Paris‌ Postdoc Seminars. January 2025‌​‌ (Paris, France)

10.1.5 Leadership​​ within the scientific community​​​‌

  • Marie-Constance Corsi is member‌ of the scientific advisory‌​‌ board of CuttingEEG
  • Laurent​​ Bougrain & Marie-Constance Corsi​​​‌ are members of the‌ Board of Directors of‌​‌ the scientific society CORTICO​​ for the promotion of​​​‌ Brain-Computer Interfaces in France.‌

10.1.6 Scientific expertise

  • Marie-Constance‌​‌ Corsi served in 2025​​ as reviewer for the​​​‌ call for proposal entitled‌ "Data IA Insitute: AI‌​‌ Modular Chairs"
  • Marie-Constance Corsi​​ served in 2025 as​​​‌ reviewer for the call‌ launched by the Dutch‌​‌ Research Council Domain Applied​​ and Engineering Sciences (NWO​​​‌ Domain AES)
  • Marie-Constance Corsi‌ served in 2025 as‌​‌ reviewer for the call​​ launched by Toulouse Initiative​​​‌ for Research Impact on‌ Society (TIRIS)
  • Marie-Constance Corsi‌​‌ has served since 2025​​ as reviewer for the​​​‌ Paris Brain Institute call‌ for Carnot Tools/Maturation projects‌​‌
  • Mario Chavez served in​​ 2025 as reviewer for​​​‌ the call ERC Consolidator‌ Grant
  • Mario Chavez served‌​‌ in 2025 as reviewer​​ for the French call​​​‌ (“bourses”) of the Ligue‌ Française contre l’Epilepsie
  • Diego‌​‌ Candia-Rivera served in 2025​​ as reviewer for the​​​‌ AAPG Generic call –‌ Agence Nationale de la‌​‌ Recherche (ANR)

10.1.7 Research​​ administration

  • Fabrizio De Vico​​​‌ Fallani has served in‌ the 2025 Inria Groupe‌​‌ de travail - Creation​​​‌ EPC COPHY
  • Fabrizio De​ Vico Fallani has served​‌ as ICM representative for​​ the European Brain Research​​​‌ Infrastructures - Ebrains
  • Marie-Constance​ Corsi has served as​‌ member of the Inria​​ Paris Doctoral Advisory Committee​​​‌ since 2025
  • Marie-Constance Corsi​ has served as member​‌ of the Inria Paris​​ Center Committee since 2025​​​‌

10.2 Teaching - Supervision​ - Juries - Educational​‌ and pedagogical outreach

10.2.1​​ Teaching

  • Fabrizio De Vico​​​‌ Fallani , MAPIMED course,​ Sorbonne Univ., Complex brain​‌ networks (3h), Paris, France​​
  • Marie-Constance Corsi , CENIR​​​‌ course, ICM, Brain-Computer Interface​ (2h), Paris, France
  • Marie-Constance​‌ Corsi , DU IA​​ Santé, Univ. Paris Cité,​​​‌ Introduction to Brain-Computer Interfaces​ (1h), Paris, France
  • Marie-Constance​‌ Corsi , Master Computational​​ Neuroscience and Neuroengineering (CNN),​​​‌ Univ. Paris-Saclay, Network science​ for understanding Brain-Computer Interfaces​‌ (3h), Saclay, France
  • Marie-Constance​​ Corsi , Master Mathématiques​​​‌ Vision Apprentissage (MVA), ENS​ Saclay, Imagerie fonctionnelle cérébrale​‌ et interface cerveau machine​​ (9h), Saclay, France
  • Laurent​​​‌ Bougrain , "Brain-computer Interfaces"​ and "Signal processing and​‌ machine learning of electroencephalographic​​ Signals" (8h), Licence and​​​‌ Master in Life Science​ and Systems Engineering, Kyushu​‌ Institute of Technology (Kyutech),​​ Kitakyushu-shi, Fukuoka, Japan

10.2.2​​​‌ Supervision

  • PhD Theses
    • Camilla​ Mannino : co-supervised by​‌ Mario Chavez & Marie-Constance​​ Corsi
    • Martin Guillemaud :​​​‌ supervised by Mario Chavez​
    • Alice Longhena : supervised​‌ by Mario Chavez
    • Cassandra​​ Dumas : co-supervised by​​​‌ Nathalie George (CNRS, ICM)​ & Marie-Constance Corsi
    • Marc​‌ Fiammante : supervised by​​ Mario Chavez
    • Sébastien Velut​​​‌ : co-supervised with Frédéric​ Dehais (ISAE-Supaero), Sylvain Chevallier​‌ (Univ. Paris-Saclay, EPI TAU)​​ & Marie-Constance Corsi
    • Jules​​​‌ Gomel : co-supervised by​ Frédéric Dehais (ISAE-Supaero) &​‌ Marie-Constance Corsi
    • Bruno Aristimunha​​ Pinto : co-supervised by​​​‌ Sylvain Chevallier (Univ. Paris-Saclay,​ EPI TAU), Raphael Camargo​‌ (Universidade Federal do ABC,​​ Brazil) & Marie-Constance Corsi​​​‌
    • Baptiste Fague : co-supervised​ by Elisa Tartaglia (EssilorLuxottica)​‌ & Marie-Constance Corsi
    • Giovanni​​ Messuti : co-supervised by​​​‌ Silvia Scarpetta (Univ of​ Salerno, Italy) & Marie-Constance​‌ Corsi
  • Master Theses
    • Pierre-Baptiste​​ Mathieu de Carvalho: co-supervised​​​‌ by Laurent Bougrain &​ Marie-Constance Corsi co-supervised (April-Sept​‌ 2025)
    • Mario Roca :​​ supervised by Marie-Constance Corsi​​​‌ (Feb-August 2025)
    • Francesco Farina​ : supervised by Mario​‌ Chavez (Feb-July 2025)
    • Laura​​ Pitti : supervised by​​​‌ Diego Candia-Rivera (September 2025-February​ 2026)
    • Giovanni Sitti :​‌ supervised by Diego Candia-Rivera​​ (September 2025-February 2026)
    • Linon​​​‌ Denis : supervised by​ Mario Chavez (October 2025-April​‌ 2026)
  • Bachelor's thesis internship​​
    • Apurba Debnath: co-supervised by​​​‌ Parul Verma (IIT Madras,​ India) and Marie-Constance Corsi​‌ (Jan-July 2025)

10.2.3 Juries​​

  • Fabrizio De Vico Fallani​​​‌ , President, PhD committee​ of D Trocellier, PhD​‌ Informatics, Univ Bordeaux
  • Fabrizio​​ De Vico Fallani ,​​​‌ Relator, PhD commitee of​ D Hajhassani, PhD Biomedical​‌ engineering, Univ Grenoble-Alpes
  • Marie-Constance​​ Corsi , admissibility jury​​​‌ member, CRCN Inria Bordeaux​
  • Marie-Constance Corsi , Examiner,​‌ PhD committee of Ambroise​​ Heurtebise (Paris-Saclay University), Saclay,​​​‌ France
  • Marie-Constance Corsi ,​ Examiner, PhD committee of​‌ Hasnae Agouram (Aix-Marseille University),​​ Marseille, France
  • Marie-Constance Corsi​​​‌ , Examiner, PhD committee​ of Edouard Ferrand (Paris-Saclay​‌ University), Saclay, France
  • Marie-Constance​​ Corsi , Examiner, PhD​​​‌ committee of Alix Lamouroux​ (Ecole nationale supérieure Mines-Télécom​‌ Atlantique), Brest, France
  • Marie-Constance​​ Corsi , Examiner, PhD​​ committee of Hanane Moumane​​​‌ (Sorbonne University), Paris, France‌
  • Marie-Constance Corsi , Examiner,‌​‌ PhD committee of Zaineb​​ Ajra (Montpellier University), Montpellier,​​​‌ France
  • Mario Chavez ,‌ Examiner, PhD committee of‌​‌ Hamed Azizollahi (Université de​​ Picardie Jules Verne), Amiens,​​​‌ France, July 2025
  • Mario‌ Chavez , Examiner, PhD‌​‌ committee of Claudio Caprioli​​ (University of Catania), Italy,​​​‌ September 2025
  • Mario Chavez‌ , Examiner, Professorship Promotion‌​‌ Committee of Nasrine Jrad​​ (Université Catolique de l’Ouest),​​​‌ Angers, France, June 2025‌
  • Laurent Bougrain , Examiner,‌​‌ PhD committee of David​​ Trocelier Bordeaux university on​​​‌ September 29, 2025, Bordeaux,‌ France
  • Laurent Bougrain ,‌​‌ Opponent, PhD committee of​​ Pex Pufvesson Lund university​​​‌ on November 21, 2025,‌ Lund, Sweeden
  • Diego Candia-Rivera‌​‌ , Examiner, PhD committee​​ of Lauren Zwienenberg (Maastricht​​​‌ University), Maastricht, The Netherlands‌

10.2.4 Educational and pedagogical‌​‌ outreach

  • Celestine Allombert Blaise​​ welcomed five schoolchildren (during​​​‌ a “stage de 3e”‌ internship week) in the‌​‌ NERV Team at the​​ Paris Brain Institute (ICM).​​​‌ The aim of this‌ initiative was to introduce‌​‌ them to neuroscience research​​ through interactive presentations, lab​​​‌ visits, and hands‑on demonstrations‌ together with team members.‌​‌

10.3 Popularization

10.3.1 Specific​​ official responsibilities in science​​​‌ outreach structures

  • Marie-Constance Corsi‌ : member of the‌​‌ organizing committee of the​​ France Brain Bee (Olympiades​​​‌ de Neurosciences)

10.3.2 Participation‌ in Live events

  • Fabrizio‌​‌ De Vico Fallani organized​​ and moderated the roundtable​​​‌ on Network Neuuroscience as‌ part of the NSF-supported‌​‌ Accelenet-Multinet international program.
  • Fabrizio​​ De Vico Fallani participated​​​‌ as a panelist in‌ a round table on‌​‌ BCIs organized by DIM​​ C-BRAINS at the Centre​​​‌ International de Conferences of‌ Sorbonne University.

10.3.3 Others‌​‌ science outreach relevant activities​​

  • Fabrizio De Vico Fallani​​​‌ , talk given during‌ "Les Matinales de l'Institut‌​‌ du Cerveau" event at​​ Paris Brain Institute
  • Marie-Constance​​​‌ Corsi , participation to‌ the "Becoming a PI"‌​‌ wokshop organized at the​​ Paris Brain Institute
  • Marie-Constance​​​‌ Corsi , presentation to‌ the "Institut de la‌​‌ gestion publique et du​​ développement économique" (IGPDE)
  • Marie-Constance​​​‌ Corsi , presentation to‌ the SCAI Spring School‌​‌ on "How can emotionally​​ intelligent AI transform society"​​​‌
  • Marie-Constance Corsi , talk‌ given during the "Les‌​‌ Mardis de la Sorbonne"​​ event
  • Mario Chavez and​​​‌ Marc Fiammante , presentation‌ of the project “NewBorn‌​‌ NeuroDigital” to the French​​ President M Emmanuel on​​​‌ the occasion of the‌ Summit for Action on‌​‌ Artificial Intelligence (AI), which​​ took place on 10​​​‌ and 11 February 2025‌ in Paris, France
  • Laurent‌​‌ Bougrain Oxford-Style Debate :​​ “Connaître le cerveau est​​​‌ nécessaire pour le développement‌ à venir de l'IA”,‌​‌ affirmative team , Forum​​ des Sciences Cognitives et​​​‌ du TAL, Nov. 26,‌ 2025, Théâtre de la‌​‌ Manufacture, Nancy.
  • Celestine Allombert​​ Blaise and Bintou Soumaoro​​​‌ have regularly contributed to‌ the team’s online communication‌​‌ by updating the official​​ website as well as​​​‌ the LinkedIn and Bluesky‌ accounts. They have published‌​‌ outreach posts highlighting recent​​ scientific articles, team news,​​​‌ and public engagement initiatives.‌

11 Scientific production

11.1‌​‌ Major publications

11.2 Publications of the​‌ year

International journals

International peer-reviewed conferences

Conferences​​​‌ without proceedings

Scientific book chapters

Edition​​​‌ (books, proceedings, special issue‌ of a journal)

  • 37‌​‌ proceedingsAutomatic Ocular Artifact​​ Correction in Electroencephalography for​​​‌ Neurofeedback.BIOSIGNALS 2025‌ - 18th International Conference‌​‌ on Bio-inspired Systems and​​​‌ Signal ProcessingPorto, Portugal​SCITEPRESS - Science and​‌ Technology PublicationsFebruary 2025​​, 773-783HALDOI​​​‌back to text

Doctoral​ dissertations and habilitation theses​‌

  • 38 thesisM.Martin​​ Guillemaud. Latent geometries​​​‌ of epileptic brain networks​ as biomarkers for seizures​‌ forecasting and outcome of​​ surgery.Sorbonne université​​​‌December 2025HAL
  • 39​ thesisA.Alice Longhena​‌. Hyperbolic representations of​​ brain networks and clinical​​​‌ applications.Sorbonne Université​May 2025HAL

Reports​‌ & preprints

Other scientific publications​​