2025​‌Activity reportProject-TeamCRONOS​​

2.1.1 Brain Signal Modeling​​

To model brain signals,​​​‌ we adopt a signal​ processing perspective, i.e. consider​‌ that at the macroscopic​​ scale we look at​​​‌ signals, general signal modeling​ approaches are sufficient for​‌ the goal of describing​​ brain dynamics. This contrasts​​​‌ with many other teams,​ which often follow a​‌ more constructivist approach, where​​ brain signal models are​​​‌ derived from a computational​ neuroscience perspective, i.e. emerge​‌ from a mathematical modeling​​ of neurons, axons and​​​‌ dendrites or assemblies of​ those (spiking neurons, neural​‌ mass models, ...). Our​​ approach gives more freedom​​​‌ to model signals and​ allows for descriptions with​‌ a very limited number​​ of parameters at the​​​‌ cost of losing some​ connection with the microscopic​‌ reality. Additionally, the reduced​​ number of parameters simplifies​​​‌ the problem of parameter​ identification. Depending on the​‌ type of modeling, we​​ may use rather simple​​​‌ phenomenological descriptions (e.g. a​ region is activated or​‌ not) or more or​​ less complicated signal models​​​‌ (sampled signals with no​ specific temporal models, various​‌ multivariate autoregressive (AR) models,​​ integro-differential equations, ...). Generally​​​‌ speaking, the more we​ will be interested in​‌ dynamic properties of the​​ signals, the more we​​​‌ will need sophisticated signal​ models.

2.1.2 Network Modeling​‌

Classical – deterministic –​​ networks only strictly exist​​​‌ at the microscopic level​ where neurons are connected​‌ together through axons. For​​ usual neural systems, such​​​‌ networks are huge, currently​ unavailable and too complex​‌ for a macroscopic view​​ of the brain. At​​​‌ such a scale, networks​ are of stochastic nature.​‌ Both nodes and edges​​ are only defined probabilistically.​​​‌ We allow ourselves to​ work either directly with​‌ these probablistic networks or​​ with deterministic networks obtained​​​‌ by thresholding the probabilistic​ ones, in which case​‌ it is important to​​ pay attention to the​​​‌ amount of approximation and​ bias introduced by the​‌ thresholding operation. Its is​​ important to note that,​​​‌ while deterministic networks have​ received a lot of​‌ attention, the domain of​​ stochastic networks or its​​​‌ tranfer to computational neurosciences​ is still in its​‌ infancy.

3.1.1​‌ Automating the detection of​​ brain state changes from​​​‌ the acquired data

Experimental​ practice in MEG, EEG,​‌ and BCI shows that​​ very valuable information can​​​‌ be obtained directly from​ the “sensor space”. This​‌ encompasses estimating the number​​ 39 or the time​​​‌ courses of sources 35​, 34 or the​‌ detection of changes in​​ the brain dynamical activity.​​​‌ This type of information​ can be an indication​‌ of a “brain state​​ change”. In particular, automating​​​‌ the detection of changes​ in brain states would​‌ allow the splitting of​​ the incoming functional data​​​‌ into segments in which​ the brain network can​‌ be considered as stationary.​​ During such stationary segments​​​‌ of data, the sources​ and their locations remain​‌ fixed, which offers a​​ means to regularize the​​​‌ solution of the inverse​ problem of source reconstruction.​‌ Furthermore, any improvement on​​ EEG signal classification has​​​‌ direct applications for BCI.​ One approach that we​‌ explore is the use​​ of autoregressive spatio-temporal models​​​‌ or lagged correlation measurements​ as a means to​‌ improve classification of MEG/EEG​​ signals. These autoregressive spatio-temporal​​​‌ models constitute a possible​ first step towards using​‌ more sophisticated causality modeling.​​

3.1.2 Better modeling of​​​‌ the spatio-temporal variability of​ brain signals

Hand in​‌ hand with the previous​​ sub-goal is the need​​​‌ to better understand the​ spatio-temporal variability of brain​‌ signals. This variability originates​​ from several factors ranging​​​‌ from the intrinsic variability​ of the brain sources​‌ (even in a same​​ subject) to important differences​​​‌ in the spatial organization​ of the cortex across​‌ subjects or to variations​​ in the way sensors​​​‌ are setup by experimenters.​ Thus, variability occurs not​‌ only across subjects, but​​ also across sessions or​​​‌ even trials for a​ same subject.

The most​‌ common approach developed to​​ cope with the noisy​​​‌ MEG and EEG signals​ is commonly referred to​‌ as evoked potentials: the​​ signal is “clocked” on​​​‌ some stimulus or reaction​ of the subject and​‌ the low amplitude signals​​ – almost completely hidden​​​‌ by background activity –​ are then averaged over​‌ multiple repetitions (trials) of​​ the experiment to improve​​​‌ the SNR. But, because​ of variability, such an​‌ averaging distorts the overall​​ activity and hides specific​​​‌ activities such as high​ frequency components. It is​‌ thus advisable to improve​​ models so that they​​​‌ can “work” in single​ event mode (i.e. without​‌ averaging). Relying on multiple​​ trials to obtain better​​​‌ statistical models of individual​ signals (with techniques such​‌ as dictionary learning, autoregressive​​ models, deep-neural networks, ...)​​​‌ would be an important​ improvement over the current​‌ state-of-the-art as it would​​ provide a more objective​​ and systematic way of​​​‌ characterizing single trial data.‌ Events will be extracted‌​‌ separately for each trial​​ without relying on averaging,​​​‌ but with the knowledge‌ of the “group” model‌​‌ (see Section 3.3).​​ This will allow the​​​‌ study of the variability‌ of brain activity across‌​‌ trials (attention, habituation are​​ e.g. known to change​​​‌ the activity). This is‌ a difficult long term‌​‌ challenge with possible short-middle​​ term advances for some​​​‌ specific cases such as‌ epileptic spikes. This research‌​‌ path is grounded by​​ some of our previous​​​‌ work 2, 6‌. Understanding variability is‌​‌ particularly important for BCI​​ as it is often​​​‌ necessary to “learn” a‌ classifier to detect the‌​‌ subject's brain state with​​ a limited dataset (because​​​‌ of time constraint). At‌ the sensor level, spatial‌​‌ patterns of activity are​​ often described by covariance/correlation​​​‌ matrices. Using Riemannian metric‌ over the space of‌​‌ symmetric definite positive matrices​​ is a powerful technique​​​‌ that has been used‌ these last years by‌​‌ the BCI community 32​​. Extending and improving​​​‌ these techniques as well‌ as finding proper low‌​‌ dimensional spaces that characterize​​ brain activity are other​​​‌ research paths we will‌ follow.

3.1.3 Adapting to‌​‌ new sensor modalities

During​​ the last few years,​​​‌ tremendous improvements have been‌ made on means to‌​‌ acquire functional brain data.​​ Yet, novelty in this​​​‌ domain continues and new‌ sensors are regularly proposed.‌​‌ These sensors can offer​​ more accurate measurements with​​​‌ improved SNR and/or some‌ ease-of-use improvements. For example,‌​‌ the use of room​​ temperature MEG sensors (such​​​‌ as those developed by‌ Mag4Health – see Section‌​‌ 7.2) promise improvements​​ in both aspects. EEG​​​‌ dry electrodes hold the‌ promises to simplify the‌​‌ setup of BCI systems,​​ but are difficult to​​​‌ master. MR machines are‌ also more and more‌​‌ powerful with either higher​​ fields (better signal quality​​​‌ and/or better spatial resolution)‌ or, on the contrary,‌​‌ lower fields (easier to​​ use and better contrasts​​​‌ in some cases). It‌ is important to continually‌​‌ adapt processing methods to​​ exploit the specificities of​​​‌ these new sensors as‌ they become available to‌​‌ the community.

3.2.1 Source​​ modeling: anatomical, temporal, and​​​‌ numerical constraints

Brain sources‌ refer to regions of‌​‌ interests whose properties link​​​‌ brain activity to the​ observed measurements. For example,​‌ in EEG, brain sources​​ can be modeled as​​​‌ dipoles representing the superposition​ of many synchronous and​‌ parallel neurons. The combination​​ of the electric fields​​​‌ generated by these dipoles​ gives rise to the​‌ potential differences measured at​​ the surface of the​​​‌ scalp in EEG. The​ magnetic fields generated by​‌ these same dipoles give​​ rise to MEG signals.​​​‌ Regardless of the modality,​ the task of recovering​‌ brain sources from imaging​​ data is an inverse​​​‌ problem. This inverse problem​ is ill–posed, either because​‌ of the limited number​​ of sensors (EEG and​​​‌ MEG), because of poor​ signal to noise ratio​‌ (EEG, MEG, fMRI), or​​ because of the limited​​​‌ spatial (dMRI, EEG, MEG)​ or temporal resolution (fMRI,​‌ dMRI). To recover a​​ unique and stable solution,​​​‌ simple mathematical (i.e. not​ fully grounded by biology)​‌ criteria such as minimum​​ norm or sparsity are​​​‌ used to constrain the​ source space. However, these​‌ regularization approaches do not​​ correspond to any specific​​​‌ anatomical or physiological properties​ of the brain and​‌ are therefore quite arbitrary.​​ The challenge we address​​​‌ here is to increase​ the amount of subject​‌ specific anatomical and physiological​​ constraints taken into account​​​‌ for the recovery of​ brain activity from non–invasive​‌ imaging. More specifically, we​​ will investigate how brain​​​‌ networks, and in particular​ their associated delays, can​‌ be used to constrain​​ the inverse problem. Previous​​​‌ work has already shown​ the potential of temporal​‌ regularization based on brain​​ connectivity 4, but​​​‌ the topic remains largely​ open to identifying new​‌ models grounded in physiological​​ data. Another difficulty is​​​‌ the dynamic aspect of​ brain networks: we first​‌ assume that a stationary​​ brain state has been​​​‌ identified (e.g. using methods​ from Section 3.1).​‌ Eventually (as a long​​ term objective), transitions between​​​‌ stationary brain state models​ could be directly modeled​‌ in the networks themselves​​ and estimated from the​​​‌ data. The validation of​ the proposed models is​‌ both challenging and essential​​ and will be explored​​​‌ with our clinical partners​ (see Section 8).​‌ Finally, non-traditional imaging modalities​​ will also be considered.​​​‌ For example, accurate modeling​ of electromyography, in collaboration​‌ with Neurodec, can​​ provide important timing information​​​‌ of sources.

3.2.2 Unified​ network models explaining various​‌ brain measurements

In the​​ previous section, brain activity​​​‌ is estimated from one​ modality using information from​‌ the others as constraints​​ or priors. This obviously​​​‌ favors one modality over​ the others and propagates​‌ errors of processing pipelines​​ via the constraints. This​​​‌ modus operandi, while suboptimal,​ is historically justified: analysis​‌ pipelines have been developed​​ independently for each modality​​​‌ by different communities. Given​ the complementary nature of​‌ the different modalities, it​​ seems relevant to instead​​​‌ construct one global pipeline​ encompassing all brain measurement​‌ modalities into a single​​ unified framework. With brain​​​‌ networks arising as a​ central concept in the​‌ neuroscientific community, we propose​​ to make it the​​​‌ basis of such a​ global model. Doing so​‌ will enable the recovery​​ of brain network dynamics​​ from non–invasive multi–modal data,​​​‌ an important open problem‌ in neuroscience. To achieve‌​‌ this objective, we propose​​ to devise forward models​​​‌ describing the link between‌ brain networks and each‌​‌ of the various imaging​​ modalities. Specifically, effort is​​​‌ needed to evaluate different‌ formalisms describing brain networks‌​‌ that differ in the​​ way nodes and their​​​‌ interactions are defined. For‌ example, we investigate the‌​‌ relationship between electrical (EEG/MEG),​​ architectural (dMRI) and metabolic​​​‌ (fMRI) activities. Separated models‌ – either electric or‌​‌ hemodynamic – have been​​ proposed for some time,​​​‌ but coupling them is‌ an important challenge all‌​‌ the more that the​​ role of some neural​​​‌ constituents (such as glia)‌ is currently not well‌​‌ understood. Such models have​​ already been – at​​​‌ least partly as in‌ 33 – devised. The‌​‌ main challenge is to​​ define one that is​​​‌ sufficiently simple and well‌ spatialized (in particular to‌​‌ incorporate connectivity) in order​​ to be useful for​​​‌ our purpose. We will‌ also consider electromyography in‌​‌ this context, extending brain​​ networks to the peripherial​​​‌ nervous system.

3.2.3 Global‌ inverse problem using all‌​‌ measurements altogether

Models established​​ in the previous section​​​‌ can be used in‌ a global inverse problem‌​‌ involving all the measurements​​ modalities to recover the​​​‌ specific brain network involved‌ in a task. In‌​‌ all cases, we need​​ to solve an inverse​​​‌ problem over a complex‌ network model with a‌​‌ sparse prior as we​​ need to find “simple”​​​‌ networks that can explain‌ our data. Depending on‌​‌ the way networks are​​ modeled, some difficulties may​​​‌ arise. For example, the‌ model defined in 36‌​‌, 37 leads to​​ combinatorial explosion when used​​​‌ for source localization. Finding‌ appropriate models that facilitate‌​‌ the resolution of this​​ inverse problem may require​​​‌ some iterations between this‌ sub-task and the previous‌​‌ one.

3.3.1 Matching‌​‌ individual subject models

Reconstructions​​ obtained in Section 3.2​​​‌ will certainly differ even‌ for different measurement sessions‌​‌ obtained with a same​​ subject. An important problem​​​‌ to solve is the‌ matching of instances of‌​‌ such reconstructions (whether they​​ consist in brain networks​​​‌ or spatio-temporal autoregressive models).‌ In general, the relative‌​‌ positions of functional brain​​ areas are fairly well​​​‌ known. However, temporal variations‌ (lag, duration) in these‌​‌ models make the matching​​​‌ problem rather complicated. Finding​ and using some invariants​‌ of the models is​​ one path to find​​​‌ a good compromise that​ would allow for enough​‌ flexibility in the matching​​ process while avoiding the​​​‌ combinatorial complexity that e.g.​ graph matching problems usually​‌ exhibit. Another – long​​ term – possibility would​​​‌ be to directly solve​ “group problems”, but this​‌ would complexify even more​​ the modeling problem and​​​‌ might have the same​ drawbacks as averaging if​‌ not properly done (i.e.​​ hiding some activity or​​​‌ the intrinsic variability of​ the studied phenomenon).

3.3.2​‌ Statistics over brain networks​​

Once the matching of​​​‌ single subject models has​ been solved, statistics will​‌ be necessary to assess​​ the significance of the​​​‌ various model elements (e.g.​ in the case of​‌ dynamic network models, brain​​ areas and their connections).​​​‌ Such statistics can also​ be used as a​‌ basis to derive “group​​ statistical models” describing a​​​‌ family of task-related models​ that can in turn​‌ be used to constrain​​ models used in Section​​​‌ 3.2 and, through a​ forward model to reduce​‌ the dimension of the​​ search space at sensor​​​‌ level (see Section 3.1​). The BCI community​‌ is still divided on​​ the actual benefits in​​​‌ terms of accuracy on​ using the source space​‌ (v.s. the sensor space)​​ for classifying brain signals.​​​‌ The statistical tools developed​ in this sub-axis may​‌ help to address this​​ problem.

5.1.1 BCI-VIZAPP​​​‌

• Name:
  Real time EEG‌ applications
• Keyword:
  EEG
• Functional‌​‌ Description:
  BCI-VIZAPP is a​​ software suite for designing​​​‌ real-time EEG applications such‌ as BCIs or neurofeedback‌​‌ applications. It has been​​ been developed to build​​​‌ a virtual keyboard for‌ typing text and a‌​‌ photodiode monitoring application for​​ checking timing issues, but​​​‌ can now be also‌ used in other tasks‌​‌ such as EEG monitoring.​​ Originally, it was designed​​​‌ to delegate signal acquisition‌ and processing to OpenViBE‌​‌ but has recently been​​ extended to get some​​​‌ of these capabilities. This‌ allows for more integrated‌​‌ and robust applications but​​ also opens up new​​​‌ algorithmic opportunities, such as‌ real time parameter modification,‌​‌ more controlled interfaces, ...​​
• News of the Year:​​​‌
  Bci-Vizapp has been enriched‌ with numerous features, notably‌​‌ to read and save​​ certain files created with​​​‌ OpenViBE or with MNE-python,‌ thus allowing an easier‌​‌ communication with them. Bci-Vizapp​​ is also the software​​​‌ base used (or which‌ we aim to use‌​‌ it in the long​​ term) for different current​​​‌ (Demagus and ConnectTC) or‌ past (Techicopa) contracts and‌​‌ has integrated different elements​​ to support them.
• Contact:​​​‌
  Théodore Papadopoulo
• Participants:
  Nathanael‌ Foy, Théodore Papadopoulo, Maureen‌​‌ Clerc Gallagher, Come Le​​ Breton, Evgenia Kartsaki, Romain​​​‌ Lacroix, Jean-Luc Szpyrka, Julien‌ Wintz, 5 anonymous participants‌​‌

5.1.2 Tractography.jl

• Keywords:
  GPGPU,​​ Diffusion MRI
• Functional Description:​​​‌
  Tractography.jl is a high-performance‌ Julia package for brain‌​‌ tractography that leverages parallel​​ computing and specialized hardware​​​‌ (e.g., GPUs) to reconstruct‌ white matter fiber bundles‌​‌ from diffusion-weighted MRI data.​​ This enables researchers to​​​‌ study the structural connectivity‌ of the brain at‌​‌ unprecedented scales.
• URL:
  https://­github.­com/­cronos-inria/­Tractography.­jl​​
• Contact:
  Romain Veltz
• Participants:​​​‌
  Romain Veltz, Samuel Deslauriers-Gauthier,‌ Evgenia Kartsaki

5.1.3 tractography‌​‌

• Name:
  tractography
• Keywords:
  Tractography,​​ Diffusion MRI
• Scientific Description:​​​‌

  The tractography package provides‌ a high-performance computational framework‌​‌ for reconstructing the complex​​ 3D trajectories of white​​​‌ matter pathways from diffusion‌ Magnetic Resonance Imaging (dMRI)‌​‌ data.

  This software addresses​​​‌ the computational bottlenecks commonly​ associated with structural connectomics​‌ and neuroimaging research. Specifically,​​ the package implements advanced​​​‌ tractography methods grounded in​ both stochastic differential equations​‌ and partial differential equations​​ reformulations of the tractography​​​‌ problem, enabling a robust​ and precise mapping of​‌ neural topology. To handle​​ the immense computational demands​​​‌ of large-scale, high-resolution dMRI​ datasets, the library is​‌ heavily optimized for parallel​​ execution, seamlessly supporting both​​​‌ multi-core CPU and hardware-accelerated​ GPU architectures.

• Functional Description:​‌
  `tractography` is a Python​​ package designed for performing​​​‌ tractography, the process of​ reconstructing nerve fiber pathways​‌ from diffusion MRI (dMRI)​​ data. Developed by the​​​‌ Inria CRONOS team, the​ package implements various tractography​‌ algorithms and is optimized​​ for high performance, supporting​​​‌ execution on both CPUs​ and GPUs.
• URL:
  https://­gitlab.­inria.­fr/­cronos/­software/­tractography​‌
• Contact:
  Samuel Deslauriers-Gauthier
• Participants:​​
  Samuel Deslauriers-Gauthier, Romain Veltz,​​​‌ Evgenia Kartsaki

5.1.4 BifurcationKit​

• Name:
  Automatic computation of​‌ numerical bifurcation diagrams
• Keywords:​​
  Bifurcation, GPU
• Functional Description:​​​‌

  This Julia package aims​ at performing automatic bifurcation​‌ analysis of possibly large​​ dimensional equations function of​​​‌ a real parameter by​ taking advantage of iterative​‌ methods, dense / sparse​​ formulation and specific hardware​​​‌ (e.g. GPU).

  It incorporates​ continuation algorithms (PALC, deflated​‌ continuation, ...) based on​​ a Newton-Krylov method to​​​‌ correct the predictor step​ and a Matrix-Free/Dense/Sparse eigensolver​‌ is used to compute​​ stability and bifurcation points.​​​‌

  The package can also​ seek for periodic orbits​‌ of Cauchy problems. It​​ is by now one​​​‌ of the few software​ programs which provide shooting​‌ methods and methods based​​ on finite differences or​​​‌ collocation to compute periodic​ orbits.

  The current package​‌ focuses on large-scale, multi-hardware​​ nonlinear problems and is​​​‌ able to use both​ Matrix and Matrix Free​‌ methods on GPU.

• News​​ of the Year:
  Development​​​‌ of a Proof of​ Concept (POC) on boundary​‌ value problems (BVP) (with​​ Inria SAM SED). Improved​​​‌ calculation of normal shapes​ of periodic orbits. Correction​‌ of the calculation of​​ normal shapes of equilibrium​​​‌ points.
• URL:
  https://­github.­com/­rveltz/­BifurcationKit.­jl
• Publication:​
  hal-02902346
• Contact:
  Romain Veltz​‌
• Participant:
  an anonymous participant​​

5.1.5 SynapseElife

• Keyword:
  Markov​​​‌ model
• Functional Description:

  This​ is the main library,​‌ written in Julia language,​​ for the simulation of​​​‌ a synapse model. The​ associated publication is 10.7554/eLife.80152.​‌ It implements a model​​ of excitatory synapse in​​​‌ the rat hippocampus.

  This​ code allows to replicate​‌ the results of the​​ paper. It is also​​​‌ used by the Julia​ community to benchmark the​‌ methods for simulating piecewise​​ deterministic Markov processes.

• URL:​​​‌
  https://­github.­com/­rveltz/­SynapseElife.­jl
• Publication:
  hal-04275554
• Contact:​
  Romain Veltz
• Participant:
  Romain​‌ Veltz

5.1.6 OpenMEEG

• Keywords:​​
  Health, Neuroimaging, Medical imaging​​​‌
• Scientific Description:
  OpenMEEG provides​ a symmetric boundary element​‌ method (BEM) implementation for​​ solving the forward problem​​​‌ of electromagnetic propagation over​ heterogeneous media made of​‌ several domains of homogeneous​​ and isotropic conductivities. OpenMEEG​​​‌ works for the quasistatic​ regime (frequencies < 100Hz​‌ and medium diameter <​​ 1m).
• Functional Description:
  OpenMEEG​​​‌ provides state-of-the art tools​ for modelling bio-electromagnetic propagation​‌ in the quasi-static regime.​​ It is based on​​​‌ the symmetric BEM for​ the EEG/MEG forward problem,​‌ with a distributed source​​ model. OpenMEEG has also​​ been used to model​​​‌ the forward problem of‌ ECoG, for modelling nerves‌​‌ or the cochlea. OpenMEEG​​ is a free, open​​​‌ software written in C++‌ with python bindings. OpenMEEG‌​‌ is used through a​​ command line interface, but​​​‌ is also interfaced in‌ graphical interfaces such as‌​‌ BrainStorm, FieldTrip or SPM.​​
• Release Contributions:
  OpenMEEG has​​​‌ had small updates and‌ bug corrections. Notably, bugs‌​‌ related to the python​​ interface and to Blas/Lapack​​​‌ implementations have been handled.‌
• URL:
  http://­openmeeg.­github.­io/
• Publications:
  inria-00467061v2‌​‌, inria-00584205v1, hal-01278377v1​​
• Contact:
  Théodore Papadopoulo
• Participants:​​​‌
  Eric Larson, Théodore Papadopoulo,‌ 8 anonymous participants

6.1.1 Electrophysiology and BCI‌​‌

6.1.2 Diffusion‌​‌ MRI

6.2.1 Understanding Brain‌ Functional Connectivity

6.2.2 Dynamical models towards​‌ Whole Brain Models

Whole​​ brains models are models​​​‌ of the whole brain​ modeled using a graph​‌ where each node is​​ used to describe a​​​‌ cortical area and where​ the edges represent the​‌ connections between the cortical​​ areas. Whole brains models​​​‌ are very attractive from​ a modeling point of​‌ view because we can​​ use data such as​​​‌ structural connectivity from diffusion​ MRI in place of​‌ the graph edges. One​​ needs to assign a​​ dynamics to each node​​​‌ in order to study‌ the structure-function mapping and‌​‌ thus study the link​​ between fMRI and MRI.​​​‌ Traditionally, heuristically derived models‌ of the dynamics of‌​‌ populations of neurons are​​ used for the dynamics​​​‌ of the nodes. However,‌ the model of neural‌​‌ populations can be rigorously​​ derived from the dynamics​​​‌ of large networks of‌ interconnected neurons making the‌​‌ link direct between microscopic​​ elements and mesoscopic components.​​​‌

6.2.3 Clinical applications​​

Alignment​​ of Brain Networks

Participants:​​​‌ Yanis Aeschlimann, Samuel‌ Deslauriers-Gauthier, Théodore Papadopoulo‌​‌, Anna Calissano [University​​ College of London].​​​‌

Every brain is unique,‌ having its structural and‌​‌ functional organization shaped by​​ both genetic and environmental​​​‌ factors over the course‌ of its development. Brain‌​‌ image studies tend to​​ produce results by averaging​​​‌ across a group of‌ subjects, under a common‌​‌ assumption that it is​​ possible to subdivide the​​​‌ cortex into homogeneous areas‌ while maintaining a correspondence‌​‌ across subjects. This project​​ questions such an assumption:​​​‌ can the structural and‌ functional properties of a‌​‌ specific region of an​​ atlas be assumed to​​​‌ be the same across‌ subjects? In this work,‌​‌ this question is addressed​​ by looking at the​​​‌ network representation of the‌ brain, with nodes corresponding‌​‌ to brain regions and​​ edges to their structural​​​‌ relationships. Structural connectivity is‌ one view of brain‌​‌ networks provided by dMRI,​​ but these networks can​​​‌ also be observed from‌ a functional point of‌​‌ view via fMRI. We​​ thus propose to simultaneously​​​‌ exploit structural and functional‌ information in the alignment‌​‌ process. This allows us​​ to explore multiple perspective​​​‌ of brain networks. Our‌ results show that the‌​‌ permutations induced by one​​ type of connectivity (structural,​​​‌ functional) are not always‌ supported by the other‌​‌ connectivity network, but when​​ considering a combined alignment,​​​‌ it is possible to‌ find permutations of regions‌​‌ which are supported by​​​‌ both connectome modalities, leading​ to an increased similarity​‌ of functional and structural​​ connectivity across subjects.

This​​​‌ work has been published​ in 9, 22​‌ and in the thesis​​ of Yanis Aeschlimann 21​​​‌.

Investigating the Cognitive​‌ Drivers of Auditory Attention​​ Detection

Participants: Joan Belo​​​‌, Maureen Clerc,​ Daniele Schön [Institut de​‌ Neurosciences des Systèmes].​​

While M/EEG-based Auditory Attention​​​‌ Detection (AAD) can identify​ which audio stream a​‌ user is focusing on,​​ performance varies significantly between​​​‌ individuals. This work investigated​ the hypothesis that executive​‌ functions—specifically sustained attention, working​​ memory, and attentional inhibition—underlie​​​‌ this variability. We designed​ a challenging paradigm using​‌ dichotic polyphonic piano excerpts​​ presented to 41 participants​​​‌ with varying musical expertise.​

Our results demonstrate that​‌ attentional inhibition is a​​ significant predictor of AAD​​​‌ performance, explaining 6% of​ reconstruction accuracy and 8%​‌ of classification accuracy. Surprisingly,​​ neither musical expertise nor​​​‌ other executive functions showed​ a significant impact. These​‌ findings indicate that cognitive​​ control mechanisms directly affect​​​‌ the robustness of neural​ auditory representations, providing crucial​‌ insights for the development​​ of next-generation, neuro-steered hearing​​​‌ aids.

This work was​ published in 10.​‌

Improving Generative Fairness in​​ VAEs on Imbalanced Data​​​‌

Participants: Aymene Mohammed Bouayed​ [Be-Ys Research], Samuel​‌ Deslauriers-Gauthier, Adrian Iaccovelli​​ [Be-Ys Research], David​​​‌ Naccache [DIENS, ENS, CNRS,​ PSL University].

Variational​‌ Autoencoders (VAE) utilizing global​​ priors typically mirror the​​​‌ class frequencies of the​ training set within the​‌ latent space. This characteristic​​ leads to the underrepresentation​​​‌ of tail classes and​ reduced generative fairness when​‌ applied to imbalanced datasets.​​ While existing methods improve​​​‌ robustness via heavy-tailed Student's​ t-distribution priors, they continue​‌ to allocate latent volume​​ proportionally to class frequency.​​​‌

In this work, we​ address this limitation by​‌ explicitly enforcing equitable latent​​ space allocation. We propose​​​‌ Conditional-VAE, a method that​ defines a per-class Student's​‌ t joint prior over​​ latent and output variables​​​‌ to prevent dominance by​ majority classes. The model​‌ is optimized using a​​ closed-form objective derived from​​​‌ the β-power divergence. Furthermore,​ to ensure class-balanced generation,​‌ we derive an equal-weight​​ latent mixture of Student's​​​‌ t-distributions. Experimental results on​ standard databases (SVHN-LT, CIFAR100-LT,​‌ and CelebA) demonstrate that​​ Conditional-VAE consistently achieves lower​​​‌ Fréchet inception distance scores​ than both VAE and​‌ Gaussian-based VAE baselines, particularly​​ under severe class imbalance.​​​‌ In per-class F1 evaluations,​ the proposed approach substantially​‌ improves generative fairness and​​ diversity compared to conditional​​​‌ Gaussian VAEs in highly​ imbalanced regimes.

This work​‌ is available as a​​ preprint 23.

CNN​​​‌ Explainability with Multivector Tucker​ Saliency Maps for Self-Supervised​‌ Models

Participants: Aymene Mohammed​​ Bouayed [Be-Ys Research],​​​‌ Samuel Deslauriers-Gauthier, Adrian​ Iaccovelli [Be-Ys Research, France]​‌, David Naccache [DIENS,​​ ENS, CNRS, PSL University]​​​‌.

Interpreting the decisions​ of Convolutional Neural Networks​‌ (CNN) is essential for​​ understanding their behavior, yet​​ it remains a significant​​​‌ challenge, particularly for self-supervised‌ models. Most existing methods‌​‌ for generating saliency maps​​ rely on reference labels,​​​‌ restricting their use to‌ supervised tasks. The EigenCAM‌​‌ approach is the only​​ notable label-independent alternative, leveraging​​​‌ singular value decomposition to‌ generate saliency maps applicable‌​‌ across CNN models, but​​ it does not fully​​​‌ exploit the tensorial structure‌ of feature maps. In‌​‌ this work, we introduce​​ the Tucker Saliency Map​​​‌ (TSM) method, which applies‌ Tucker tensor decomposition to‌​‌ better capture the inherent​​ structure of feature maps,​​​‌ producing more accurate singular‌ vectors and values. These‌​‌ are used to generate​​ high-fidelity saliency maps, effectively​​​‌ highlighting objects of interest‌ in the input. We‌​‌ further extend EigenCAM and​​ TSM into multivector variants—Multivec-EigenCAM​​​‌ and Multivector Tucker Saliency‌ Maps (MTSM)—which utilize all‌​‌ singular vectors and values,​​ further improving saliency map​​​‌ quality. Quantitative evaluations on‌ supervised classification models demonstrate‌​‌ that TSM, Multivec-EigenCAM, and​​ MTSM achieve competitive performance​​​‌ with label-dependent methods. Moreover,‌ TSM enhances interpretability by‌​‌ approximately 50% over EigenCAM​​ for both supervised and​​​‌ self-supervised models. Multivec-EigenCAM and‌ MTSM further advance state-of-the-art‌​‌ interpretability performance on self-supervised​​ models, with MTSM achieving​​​‌ the best results.

This‌ work has been published‌​‌ in 11.

FinalSpark:‌​‌ collaborative research agreement -​​ 2024-2028

Participants: Patricia Reynaud-Bouret​​​‌ [DR CNRS, LJAD],‌ Paula Pousinha [MCF, IPMC]‌​‌, Ingrid Bethus [PR,​​ IPMC], Evgenia Kartsaki​​​‌, Gilles Scarella [IG,‌ CNRS, LJAD], Gregorio‌​‌ Rebecchi [PhD student, LJAD]​​, Romain Veltz.​​​‌

This is a collaborative‌ research agreement on the‌​‌ study of the “deformation​​ of the neuronal network​​​‌ by synaptic plasticity”. G.‌ Rebecchi is a Ph.D.‌​‌ student, since October 1st,​​ 2024 under the supervision​​​‌ of P. Reynaud-Bouret and‌ I. Bethus on the‌​‌ topic described above. This​​ PhD is funded by​​​‌ CNRS and by FinalSpark.‌

1. Description of the transfer.‌​‌Knowledge transfer. This is​​ a collaborative research agreement​​​‌ between FinalSpark, Université‌ Côte d'Azur, Inria and‌​‌ CNRS. The Parties wishes​​ to collaborate in order​​​‌ to understand more precisely‌ the synaptic connectivity of‌​‌ the neurospheres and how​​ it can be manipulated.​​​‌
2. Transfer modalities. It is‌ based on a contract.‌​‌ The study is on-going.​​
3. Contribution. R. Veltz is​​​‌ one of the parties‌ on the contract as‌​‌ an expert in dynamical​​ systems and synapses models.​​​‌

Demagus: BPI Grant‌​‌ April 2022–September 2025

Participants:​​ Laura Gee, Éléonore​​​‌ Haupaix-Birgy, Noémie Gonnier‌, Côme Le Breton‌​‌, Théodore Papadopoulo,​​ Julien Wintz.

This​​​‌ grant has been obtained‌ with the startup Mag4Health,‌​‌ CNRS and INSERM (see​​ Section 3.1.3). This​​​‌ company develops a new‌ MEG machine working with‌​‌ optically pumped magnetometers (magnetic​​ sensors), which potentially means​​​‌ lower costs and better‌ measurements. Cronos is in‌​‌ charge of developing a​​ real time interface for​​​‌ signal visualization, source reconstruction‌ and epileptic spikes detection.‌​‌ The initial funding was​​ for 2 years, but​​​‌ three extensions of 6‌ months have been obtained.‌​‌ This work is related​​​‌ to the research goals​ described in Section 3.1.3​‌.

SynPlasTool

Participants: Romain Veltz​, Hélène Marie [IPMC]​‌.

• Title:
  Development of​​ the Synapse Plasticity Tool​​​‌ for prediction of synapse​ plasticity outcome in silico.​‌
• Coordinator Name :
  Hélène​​ Marie
• Partner Institution:
  IPMC​​​‌
• Date/Duration:
  2024-2026

SynPlasTool is​ funded by FC3R, obtained​‌ in collaboration with H.​​ Marie (IPMC), 37500 euros.​​​‌ The main goal is​ to make easily accessible​‌ (web platform, ...) to​​ experimentalists the model that​​​‌ we developed 10.7554/eLife.80152.​ It is mainly a​‌ programming project that relies​​ on the software described​​​‌ in 5.1.5.

8.1.1​ Associate Teams in the​‌ framework of an Inria​​ International Lab or in​​​‌ the framework of an​ Inria International Program

8.2.1 Visits of international‌​‌ scientists

• In September, L.​​ Florack and L. Smolders​​​‌ from Eindhoven Technological University‌ visited the team to‌​‌ discuss their work on​​ geodesic tractography and brain​​​‌ structure–function mapping.
• P. Ford-Dominey‌ visited the team on‌​‌ several occations to discuss​​ input-driven functional connectivity. These​​​‌ discussion led to an‌ ANR proposal currently in‌​‌ review.

ANR ChaMaNe, Mathematical Challenges​​​‌ in Neurosciences

Participants: Romain‌ Veltz.

• Coordinator Name‌​‌ :
  Delphine Salort (Sorbonne​​ Université)
• Partner Institutions:
  • Partner​​​‌ 1: Biologie Computationnelle et‌ Quantitative (CQB), France
  • Partner‌​‌ 2: LJAD, Université Nice,​​ France
• Date/Duration:
  2020-2025
• Web​​​‌ site

This project aimed‌ at a mathematical study,‌​‌ on the one hand,​​ of the intrinsic dynamics​​​‌ of a neuron and‌ their consequences, and on‌​‌ the other hand, of​​ the qualitative dynamics of​​​‌ large neural networks with‌ respect to the intrinsic‌​‌ behavior of the individual​​ neurons, the interactions between​​​‌ them, memory effects, spatial‌ structure, etc.

FHU InnovPain‌​‌ 2

Participants: Petru Isan​​, Théodore Papadopoulo,​​​‌ Romain Veltz.

• Coordinator‌ Name :
  Denys Fontaine‌​‌ (Pasteur Hospital)
• Date/Duration:
  2023–2027​​
• Web site

The FHU​​​‌ INOVPAIN is a Hospital-University‌ Federation focused on chronic‌​‌ refractory pain and innovative​​ therapeutic solutions. It involves​​​‌ 6 hospital facilities, 13‌ academic research teams and‌​‌ 12 platforms and core​​ facilities.

Inria-Inserm fellowship

Participants:​​​‌ Laura Gee, Théodore‌ Papadopoulo, Christian Bénar‌​‌.

• Date/Duration:
  2024–2027

The​​ Ph.D. of L. Gee​​​‌ is funded by an‌ Inria-Inserm fellowship (see section‌​‌ 6 for a description​​ of the work).

ANR​​​‌ NeuroMotor

Participants: Samuel Deslauriers-Gauthier‌.

• Coordinator Name :‌​‌
  François Hug (LAMHESS, Université​​ Côte d'Azur)
• Date/Duration:
  2024–2027​​​‌

There are many physical‌ disabilities with a neurological‌​‌ origin, such as stroke​​ and spinal cord injuries,​​​‌ that significantly impact a‌ person’s mobility, physical capacity,‌​‌ stamina, or dexterity. In​​ the era of personalized​​​‌ medicine, optimizing the treatment‌ of these physical disabilities‌​‌ requires: i) accessing direct​​ information on the neural​​​‌ commands that are sent‌ to the muscles, and‌​‌ ii) integrating this knowledge​​ into the development and​​​‌ assessment of rehabilitation and‌ neurotechnology aimed at restoring‌​‌ movement. The breakthrough of​​ our approach lies in​​​‌ changing the level at‌ which we observe the‌​‌ control of movement, i.e.,​​ shifting focus from the​​​‌ level of whole muscles‌ to the spinal (alpha)‌​‌ motor neurons. To achieve​​ this, we will combine​​​‌ the use of dense‌ grids of surface EMG‌​‌ electrodes with algorithms that​​ decode the firing activity​​​‌ of spinal motor neurons.‌ By unraveling the “neural‌​‌ code” for movement generation,​​ we will address critical​​​‌ gaps in our understanding‌ of the control of‌​‌ movement in health and​​ disease. In this project,​​​‌ we will decode the‌ activity of large populations‌​‌ of motor neurons from​​ different muscles. We will​​​‌ identify motor neuron synergies,‌ defined as functional groups‌​‌ of motor neurons that​​ share inputs from various​​​‌ supraspinal, spinal and sensory‌ sources. In this translational‌​‌ project, we will examine​​ the structure and plasticity​​​‌ of these synergies in‌ both healthy controls and‌​‌ patients with neurological impairments​​​‌ (stroke, spinal cord injury).​ We will also enhance​‌ the electrode design to​​ facilitate the transfer of​​​‌ these methods into clinical​ settings.

ANR spinPAIN-PATH

Participants:​‌ Romain Veltz.

• Coordinator​​ Name :
  Emmanuel Deval​​​‌ (IPMC, Université Côte d'Azur)​
• Date/Duration:
  2025–2029

The main​‌ objective of this project​​ is to better understand​​​‌ spinal pain processes, which​ still remain to be​‌ fully deciphered. It will​​ focus on the role​​​‌ of particular ion channels,​ Acid-Sensing Ion Channels (ASICs),​‌ highly expressed in the​​ spinal pain neuronal network​​​‌ but where their functions​ have yet to be​‌ fully elucidated. Indeed, if​​ their role in pain​​​‌ has been largely documented​ in peripheral sensory neurons,​‌ far less is known​​ in the central nervous​​​‌ system (CNS), particularly at​ the spinal cord level.​‌ This project is based​​ on preliminary data, demonstrating​​​‌ (i) a large expression​ of particular ASIC subunits​‌ in spinal neurons both​​ in mouse and human,​​​‌ and (ii) a potent​ analgesic effect of specific​‌ ASIC blockers injected intrathecally​​ in mice. Importantly, the​​​‌ in vivo analgesic effect​ observed in mice with​‌ the ASIC1a inhibitor, mambalgin-1,​​ can be as strong​​​‌ as that of morphine​ and partially independent of​‌ the endogenous opioid system,​​ opening new potential therapeutic​​​‌ perspectives in a context​ of world opioid crisis.​‌ In line with the​​ preliminary data, the specific​​​‌ aims are 1) to​ localize the different ASIC​‌ subunits in the different​​ neuronal populations and characterize​​​‌ their association in the​ spinal dorsal horn (SDH)​‌ of both mice and​​ humans, 2) to assess​​​‌ their overall contribution to​ the SDH network activity​‌ by combining in vitro​​ primary culture model with​​​‌ computational modeling and, 3)​ to explore the in​‌ vivo and ex vivo​​ contribution of spinal ASICs​​​‌ to long-term sensitization processes,​ with a particular focus​‌ on inhibitory interneurons, in​​ neuropathic pain condition, which​​​‌ needs new therapeutic perspectives​ and where spinal ASICs​‌ have been already involved.​​

NeuroMod​​​‌

Participants: Rodrigo Cofre Torres​, Samuel Deslauriers-Gauthier,​‌ Olivier Faugeras, Evgenia​​ Kartsaki, Théodore Papadopoulo​​​‌, Romain Veltz.​

• Direction:
  T. Papdopoulo and​‌ I. Bethus.
• Web site​​.

The NeuroMod Institute​​​‌ for Modeling in Neuroscience​ and Cognition aims at​‌ promoting modeling as an​​ approach for integrating brain​​​‌ mechanisms and cognitive functions.​ To meet this central​‌ challenge of integration in​​ cognitive science, the institute​​​‌ relies on the interdisciplinary​ resources available within the​‌ Université Côte d’Azur: more​​ than 250 researchers and​​​‌ 16 laboratories.

Université Côte​ d'Azur encourages interaction between​‌ human sciences (psychology, behavioral​​ economics, language sciences), modeling​​​‌ (computer science, mathematics, physics,​ etc.) and neuroscience (biology,​‌ neurophysiology, cognitive neuroscience, medicine,​​ etc.). NeuroMod is unique​​​‌ because of the variety​ of approaches used to​‌ integrate different models at​​ multiple levels, from neurons​​​‌ and their molecular mechanisms,​ to cognitive processes and​‌ behaviors, through neural network​​ dynamics.

NeuroMod also provides​​​‌ interdisciplinary training for Master​ students, PhD students and​‌ future faculty members. We​​ opened an international elite​​​‌ degree in Modeling for​ Neuronal and Cognitive Systems​‌ (Master of Science Mod4NeuCog)​​ and a national Master's​​ degree in Cognitive Science.​​​‌

Cronos is deeply involved‌ in this regional initiative‌​‌ through its researchers, but​​ also through its engineer​​​‌ E. Kartsaki, who devotes‌ part of its time‌​‌ to software projects of​​ the institute.

PSI ion​​​‌ channels

Participants: Théodore Papadopoulo‌, Romain Veltz.‌​‌

• Coordinator Name :
  M.​​ Mantegazza, E. Lingueglia (IPMC,​​​‌ Université Côte d'Azur)
• Date/Duration:‌
  2025–2031

The Programme Stratégique‌​‌ IdEx (PSI) Ion Channels​​ capitalizes on Université Côte​​​‌ d’Azur’s teams already present‌ in former LABEX ICST‌​‌ 1.0 and 2.0 projects​​ (F. Lesage, M. Mantegazza,​​​‌ E. Lingueglia & E.‌ Deval, and E. Honoré‌​‌ from IPMC, G. Sandoz​​ from iBV and L.​​​‌ Counillon from LP2M) to‌ include new teams (from‌​‌ IPMC (H. Marie/J. Barik,​​ M. Chami/A. DaCosta, P.​​​‌ Blancou/T. Simon) and iBV‌ (O. Soriani), strengthening basic‌​‌ and translational research in​​ biology, but also from​​​‌ Inria (T. Papadopoulo) and‌ the J. A. Dieudonné‌​‌ laboratory (P. Reynaud-Bouret) for​​ mathematical modeling, and UR2CA​​​‌ (D. Fontaine/M. Lanteri-Minet) for‌ clinics) in order to‌​‌ fully exploit the University’s​​ strengths and extend its​​​‌ scope to build a‌ multidisciplinary approach targeting the‌​‌ pathophysiology of ion channels​​ and excitability5.​​​‌

The PSI has several‌ objectives:

• Train PhD students‌​‌ offering competitive PhD fellowships.​​
• Boost Université Côte d’Azur’s​​​‌ International development and visibility.‌
• Develop technological innovations.
• Strengthen‌​‌ interaction with industry and​​ the University Hospital.
• Help​​​‌ the evolution of Université‌ Côte d’Azur’s teams.

The‌​‌ landscape of the PSI​​ Ion Channels will cover​​​‌ from basic science to‌ translation and innovation through‌​‌ cutting edge techniques (ex​​ vivo & in vivo​​​‌ electrophysiology, neurostimulation, optopharmacology, transcriptomics,‌ high-speed imaging), integrative science‌​‌ (neuroinflammation, mouse and zebrafish​​ models, computational modeling and​​​‌ AI), pathophysiology (pain, migraine,‌ cancer, heart disease, autism,‌​‌ intellectual disabilities, Alzheimer's disease,​​ mental disorders, epilepsy) and​​​‌ therapy (neurostimulation, pharmacology, precision‌ medicine).

9.1.1 Scientific​​​‌ events: organisation

9.1.2 Journal

9.1.3​​ Invited talks

• Romain Veltz​​​‌ , “Theoretical / numerical​ study of modulated traveling​‌ pulses in inhibition stabilized​​ networks”, ANR ChaMaNe, February​​​‌ 2025
• Théodore Papadopoulo “Brain​ Computer Interfaces… a field​‌ with many challenges and​​ opportunities”, MOMI workshop, Sophia​​​‌ Antipolis, May 26th, 2025.​
• Romain Veltz , “Pros​‌ / cons of mean​​ field representations of large​​​‌ size networks of spiking​ neurons”, CAMDAM workshop, Montreal,​‌ Canada, June 4th, 2025.​​
• Théodore Papadopoulo “Estimating and​​​‌ Exploiting Brain Dynamics”, BMW​ workshop, Frejus, June 17th,​‌ 2025.
• Romain Veltz ,​​ “SYNPLASTOOL : development of​​​‌ a GUI tool for​ the prediction of synapse​‌ plasticity outcome in silico”,​​ Replacement in Neuroscience, November​​​‌ 2025 (online, 700 persons).​
• Rodrigo Cofre , “Computational​‌ and Neurobiological Modeling of​​ Structure-Function Coupling in Different​​​‌ Consciousness States”, presented at​ the NeuroMod Meeting 2025,​‌ Antibes, July 8th, 2025.​​
• Samuel Deslauriers-Gauthier presented at​​​‌ the Neuromod Institute Open​ Day.

9.1.4 Leadership within​‌ the scientific community

• Olivier​​ Faugeras is a member​​​‌ of the French Academy​ of Sciences.
• Rachid Deriche​‌ is a member of​​ Academia Europaea.

9.1.5 Scientific​​​‌ expertise

• Théodore Papadopoulo was​ member of a panel​‌ in the ANR-NSF CNS​​ grant selection.
• Théodore Papadopoulo​​​‌ was selected to be​ a member of a​‌ EIC Pathfinder grant selection​​ panel. Due to a​​​‌ conflict of interest, he​ had to resign.
• Théodore​‌ Papadopoulo reviewed project proposals​​ for Université de Toulouse​​​‌ and Université Côte d'Azur.​
• Théodore Papadopoulo is the​‌ scientific referent of Yuxiu​​ Shao, recently hired at​​​‌ Université Côte d'Azur on​ a NeuroMod-LJAD CPJ (Chaire​‌ de Professeur Junior) position.​​
• Samuel Deslauriers-Gauthier reviewed project​​​‌ proposals for the Institut​ des Neurosciences cliniques de​‌ Rennes.

9.1.6 Research administration​​

• Théodore Papadopoulo is the​​​‌ director of the NeuroMod​ Institute.
• Théodore Papadopoulo​‌ is a member of​​ the Neuromod scientific council​​​‌ and represents Neuromod in​ the EUR Healthy (EUR:​‌ Universitary Research School) and​​ in the Maison de​​​‌ la Simulation of Université​ Côte d'Azur.
• Théodore Papadopoulo​‌ is a member of​​ the steering committee of​​​‌ the IHU RespiERA.​
• Théodore Papadopoulo is the​‌ alternate Inria representative at​​ the CRBSP (Comité de​​​‌ la recherche en matière​ biomédicale et de santé​‌ publique) of the University​​ Hospital of Nice.

9.2.1 Teaching​​

• Master: Théodore PapadopouloInverse​​​‌ problems for brain functional​ imaging, 3h ETD,​‌ M2, Mathématiques, Vision et​​ Apprentissage, ENS Cachan, France.​​​‌
• Master: Théodore Papadopoulo and​ Samuel Deslauriers-Gauthier , Functional​‌ Brain Imaging, each​​ 15h ETD, M1,M2 in​​​‌ the MSc Mod4NeuCog of​ Université Côte d'Azur.
• Master:​‌ Théodore Papadopoulo and Samuel​​ Deslauriers-Gauthier , Application of​​​‌ machine learning to MRI,​ electophysiology & brain computer​‌ interfaces, each 10h​​ ETD, M1, M2 in​​​‌ the MSc Data Science​ and Artificial Intelligence of​‌ Université Côte d'Azur.
• Master:​​ Samuel Deslauriers-Gauthier and Romain​​​‌ Veltz , Introduction to​ Python programming and simulation​‌, each 15h ETD,​​ M1, MSc Mod4NeuCog of​​ Université Côte d'Azur.
• Master:​​​‌ Romain Veltz , Mathematical‌ methods for neuroscience,‌​‌ 24h ETD, M2, Mathématiques,​​ Vision et Apprentissage, ENS​​​‌ Cachan, France.
• Master: Rodrigo‌ Cofre , Scientific Communication‌​‌, 24h ETD, M2,​​ MSc Mod4NeuCog of Université​​​‌ Côte d'Azur, France.
• Master:‌ Rodrigo Cofre , Introduction‌​‌ to modeling in neuroscience​​ and cognition, 8h​​​‌ ETD, M1, MSc Mod4NeuCog‌ of Université Côte d'Azur,‌​‌ France.
• Master: Rodrigo Cofre​​ , Digital expertise I​​​‌, 8h ETD, M1,‌ MSc SmartEdTech of Université‌​‌ Côte d'Azur, France.
• Master:​​ Rodrigo Cofre , Digital​​​‌ expertise II, 8h‌ ETD, M2, MSc SmartEdTech‌​‌ of Université Côte d'Azur,​​ France.

9.2.2 Supervision

• PhD​​​‌ defended in December: Yanis‌ Aeschlimann , "Brain networks‌​‌ from simultaneous modelling of​​ functional MRI and diffusion​​​‌ MRI", started in Oct.‌ 2022. Supervisors: Samuel Deslauriers-Gauthier‌​‌ , Théodore Papadopoulo  21​​.
• PhD in progress:​​​‌ Petru Isan , "Cerebral‌ eletrophysiological exploitation of in‌​‌ vivo evoked potentials to​​ guide the resection of​​​‌ adult brain tumors", started‌ in Oct. 2023. Supervisor:‌​‌ F. Almairac. Co-Supervisors: Théodore​​ Papadopoulo , Samuel Deslauriers-Gauthier​​​‌ .
• PhD in progress:‌ Émeline Manka , "Modeling‌​‌ of evoked potentials by​​ direct current stimulation and​​​‌ their links to electromyography",‌ started in Oct. 2024.‌​‌ Supervisors: Samuel Deslauriers-Gauthier ,​​ Théodore Papadopoulo .
• PhD​​​‌ in progress: Laura Gee‌ , "Exploring and exploiting‌​‌ the new capabilities of​​ room temperature MEG sensors",​​​‌ started in Dec. 2024.‌ Supervisors: Théodore Papadopoulo ,‌​‌ C.Bénar (INSERM, Marseille).
• PhD​​ in progress: A. Bouayed,​​​‌ "Fast, interpretable and fair‌ image generation on imbalanced‌​‌ data", to be defended​​ in March 2026. Supervisors:​​​‌ D. Naccache, A. Iaccovelli,‌ Samuel Deslauriers-Gauthier .
• PhD‌​‌ in progress: P. Castro,​​ "New approaches in modeling​​​‌ and fMRI data analysis‌ of the brain in‌​‌ different states of consciousness",​​ to be defended in​​​‌ September 2027. Supervisors: B.‌ Jarraya (NeuroSpin, Paris), A.‌​‌ Destexhe (NeuroPsi, Paris), Rodrigo​​ Cofre .
• Romain Veltz​​​‌ supervised the M2 student‌ A. Ackay on "Bifurcation‌​‌ Theory in Whole-Brain Models:​​ Exploring Cortical Dynamics for​​​‌ Wake and Sleep State‌ Transitions", November 2024 to‌​‌ February 2025.
• Samuel Deslauriers-Gauthier​​ and Romain Veltz supervised​​​‌ the M2 internship (InterMaths‌ Network) of H. Harshit‌​‌ on "A Finite Volume​​ Framework for Probabilistic Tractography",​​​‌ April 2025 to August‌ 2025.
• Samuel Deslauriers-Gauthier and‌​‌ Rodrigo Cofre supervised the​​ M2 internship (MSc Mod4NeuCog​​​‌ of Université Côte d'Azur)‌ of P. Rice on‌​‌ "Optimization of whole brain​​ models to simulate the​​​‌ effect of anesthesia".
• Samuel‌ Deslauriers-Gauthier supervised the M2‌​‌ internship (MSc Mod4NeuCog of​​ Université Côte d'Azur) of​​​‌ T. Stei on "Predicting‌ task-related functional connectivity from‌​‌ structural connectivity".
• Théodore Papadopoulo​​ supervised the M2 student​​​‌ J. Feriau on "Uncertainty‌ propagation: From electroencephalogram measurements‌​‌ to BCI classification".
• Rodrigo​​ Cofre supervised the M1​​​‌ student D. Zuniga on‌ "Dynamical Structure-Function Correlations of‌​‌ fMRI Human Brain Signals​​ under LSD".
• Samuel Deslauriers-Gauthier​​​‌ supervised the M1 internship‌ (MSc Mod4NeuCog of Université‌​‌ Côte d'Azur) of M.​​ Shaw on "Inverse problems​​​‌ in electromyography (EMG)".

9.2.3‌ Juries

• Olivier Faugeras was‌​‌ a member of several​​ committees awarding prizes from​​​‌ the French Academy of‌ Sciences.
• Théodore Papadopoulo participated‌​‌ as a reviewer in​​​‌ the PhD jury of​ I. Siviero at Université​‌ of Verona on June​​ 19th, 2025.
• Théodore Papadopoulo​​​‌ participated as a reviewer​ in the PhD jury​‌ of S. Reynaud at​​ IMT Atlantique in Brest​​​‌ on September 29th, 2025.​
• Théodore Papadopoulo participated as​‌ a reviewer in the​​ PhD jury of H.​​​‌ Agouram at University of​ Aix-Marseille on December 9th,​‌ 2025.
• Théodore Papadopoulo and​​ Samuel Deslauriers-Gauthier participated in​​​‌ the PhD jury of​ Y. Aeschlimann at Université​‌ Côte d'Azur on December​​ 15th, 2025.
• Romain Veltz​​​‌ participated in the PhD​ jury of A. Rossi​‌ at Université Côte d'Azur​​ on September 25th, 2025.​​​‌
• Samuel Deslauriers-Gauthier participated in​ the PhD jury of​‌ L. Smolders at Eindhoven​​ University of Technology on​​​‌ January 12th, 2025.
• Samuel​ Deslauriers-Gauthier was part of​‌ the M2 jury for​​ the MSc Mod4NeuCog of​​​‌ Université Côte d'Azur.
• Rodrigo​ Cofre participated in the​‌ Comité de suivi de​​ thèse (CST) of I.​​​‌ Mindlin: PhD student at​ the Sorbonne University (ED3C​‌ Cerveau - Cognition -​​ Comportement), PhD under the​​​‌ direction of J. Sitt,​ ICM Paris.
• Rodrigo Cofre​‌ participated in the CST​​ of C. Picard: PhD​​​‌ student at the Paris-Saclay​ University (Ècole doctorale Biosigne),​‌ under the direction of​​ V. Ego-Stern, NeusoPsi Saclay.​​​‌
• Rodrigo Cofre participated in​ the CST of L.​‌ Martineau: PhD student at​​ IRMA Strasbourg University (Ècole​​​‌ doctorale Mathèmatiques, Sciences de​ l'information et de l'ingenieur),​‌ under the direction of​​ S. Geffray et C.​​​‌ Pouzat.
• Rodrigo Cofre participated​ in the CST of​‌ T. Hardy: PhD student​​ at Université Paris-Cité (INCC​​​‌ UMR 8002, CNRS, 75006​ Paris), under the direction​‌ of C. Sergent.
• Théodore​​ Papadopoulo participated in the​​​‌ CST of L. Abdalah:​ PhD student at Université​‌ Côte d'Azur, under the​​ direction of V. Zarzoso​​​‌ and W. Da Cruz​ Freitas.
• Romain Veltz participated​‌ in the CST of​​ A. Bavoil: PhD student​​​‌ at Université Côte d'Azur,​ under the direction of​‌ J.B. Caillay and A.​​ Nême.
• Romain Veltz participated​​​‌ in the CST of​ P. Izan: PhD student​‌ at Université Côte d'Azur,​​ under the direction of​​​‌ F. Almairac and Samuel​ Deslauriers-Gauthier .
• Romain Veltz​‌ participated in the CST​​ of G. Rebecchi: PhD​​​‌ student at Université Côte​ d'Azur, under the direction​‌ of P. Reynaud-Bouret and​​ I. Bethus.

9.3.1 Specific official responsibilities​ in science outreach structures​‌

• Romain Veltz is Science​​ Outreach Officer at the​​​‌ Inria Centre at Université​ Côte d'Azur since 2025.​‌ He was responsible for​​ the Chiche program,​​​‌ organized the Café'In, coordinated​ Intro, and ran the​‌ one-week MATHC2+ internship for​​ 10th-grade students.

9.3.2 Participation​​​‌ in Live events

• Romain​ Veltz has given a​‌ one hour conference to​​ the Fête de la​​​‌ Science at Juan-les-Pins in​ front of a general​‌ audience on October 11th,​​ 2025.
• Rodrigo Cofre has​​​‌ given a one hour​ talk at Cafe'In INRIA​‌ "Explorer la conscience avec​​ des drogues : De​​​‌ l’anesthésie générale aux psychédéliques​ en thérapie? " on​‌ April 24th, 2025.
• Rodrigo​​ Cofre has given a​​​‌ one hour conference to​ the Fête de la​‌ Science at Juan-les-Pins in​​ front of a general​​ audience on October 10th,​​​‌ 2025.
• Romain Veltz gave‌ 8 chiches (one hour‌​‌ each) to 10th-grade students.​​
• Théodore Papadopoulo and I.​​​‌ Bethus gave an interview‌ at BFM TV Côte‌​‌ d'Azur in the contexts​​ Brain's week and NeuroMod.​​​‌

International journals

• 9​​ articleY.Yanis Aeschlimann​​​‌, A.Anna Calissano‌, T.Theodore Papadopoulo‌​‌ and S.Samuel Deslauriers-Gauthier​​. Inter-individual alignment of​​​‌ multimodal brain networks with‌ anatomical constraints.Network‌​‌ Neuroscience2025HALDOI​​back to text
• 10​​​‌ articleJ.Joan Belo‌, M.Maureen Clerc‌​‌ and D.Daniele Schön​​. Attentional inhibition ability​​​‌ predicts neural representation during‌ challenging auditory streaming.‌​‌Cognitive, Affective, and Behavioral​​​‌ Neuroscience2025HALDOI​back to text
• 11​‌ articleA. M.Aymene​​ Mohammed Bouayed, S.​​​‌Samuel Deslauriers-Gauthier, A.​Adrian Iaccovelli and D.​‌David Naccache. CNN​​ Explainability with Multivector Tucker​​​‌ Saliency Maps for Self-Supervised​ Models.Transactions on​‌ Machine Learning Research Journal​​February 2025HALback​​​‌ to text
• 12 article​O.Olivier Faugeras and​‌ E.Etienne Tanré.​​ Universality of the mean-field​​​‌ equations of networks of​ Hopfield-like neurons.Journal​‌ of Mathematical Biology91​​4September 2025,​​​‌ 51In press. HAL​DOIback to text​‌
• 13 articleG.Guylaine​​ Hoffner, P.Pablo​​​‌ Castro, L.Lynn​ Uhrig, C. M.​‌Camilo Miguel Signorelli,​​ M.Morgan Dupont,​​​‌ J.Jordy Tasserie,​ A.Alain Destexhe,​‌ R.Rodrigo Cofre,​​ J.Jacobo Sitt and​​​‌ B.Béchir Jarraya.​ Transcranial direct current stimulation​‌ modulates primate brain dynamics​​ across states of consciousness​​​‌.eLife13October​ 2025, RP101688HAL​‌DOIback to text​​
• 14 articleA.Andrea​​​‌ Luppi, L.Lynn​ Uhrig, J.Jordy​‌ Tasserie, P.Pedro​​ Mediano, F.Fernando​​​‌ Rosas, S. P.​S. Parker Singleton,​‌ D.Daniel Gutierrez-Barragan,​​ S.Silvia Gini,​​​‌ P.Pablo Castro,​ C.Camilo Signorelli,​‌ D.Daniel Golkowski,​​ A.Andreas Ranft,​​​‌ R.Rüdiger Ilg,​ D.Denis Jordan,​‌ K.Kanako Muta,​​ J.Junichi Hata,​​​‌ H.Hideyuki Okano,​ Z.-Q.Zhen-Qi Liu,​‌ Y.Yohan Yee,​​ A.Alain Destexhe,​​​‌ R.Rodrigo Cofre,​ D.David Menon,​‌ A.Alessandro Gozzi,​​ B.Bechir Jarraya and​​​‌ E.Emmanuel Stamatakis.​ Convergent transcriptomic and connectomic​‌ controllers of information integration​​ and its anaesthetic breakdown​​​‌ across mammalian brains.​Nature Human BehaviourJanuary​‌ 2026HALDOI
• 15​​ articleA.Aurora Rossi​​​‌, Y.Yanis Aeschlimann​, E.Emanuele Natale​‌, S.Samuel Deslauriers-Gauthier​​ and P.Peter Ford​​​‌ Dominey. Characterizing Dynamic​ Functional Connectivity Subnetwork Contributions​‌ in Narrative Classification with​​ Shapley Values.Network​​​‌ Neuroscience2025, 1-25​HALDOIback to​‌ text
• 16 articleM.​​Michael Tangermann, S.​​​‌Sylvain Chevallier, M.​Matthias Dold, P.​‌Pierre Guetschel, R.​​Reinmar Kobler, T.​​​‌Theodore Papadopoulo and J.​Jordy Thielen. Learning​‌ from small datasets --​​ review of workshop 6​​​‌ of the 10th International​ BCI Meeting 2023.​‌Journal of Neural Engineering​​223June 2025​​​‌, 033001HALDOI​back to text

International​‌ peer-reviewed conferences

• 17 inproceedings​​O.Olivier Bisson,​​​‌ Y.Yanis Aeschlimann,​ S.Samuel Deslauriers-Gauthier and​‌ X.Xavier Pennec.​​ Log-Euclidean Frameworks for Smooth​​​‌ Brain Connectivity Trajectories.​Lecture Notes in Computer​‌ Science (LNCS)GSI'25 -​​ International Conference on Geometric​​​‌ Science of Information16035​Saint-Malo (France), France2025​‌, 194–204HALDOI​​back to text

Conferences​​​‌ without proceedings

• 18 inproceedings​ Y.Yanis Aeschlimann,​‌ S.Samuel Deslauriers-Gauthier and​​ R.Romain Veltz.​​​‌ GPU tractography: What can​ we learn from half​‌ a trillion streamlines? International​​ Society for Tractography Conference​​ - IST 2025 Bordeaux​​​‌ (France), France October 2025‌ HAL back to text‌​‌ back to text
• 19​​ inproceedingsS.Samuel Deslauriers-Gauthier​​​‌ and R.Romain Veltz‌. A principled mathematical‌​‌ study of the limit​​ of fiber tractography.​​​‌International Society for Tractography‌ Conference - IST 2025‌​‌Bordeaux (France), FranceOctober​​ 2025HALback to​​​‌ textback to text‌

Scientific book chapters

• 20‌​‌ inbookE.Etienne St-Onge​​, G.Gabriel Girard​​​‌, K.Kurt Schilling‌, A.Alessandro Daducci‌​‌, S.Samuel Deslauriers-Gauthier​​, L.Laurent Petit​​​‌ and M.Maxime Descoteaux‌. Improving tractography using‌​‌ anatomical priors &amp; multimodal​​ integration.Handbook of​​​‌ Diffusion MR TractographyElsevier‌2025, 347-362HAL‌​‌DOIback to text​​

Doctoral dissertations and habilitation​​​‌ theses

• 21 thesisY.‌Yanis Aeschlimann. Multimodal‌​‌ brain network alignment to​​ correct for inter-subject variability​​​‌ with an application to‌ structure-function mapping.Université‌​‌ Côte d'AzurDecember 2025​​HALback to text​​​‌back to text

Reports‌ & preprints

• 22 misc‌​‌Y.Yanis Aeschlimann,​​ A.Anna Calissano,​​​‌ T.Théodore Papadopoulo and‌ S.Samuel Deslauriers-Gauthier.‌​‌ Brain network alignment using​​ structural and functional connectivity​​​‌ with anatomical constraints.‌January 2025HALback‌​‌ to text
• 23 misc​​A. M.Aymene Mohammed​​​‌ Bouayed, S.Samuel‌ Deslauriers-Gauthier, A.Adrian‌​‌ Iaccovelli and D.David​​ Naccache. Conditional-${\mathrm{t‌}}^3$VAE: Equitable Latent‌ Space Allocation for Fair‌​‌ Generation.September 2025​​HALback to text​​​‌
• 24 miscR.Rui‌ Dai, R.Rodrigo‌​‌ Cofre, C.Christopher​​ Timmermann, R. L.​​​‌Robin L Carhart-Harris,‌ A. G.Anthony G‌​‌ Hudetz, Z.Zirui​​ Huang and G. A.​​​‌George A Mashour.‌ A Mesoscale Framework for‌​‌ Psychedelic Drug Action in​​ the Human Brain.​​​‌November 2025HALDOI‌back to text
• 25‌​‌ miscP.Pascal Helson​​, E.Etienne Tanré​​​‌ and R.Romain Veltz‌. Mean-field analysis of‌​‌ a neural network with​​ stochastic STDP.2025​​​‌HALback to text‌
• 26 miscH. M.‌​‌Hugo M Martin,​​ R.Rodrigo Cofre and​​​‌ A.Alain Destexhe.‌ Simulated 5-HT2A receptor activation‌​‌ accounts for the high​​ complexity of brain activity​​​‌ during psychedelic states.‌October 2025HALback‌​‌ to text
• 27 misc​​I.Iván Mindlin,​​​‌ C.Carlos Coronel-Oliveros,‌ J. D.Jacobo D‌​‌ Sitt, R.Rodrigo​​ Cofre, A.Andrea​​​‌ Luppi, T.Thomas‌ Andrillon, Y. S.‌​‌Yonatan Sanz Perl and​​ R.Rubén Herzog.​​​‌ The impact of homeostatic‌ inhibitory plasticity in a‌​‌ generative biophysical model.​​January 2026HALDOI​​​‌back to text
• 28‌ miscR.Romain Veltz‌​‌. Analysis of a​​ mean-field limit of interacting​​​‌ two-dimensional nonlinear integrate-and-fire neurons‌.2025HALDOI‌​‌back to text

Other​​ scientific publications

• 29 inproceedings​​​‌H.Harshit Harshit,‌ S.Samuel Deslauriers-Gauthier and‌​‌ R.Romain Veltz.​​ A Finite Volume Framework​​​‌ for Probabilistic Tractography: Solving‌ the Fiber Transport PDE‌​‌.Neuromod meeting, 2025​​Antibes (06), FranceJuly​​​‌ 2025HALback to‌ text
• 30 inproceedingsR.‌​‌Romain Veltz and S.​​​‌Samuel Deslauriers-Gauthier. A​ mathematical study of the​‌ limit of fiber tractography​​.Neuromod meeting, 2025​​​‌Antibes (06), FranceJuly​ 2025HALback to​‌ text

Scientific popularization

• 31​​ inproceedingsE.Emeline Manka​​​‌, S.Samuel Deslauriers-Gauthier​, T.Théodore Papadopoulo​‌, P.Petru Isan​​ and F.Fabien Almairac​​​‌. Modeling of the​ stimulation artifact and brain​‌ evoked potential elicited by​​ direct stimulation: OHBM 2025,​​​‌ 24-28 June, Brisbane, Australia​.OHBM 2024 -​‌ Organization of Human Brain​​ MappingBrisbane (AU), Australia​​​‌June 2025HALback​ to text