Phd defense on 15-12-2025

ED Mathématiques et Informatique

• Exploring and designing interactive physical-digital tools for quantum physics education

  by Vincent CASAMAYOU (LaBRI - Laboratoire Bordelais de Recherche en Informatique)

  The defense will take place at 14h00 - Ada Lovelace 200 Av. de la Vieille Tour, Centre Inria de l'Université de Bordeaux, 33405 Talence

  in front of the jury composed of

  • Martin HACHET - Directeur de recherche - Inria - Directeur de these
  • Iza MARFISI - Professeure des universités - Le Mans Université - Rapporteur
  • Christophe DAUSSY - Maître de conférences - Université Sorbonne Paris Nord - Rapporteur
  • Lionel CANIONI - Professeur - Université de Bordeaux - CoDirecteur de these
  • Christophe SCHLICK - Professeur - Université de Bordeaux - Examinateur
  • Manuel JOFFRE - Directeur de recherche - CNRS - Examinateur

  Summary

  Quantum physics is currently evolving with the emergence of new quantum technologies that open a wide range of opportunities for future concrete applications. However, the actual development of such promising applications requires new generations of students to be trained, not only as researchers, but also as engineers and technicians. To reach this goal, education in quantum physics plays central role and should evolve to meet these emerging needs. The problem with quantum physics education lies in underlying concepts that are often abstract, counterintuitive and theories based on a complex mathematical ground. Consequently, teaching quantum physics remains challenging. Among the tools that help students to build complex knowledge, experimental practices and digital materials have shown undeniable benefits toward student conceptual understanding. In this thesis, we explore how quantum physics can leverage digital tools to foster experimental practices. We propose new approaches to practice and learn quantum physics based on a real-time digital simulation of quantum optics experimental setups through multiples format (tangible and completely digital). With these new tools, simulation parameters can be directly manipulated using actionable components, mimicking actual quantum optics experiments. Over and above the practical advantages (cost, implementation, danger), this approach enables the simulation to be enhanced by the addition of digital pedagogical information that links the concrete experiment to its theoretical foundation. The first part of this thesis focuses on the evolution of teaching practices in quantum physics, as well as the use of digital materials in education. Then, it presents a series of three projects, co-designed with the experts in education and physics. The first project, called HOBIT, consists in a setup that mixes tangible interaction and augmented reality to simulate an optical bench where learners manipulate physical reproductions of optics elements to perform wave optics experiments. The second project, SHIRE, is an extension of the HOBIT experience on a full digital software that can be run at home or in classrooms on learner's personal computers. Despite the absence of tangible interactions, this approach expands the range of pedagogical scenarios in which the simulation can take place and opens new opportunities. This project also includes a user-study of the software with physics teachers. Finally, HQBIT, discussed in the final part of this thesis, integrates quantum physics simulation and pedagogical augmentations into the two previous projects. The objective is to develop a hybrid digital tool that can effectively help the training of a new generation of students in quantum optics and experimentation. A user-study with students in quantum physics completement this project and provides a first evaluation of this tool. To conclude this thesis, I explore the perspectives that HQBIT brings in terms of collaboration, assessments of its learning outcomes, and evolution towards quantum technologies.

• Model-based approaches and reinforcement learning for robust and autonomous locomotion of humanoid robots in dynamic contexts

  by Marc DUCLUSAUD (LaBRI - Laboratoire Bordelais de Recherche en Informatique)

  The defense will take place at 10h00 - 178 351 Cr de la Libération, Bâtiment A30, 33405 Talence

  in front of the jury composed of

  • Olivier LY - Professeur des universités - LaBRI Université de Bordeaux - Directeur de these
  • Vincent PADOIS - Directeur de recherche - INRIA Université de Bordeaux - CoDirecteur de these
  • Stéphane CARON - Chargé de recherche - INRIA Paris - Rapporteur
  • Marcos MAXIMO - Directeur de recherche - Aeronautics Institute of Technology (ITA) - Rapporteur
  • Pierre BLAZEVIC - Professeur des universités - ISTY UVSQ Paris-Saclay - Examinateur

  Summary

  This thesis is set in a period of unprecedented growth in humanoid robotics, driven both by rapid technological advances and by a growing enthusiasm from private actors and the general public. In such a context, the recent progress in humanoid robot locomotion, together with their increasing social acceptance, seems to announce their forthcoming integration into real environments alongside humans. Faced with this ambition, a major challenge remains: ensuring the robustness and autonomy of locomotion in the dynamic contexts in which these robots will have to operate. To address this challenge, this work combines physical modeling and reinforcement learning, leveraging the complementary advantages of these two paradigms — the stability guarantees provided by modeling and the adaptability arising from learning. This manuscript begins with a state of the art of control and learning methods applied to humanoid locomotion, aimed at identifying the most promising approaches to reconcile robustness, adaptability, and dynamic realism. On these foundations, the PlaCo software, dedicated to motion planning and robot control, is developed. It aims to abstract the complexity of the optimization formulations required for trajectory generation, while maintaining performance compatible with real-time execution. This framework is then used to design and deploy on the humanoid robot Sigmaban a walking controller based on the Linear Inverted Pendulum Model (LIPM). This development highlights the ability of this model to produce coherent trajectories in real time, while revealing the practical limitations encountered on a real platform. To overcome these limitations and enable dynamic adaptation to disturbances, a reinforcement learning agent dedicated to fall recovery is developed. Trained in simulation, this agent is successfully transferred to the real robot, demonstrating a significant gain in autonomy. However, the difficulty of this transfer highlights the central issue of the gap between simulated and real environments. This observation leads to an investigation of how this gap can be reduced by improving simulation fidelity. A detailed study of friction phenomena in servo actuators is therefore carried out, showing how a more accurate consideration of these effects improves both the quality of simulation and the transferability of control policies.

ED Sciences Chimiques

• Physico-chemical aging processes and hydration properties of secondary aerosols at the single particle scale

  by Gwendoline BOURDON (Institut des Sciences Moléculaires)

  The defense will take place at 9h30 - Salle de conférences, 3ème étage EST Université de Bordeaux, Campus Peixotto, Bâtiment A12, ISM, UMR 5255 CNRS, 351 cours de la Libération, 33405 Talence Cedex, France

  in front of the jury composed of

  • Sophie SOBANSKA - Directrice de recherche - Université de Bordeaux - CoDirecteur de these
  • Eric VILLENAVE - Professeur des universités - Université de Bordeaux - CoDirecteur de these
  • Céline TOUBIN - Professeure des universités - Université de Lille - Rapporteur
  • Marcello BRIGANTE - Professeur des universités - Université Clermont Auvergne - Rapporteur
  • Stéphane COUSSAN - Chargé de recherche - Université de Marseille - Examinateur
  • Valérie RAVAINE - Professeure des universités - Bordeaux INP - Examinateur

  Summary

  For several decades, atmospheric aerosols have been at the forefront of societal and scientific concerns due to their significant impacts on climate change and human health. Given the inherent complexity of aerosols—arising from their variable composition and morphology—as well as the diversity of their physicochemical properties (e.g., hygroscopicity, viscosity, pH) and the multitude of atmospheric processes in which they participate (such as photo-oxidation and condensation), a comprehensive understanding of their behavior requires the implementation of complementary experimental and theoretical approaches across multiple spatial and temporal scales. This multiscale strategy is essential for achieving a detailed characterization of aerosol systems and for accurately assessing their atmospheric fate and environmental impacts. At the single-particle level, levitation techniques offer unique advantages for investigating the physicochemical properties and transformation processes of aerosols, notably by eliminating substrate-induced artifacts and enabling fine control of environmental parameters (e.g., particle suspension, temperature, and relative humidity). Within this context, the present PhD thesis aimed to elucidate the chemical aging mechanisms (e.g., photolysis, oxidation) and hydration properties of model biogenic secondary organic aerosols (BSOAs), generated via the photo-oxidation of α-pinene, with the ultimate goal of improving the current understanding of their fate and impact in the atmosphere. The interaction of individual suspended particles with water vapor was characterized through two complementary approaches: monitoring the evolution of hygroscopic growth as a function of relative humidity, and analyzing water evaporation kinetics under fixed humidity conditions. To address these objectives, we used three original experimental setups. The coupling of acoustic or optical levitation with Raman microspectrometry enabled real-time, in situ analysis of the chemical composition and microphysical state of levitated particles. This configuration was used for both hydration studies and aging experiments of BSOAs. In particular, the acoustic levitation system developed at ISM facilitated the investigation of hydration behavior and photolytic degradation of individual cis-pinonic acid particles. The optical levitation setup was optimized at Hiroshima University (Japan) as part of a collaborative initiative. It allowed for the trapping and chemical analysis of secondary organic aerosols formed in situ from gaseous precursors, for the first time at the single-particle scale. In parallel, a new coupling between acoustic levitation and high-resolution proton-transfer-reaction time-of-flight mass spectrometry (PTR-TOF-MS) was developed during this thesis. This system enabled to observe the direct transfer of volatile organic compounds (VOCs) released from levitated particles during the photolysis of cis-pinonic acid. This innovative approach opens new perspectives for the investigation of dynamic gas-particle exchanges at the single-particle scale, thereby contributing to a more comprehensive understanding of atmospheric aerosol aging mechanisms.

• Exploring Chirality in Aromatic Oligoamide Foldamers: From Molecular Design to Chiroptical Probing in Solution and on Surfaces

  by Rita BORGES ANASTACIO (Institut des Sciences Moléculaires)

  The defense will take place at 13h30 - Amphithéâtre European Institute of Chemistry and Biology (IECB) 2 Rue Robert Escarpit, 33600 Pessac

  in front of the jury composed of

  • Céline OLIVIER - Directrice de recherche - Université de Bordeaux - Directeur de these
  • Jeanne CRASSOUS - Directrice de recherche - CNRS Institut des Sciences Chimiques de Rennes - Rapporteur
  • Yann FERRAND - Directeur de recherche - Université de Bordeaux - Directeur de these
  • Benoît CHAMPAGNE - Professeur des universités - Namur Institute of Structured Matter - Rapporteur
  • Sébastien GOEB - Directeur de recherche - UMR CNRS - MOLTECH-Anjou - Examinateur
  • Vincent RODRIGUEZ - Professeur des universités - Université de Bordeaux - Examinateur

  Summary

  Chirality plays a crucial role at the molecular level in diverse fields such as biology, biochemistry, catalysis and pharmacology. To probe chirality in solution, well-established spectroscopic techniques such as optical rotatory dispersion (ORD) and electronic circular dichroism (ECD) are commonly employed. However, these linear optical methods face limitations in the direct detection of chirality on surfaces and interfaces. In contrast, nonlinear optical processes such as second-harmonic generation (SHG) are inherently surface-sensitive. The SHG responses of enantiomeric helices of variable lengths were first evaluated in solution using hyper-Rayleigh scattering (HRS) with quinoline-based oligoamide foldamers (QOFs) as chiral molecular models. These studies confirmed that HRS can serve as a powerful complementary chiroptical method, providing additional structural information compared to classical techniques, in particular, sensitivity to the number of helical turns. To extend this approach to the solid state, chiral thiolated QOFs of variable length and electronic structure were prepared and immobilized on gold substrates. Homogeneous monolayers were obtained, and their chiral signatures were investigated using reflectivity-SHG (an adapted method for detecting SHG on surfaces). Comparable features to those observed in solution emerged, in particular, noticeable differences in the responses of enantiomers and their corresponding racemic mixture when probed with linearly polarized light. Furthermore, the spin-filtering properties of these systems, known as the chirality-induced spin selectivity (CISS) effect, were evaluated via magnetic conductive atomic force microscopy, revealing additional aspects of surface chirality. As part of a related study, a short rigid helicene segment was introduced as an alternative chiral inducer to bias the handedness of QOF sequences. Its effect was first investigated in sequences exclusively forming single helices, revealing solvent-dependent diastereoselectivity. This approach was subsequently extended to foldamer sequences capable of assembling into double-helical architectures, enabling the study of chirality in higher-order helices.

• Development of helical aromatic oligoamide foldamers for protein recognition

  by Florian SANCHEZ (Institut de Chimie & de Biologie des Membranes & des Nano-objets)

  The defense will take place at 14h00 - L0.010 Butenandtstr. 5-13, Haus L 81377 Munich

  in front of the jury composed of

  • Lucile FISCHER - Chargée de recherche - Université de Bordeaux - Directeur de these
  • Vladimir TORBEEV - Professeur - Université de Strasbourg - Rapporteur
  • Evgeny KATAEV - Professeur - Université Friedrich Alexander Erlangen-Nuremberg - Rapporteur
  • Ivan HUC - Professeur - Université Ludwig-Maximilian Munich - CoDirecteur de these
  • Anne SCHÜTZ - Professeure - Université Ludwig-Maximilian Munich - Examinateur

  Summary

  Protein-protein interactions (PPIs) are key regulators of cellular signaling and represent promising yet complex therapeutic targets. Advances in chemical biology and structural computational tools now enable the design of molecules to modulate these interfaces. Inspired by natural biopolymer folding, aromatic oligoamide foldamers (AOFs) offer synthetically accessible, stable helical frameworks with predictable side-chain orientation, making them attractive candidates for functional design and drug discovery. This thesis presents the design and synthesis of novel building blocks for AOFs, enhancing the chemical diversity of the quinoline (Q) scaffold through targeted substitutions at positions 4, 5, 6, and 4,6. New methodologies enabled the incorporation of diverse biogenic side chains: cationic, anionic, polar, and hydrophobic. These new building blocks broaden the capabilities of AOFs for biomolecular recognition. Mechanistic studies of solid phase foldamer synthesis (SPFS) led to optimised protocols for automated synthesiser, allowing efficient parallel production of up to three AOF sequences. Making use of the newly established protocols and biogenic side chains, an AOF was designed and synthesised to randomly recognise a library of protein (affitin and affibody), clones were identified and submicromolar binding was observed for the racemic Q12 AOF candidate. We discovered an AOF capable of binding two structurally distinct protein scaffolds: β-sheet-based (affiitn) and α-helical (affibody) selected via mRNA and phage display, respectively. Solution studies enabled truncation of the AOF to its minimal binding epitope. Notably, the AOF exhibits enantioselective recognition of both targets, driven by its intrinsic P-handedness. In an attempt to target the bio-relevant interaction between linear di-ubiquitin and the coiled-coil domain of NEMO, AOFs candidates were designed to mimic the interaction of the coiled-coil on the surface of the ubiquitin. Each AOF incorporated covalent linkers via triphosgene activation and an activated disulphide for site-specific ligation. Solution-state NMR confirmed local environmental perturbations upon ligation, although initial crystallisation attempts were unsuccessful. Modifications to linker length and side-chain composition enabled reproducible crystal formation, yet structural resolution remained limited due to poor diffraction and asymmetric unit complexity. In conclusion, this work highlights the synthetic versatility and functional potential of aromatic oligoamide foldamers (AOFs). The integration of diverse biogenic side chains and streamlined automated protocols enabled the generation of AOFs with submicromolar affinities for distinct protein targets. These findings underscore the promise of AOFs as adaptable molecular tools for protein recognition and interface mimicry in biologically relevant contexts.

• Electrochemiluminescent imaging of individual objects

  by Leslie Rocio ARIAS ARANDA (Institut des Sciences Moléculaires)

  The defense will take place at 14h00 - ENSMAC or ISM A12 Univ. Bordeaux, CNRS UMR 5255, Bordeaux INP, Site ENSMAC, 33607 Pessac

  in front of the jury composed of

  • Laurent BOUFFIER - Directeur de recherche - Université de Bordeaux - Directeur de these
  • Neso SOJIC - Professor - Université de Bordeaux - CoDirecteur de these
  • Florence LAGARDE - Chargée de recherche - Université Claude Bernard de Lyon - Examinateur
  • Laurence VIGNAU - Professeure - Bordeaux INP, Université de Grenoble, IMS - Examinateur
  • Jérôme CHAUVIN - Professeur - Univeristé Grenoble Alpes - Rapporteur
  • Benoit LIMOGES - Directeur de recherche - ITOYS, Université de Paris Cité - Rapporteur

  Summary

  A combination of electrochemiluminescence (ECL) and bipolar electrochemistry (BP) has enabled the development of innovative analytical systems. ECL, which involves light emission from exergonic redox reactions, is a highly sensitive and versatile optical detection method. BP, on the other hand, allows remote electrode activation without direct wiring, promoting system simplification and miniaturization. Their integration has led to BP-ECL devices capable of generating light signals remotely, with high precision and efficiency, expanding their use in biochemistry, environmental monitoring, and portable diagnostics. This thesis explores the synergy between ECL and BP by studying their theoretical bases and applying them to the design of advanced experimental systems. First, the fundamental mechanisms of ECL generation, via annihilation and co-reactant pathways, are studied, alongside the criteria for a material to function as a bipolar electrode (BPE). Based on this, static BP-ECL configurations were developed to study how the spatial arrangement of BPEs affects ECL emission. It was shown that a frontal orientation of two BPEs promotes redox species mass transport and enhances light emission. Furthermore, it was possible to find another strategy for polarization mapping of BPE. Moreover, BPE rotation was introduced to analyse hydrodynamic effects on ECL via the co-reactant mechanism. Using a rotating disk bipolar electrode (RDBPE), unusual emission patterns such as "croissant" or ring shapes were observed. These patterns depended on the applied electric field, rotation speed, and reagent stoichiometry. Rotation studies were further extended to ECL via the annihilation mechanism, using alternating current at the feed electrodes to polarize the RDBPE. This enabled control over both intensity and spatial symmetry of the ECL in real-time. An achievement was the development of a 2D graphene rotor actuated by magnetohydrodynamic effects, capable of efficient and tunable ECL emission. This system integrates chemical, electrical, and mechanical controls on a single platform. Additionally, an endogenous BP strategy based on spontaneous redox reactions was demonstrated, enabling ECL without external power sources and achieving multicolor light emission. Overall, this work provides in-depth insights into the BP–ECL synergy and highlights its relevance for future technologies, from sensor development to autonomous optical systems.

• Ultra-bright Dye-based Fluorescent Organic Nanoparticles (dFONs) for 3D Bioimaging

  by Eleonore KUREK (Institut des Sciences Moléculaires)

  The defense will take place at 9h30 - Amphithéâtre B Bâtiment A29 351 Cours de la Libération 33405 TALENCE cedex

  in front of the jury composed of

  • Bo LAURSEN - Professeur des universités - Université de Copenhague - Kemisk Institut - Rapporteur
  • Suzanne FERY-FORGUES - Directrice de recherche - Université de Toulouse 3 Paul-Sabatier - SPCMIB (UMR 5068) - Rapporteur
  • Rémi MéTIVIER - Directeur de recherche - ENS Paris Saclay - PPSM (UMR8531) - Examinateur
  • Véronique JUBERA - Professeure des universités - Université de Bordeaux - ICMCB (UMR5026) - Examinateur

  Summary

  Fluorescence bioimaging employs fluorescent probes to enhance contrast in biological samples. In thick tissues and whole organisms, imaging depth is limited by scattering and absorption, but losses can be mitigated by increasing probe brightness, shifting emission toward the red to near-infrared (NIR) biooptical window, or using two-photon excitation for deeper penetration. Fluorescent nanoparticles made from organic dyes are increasingly valued for their high brightness, photostability, and hold promise for improved bio and eco compatibility relative to inorganic probes like quantum dots or lanthanide-doped particles. Unlike dye-doped matrices, dye-based fluorescent organic nanoparticles (dFONs) are entirely self-stabilized dye aggregates formed via nanoprecipitation from polar and polarizable dyes (PPDs). Their maximized dye content, coupled with smart molecular engineering to minimize aggregation caused quenching yield large brightness values, rivaling or even surpassing that of well-established inorganic quantum dots. This work focuses on the development of dye-based fluorescent organic nanoparticles with enhanced properties for bioimaging. The first section presents the design, characterization, and application of red-emitting dFONs for three-dimensional two-photon single-particle tracking (SPT). Built from based a two-photon-absorbing quadrupolar dye, and engineered for enhanced photostability, these dFONs were found to withstand intense excitation laser intensities for several minutes and, due to their small size and brightness, achieved better localization accuracy than standard inorganic SPT probes. The subsequent sections describe binary fluorescent organic nanoparticles (BFONs) composed of two distinct dyes - selected for excited-state energy transfer such as Förster Resonance Energy Transfer (FRET). The objective was to develop probes that combine red emission with a high fluorescence quantum yield - and thus high brightness - in an attempt to overcome the constraints imposed by the energy gap law. BFONs made from green- and red-emitting bis-bipolar dyes were prepared through sequential nanoprecipitation to form core-shell structures, enabling NanoInterfacial Emission Enhancement (NIEE) - an increase in the acceptor's fluorescence quantum yield upon donor-mediated excitation. Tuning the donor-to-acceptor ratio demonstrated a direct dependence of NIEE on BFONs composition. Live-cell confocal fluorescence microscopy showed that these BFONs maintained their photophysical properties - and thus their structural integrity - over 48h in vitro. Finally, BFONs composed of orange- and red-emitting quadrupolar dyes were prepared in a single step to achieve a uniform dye distribution within particles. These systems exhibited highly efficient FRET and a novel enhancement - Environmental Emission Enhancement (EEE) - matching NIEE values of core-shell BFONs, while requiring only a fraction of the acceptor content. This study underscores the role of molecular engineering and the potential of synergistic multi-dye integration in producing brighter, red-shifted dFONs - providing insights for the development of next generation fluorescent probes.

• Development and characterization of UHTC ceramic materials with specific reinforcements

  by Timothe MENARD (Laboratoire des Composites ThermoStructuraux)

  The defense will take place at 10h00 - Amphithéatre LCTS, 3 allée de la Boétie, 33600, Pessac

  in front of the jury composed of

  • Laurence MAILLé - Professeure des universités - Université de Bordeaux - Directeur de these
  • Frédéric BERNARD - Professeur des universités - Université de Bourgogne Europe - Rapporteur
  • Emmanuel DE BILBAO - Professeur des universités - Université Orléans - Rapporteur
  • Pierre-Jean PANTEIX - Ingénieur de recherche - Institut Jean Lamour - Examinateur
  • Romain LUCAS-ROPER - Professeur des universités - IRCER, Université de Limoges - Examinateur
  • Aurélie JULIAN JANKOWIAK - Ingénieure de recherche - ONERA - Examinateur
  • Aurélie QUET - Ingénieure de recherche - CEA DAM - Le Ripault - Examinateur
  • Francis REBILLAT - Professeur des universités - Université de Bordeaux - CoDirecteur de these

  Summary

  During atmospheric re-entry or hypersonic flight, the leading edges of vehicles are subjected to temperatures that can exceed 2000 °C. Carbon/carbon (Cf/Cm) composites are currently the benchmark materials for high-temperature space applications. However, their low resistance to oxidation above 400°C severely limits their service life. Ultra-high temperature ceramics (UHTC), with a melting point exceeding 3000°C, appear to be promising candidates for these extreme applications. The most widely studied systems are mainly carbides (ZrC, HfC) and borides (ZrB2, HfB2). Nevertheless, the oxidation of these materials, although passive, leads to the formation of porous oxides. In order to improve their behaviour, the addition of SiC promotes the formation of a glassy layer of SiO2, capable of filling the porosity and limiting the diffusion of oxidising species. The objective of this thesis is to develop materials that can withstand very high thermal and material fluxes. Flash sintering is the method used to densify UHTCs from the Zr–Si–C–B system, either reactively or non-reactively. In order to increase the toughness of the material, discontinuous reinforcements have been incorporated. The fibre/matrix interaction is a key factor that has been considered. Oxidation tests under high thermal and material flux, carried out using an oxyacetylene torch and a plasma torch, were used to evaluate the influence of microstructure and composition on oxidation resistance. Four-point bending mechanical tests were also conducted to describe the damageability of the composites produced.

ED Sciences de la Vie et de la Santé

• Effects of Type 2 Diabetes on Cardiac Pericyte Survival, Renewal and Phenotype.

  by Célia BOURGUIGNON (Biologie des maladies cardiovasculaires)

  The defense will take place at 14h00 - Salle de réunion de l'INSERM 1 av. Magellan, 33600 Pessac

  in front of the jury composed of

  • Marie-Ange RENAULT - Chargée de recherche - Université de Bordeaux - Directeur de these
  • Jean-François FAIVRE - Professeur des universités - Université de Poitiers - Laboratoire PRéTI - Examinateur
  • Elizabeth JONES - Professeure - Centre for Molecular and Vascular Biology de Louvain - Rapporteur
  • Eric CAMERER - Directeur de recherche - PARCC Paris – Cardiovascular Research Center - Rapporteur

  Summary

  Introduction. Pericytes are mural cells essential for maintaining vascular integrity. Type 2 Diabetes (T2D) is a metabolic disorder often associated with microvascular abnormalities. While its harmful effects on endothelial cells are well studied, the impact of T2D on cardiac pericytes remains poorly explored. Objective. This study aims to investigate how T2D affects cardiac pericyte biology by assessing their survival, regenerative capacity, and transcriptomic profile. Results. Cardiac pericyte density was assessed using two mouse models of T2D : in C57BL6/J mice subjected to a fed a high-fat diet combined with streptozotocin injections (HFD/STZ) for 3 or 6 months, and in 3-month-old Leprdb/db mice, harboring a spontaneous mutation resulting in leptin receptor deficiency. A significant reduction in cardiac pericyte was observed in both models as early as 3 months, suggesting that T2D negatively affects either the survival or renewal of these cells. In contrast, only brain pericytes in HFD/STZ mice treated for 6 months were reduced, indicating that the brain may be more resistant or better protected against T2D-related stress than the heart. In vitro, palmitate-induced lipotoxicity triggered necrosis in primary murine cardiac pericytes, whereas exposure to supraphysiological glucose concentrations, mimicking hyperglycemia, had no effect on their survival. Furthermore, in vivo, a high-fat diet regimen alone was sufficient to reproduce the pericyte loss observed in diabetic hearts, identifying lipotoxicity as a primary driver of this loss. To explore whether T2D impacts cardiac pericyte renewal, regenerative capacity was first assessed under non-diabetic conditions using an inducible pericyte-depletion mouse model (Pdgfrb-CreERT2;RosaDTA). In this model, cardiac pericytes efficiently reappeared within 14 to 21 days after their depletion. A similar phenomenon was observed in the brain, although the recovery was slower. Moreover, neither hyperglycemia nor hyperlipidemia had a negative impact on cell proliferation in the heart, 21 days after pericyte depletion. These findings suggest that T2D specifically compromises cardiac pericyte survival rather than their renewal. Unexpectedly, in both Pdgfrb-CreERT2;RosaDTA and HFD/STZ mice models, cardiac pericyte repopulation was not driven by proliferation of existing pericytes, but rather by proliferation of non-pericyte cells, predominantly endothelial cells. Finally, lipotoxic stress also altered the pericyte phenotype in vitro by inhibiting PDGF, NOTCH and TGF-β signaling pathways, which are critical for pericyte homeostasis. These changes were associated with reduced expression of Angpt1, indicating potential disruption in pericyte-endothelial cell communication. Conclusion. This study demonstrates that T2D, mainly through lipotoxicity mechanisms, impairs both the survival and phenotype of cardiac pericytes. These alterations likely contribute to the microvascular dysfunction observed in the diabetic heart. Importantly, these findings highlight the critical role of cardiac pericytes in T2D pathophysiology and suggest new therapeutic strategies aimed at preserving their function to maintain vascular integrity. Notably, in the Pdgfrb-CreERT2;RosaDTA model, pericyte reappearance was associated with decreased vascular permeability and reduced endothelial activation.

• CISCO : Characterization of PTEN expression Cis-regulatory elements involved in Cowden disease

  by Thibaut MATIS (BoRdeaux Institute of onCology)

  The defense will take place at 14h00 - Amphithéatre 12 Campus Carreire 146 rue Léo Saignat 33000 Bordeaux

  in front of the jury composed of

  • Nicolas SEVENET - Professeur des universités - praticien hospitalier - Université de Bordeaux - Directeur de these
  • Caroline ROORYCK-THAMBO - Professeure des universités - praticienne hospitalière - Université de Bordeaux - Examinateur
  • Sylvain FOISSAC - Chargé de recherche - INRAE - Rapporteur
  • Frederic CAUX - Professeur des universités - praticien hospitalier - Université Sorbonne Paris Nord - Examinateur
  • Martin TEICHMANN - Professeur des universités - Université de Bordeaux - Examinateur
  • Estelle COLIN - Professeure des universités - praticienne hospitalière - Université d'Angers - Rapporteur

  Summary

  Cowden disease, which belongs to the spectrum of PTEN Hamartoma Tumor Syndromes (PHTS), is an autosomal dominant genetic disorder caused by germline pathogenic variants (GPV) in the PTEN tumor suppressor gene. PTEN encodes a phosphatase that regulates the PI3K–AKT pathway, controlling cell proliferation, survival, and metabolism. About 20% of patients with a typical phenotype do not present a GPV in the coding sequences of PTEN, suggesting genetic or allelic heterogeneity, particularly involving distal cis-regulatory alterations. Negative results from analyses of promoter alterations, methylation, or exomes, as well as the presence of intragenic Alu elements, support the latter hypothesis. Genes are regulated by long-range cis-regulatory elements such as enhancers. While enhancer–promoter interaction defects are well established in some genetic diseases (e.g., β-thalassemia, developmental disorders), their implication in cancer predisposition syndromes remains poorly explored. The PTEN cis-regulatory elements and the architecture of its topologically associated domain (TAD) had until now remained poorly characterized. The aims of this thesis and the CISCO project (Characterization of PTEN expression Cis-regulatory elements involved in Cowden disease) were therefore to map chromatin interactions at the PTEN locus, to identify and functionally characterize enhancers in two reference cell lines (lymphocytic and mammary), and to assess the integrity of the TAD in tumor cell lines lacking PTEN expression (due to pathogenic variants (PV) or chromosomal rearrangements). The approach combined multi-omic datasets from ENCODE (chromatin accessibility, histone marks, eRNA transcription) with functional assays (CRISPRi) and 3D genome mapping (Hi-C and Tiled-C). In total, eleven enhancers were identified, organized around seven contact hubs, some interacting directly with the PTEN promoter, others interacting with neighboring enhancers, forming a complex regulatory network. While most enhancers and enhancer–promoter interactions were cell-type dependent, one enhancer (PE_11) was highly conserved across all cell lines. The CTCF proteins and cohesin contributed complementarily to the formation of these interactions, although about half occurred independently of their binding, suggesting alternative structural mechanisms. At high resolution, active enhancer marks localized near, but not directly overlapping, with contact hotspots, highlighting genomic regions that structurally bring enhancers close to the promoter but lack transcriptional function. Furthermore, eRNA transcription was oriented toward the PTEN promoter, indicating a functional directionality of enhancer–promoter interactions. Finally, the 3D architecture of the PTEN locus remained largely stable in the presence of exonic PVs, with hubs and enhancer–promoter interactions preserved. Only intra-TAD structural rearrangements led to notable alterations. Overall, this work highlights a conserved 3D architecture ensuring both stability and flexibility in enhancer engagement at the PTEN locus. Understanding PTEN regulation and identifying key regulatory regions open avenues for applications in medical genetics, particularly for PHTS patients without identifiable GPVs.

• A reverse clinical translation platform for the molecular stratification of vascular progression in Systemic Sclerosis (SSc)

  by Stefano DI DONATO (Immunologie Conceptuelle, Expérimentale et Translationnelle)

  The defense will take place at 15h30 - Salle Module 2.6 2ᵉ étage du bâtiment CROUS – Site de Carreire, 2 Rue Dr Hoffmann Martinot, 33076 Bordeaux

  in front of the jury composed of

  • Marie-Elise TRUCHETET - Professeure des universités - praticienne hospitalière - Université de Bordeaux - Directeur de these
  • Amélie SERVETTAZ - Professeure des universités - praticienne hospitalière - Université de Reims Champagne-Ardenne (URCA) - Rapporteur
  • Yannick ALLANORE - Professeur des universités - praticien hospitalier - Université Paris Cité - Examinateur
  • Vincent SOBANSKI - Professeur des universités - praticien hospitalier - Université de Lille - Rapporteur

  Summary

  Systemic sclerosis (SSc) is a complex autoimmune disease characterized by early vascular injury, persistent immune activation, and progressive fibrosis. This thesis combines clinical and experimental approaches to elucidate the central role of endothelial dysfunction in SSc pathogenesis. In the clinical section, we developed the DAVIX, a composite score designed to quantify vascular damage using magnetic resonance imaging data. This work identified the most representative components of systemic vascular injury and demonstrated the prognostic relevance of the DAVIX in patient stratification. In the experimental section, we investigated the molecular mechanisms underlying endothelial dysfunction, focusing on the cGAS–STING pathway. Using flow cytometry, we characterized circulating endothelial cells (CEC) and endothelial progenitor cells (EPC), revealing an imbalance between vascular injury and repair. In vitro studies on primary endothelial cells showed aberrant STING activation associated with mitochondrial stress and an enhanced interferon signature. Altogether, these findings support an integrated model in which the endothelium acts both as a target and a driver of immune-mediated vascular injury. This work provides a mechanistic framework for vascular damage in SSc and suggests novel therapeutic targets within endothelial immunity.

• Developing the new paradigm of ‘meta-plant pathogens' by studying the compositional and metabolomic dynamics of a Meta-Fusarium sp. exposed to abiotic and biotic perturbations

  by Valentin FIEVET (MycSA - Mycologie et Sécurité des Aliments)

  The defense will take place at 10h00 - Amphithéâtre de l'ISVV Institut des Sciences de la Vigne et Vin - 210 Chemin de Leysotte 33140 Villenave d'Ornon

  in front of the jury composed of

  • Florence RICHARD-FORGET - Directrice de recherche - Université de Bordeaux - Directeur de these
  • Claire LE HENAFF-LE MARREC - Professeure des universités - Université de Bordeaux - Examinateur
  • Caroline STRUB - Professeure des universités - Université de Montpellier - Examinateur
  • Nicolas CREUSOT - Chargé de recherche - INRAE Bordeaux-Nouvelle-Aquitaine - Examinateur
  • Aymé SPOR - Chargé de recherche - INRAE Bourgogne Franche-Comté - Rapporteur
  • Kris AUDENAERT - Full professor - Ghent University - Rapporteur

  Summary

  Fusarium species are causative agents of Fusarium head blight (FHB), a devastating fungal disease affecting cereal crops worldwide. FHB causes yield losses and grain contamination with mycotoxins (type A and B trichothecenes, zearalenone, enniatins, beauvericin), posing significant health and food safety concerns. Several Fusarium species occupying the same ecological niche during infection likely interact, modulating FHB outcomes, including symptoms and mycotoxin contamination. However, most of the studies so far focused on a single Fusarium species, particularly F. graminearum, which has proven to be insufficient to develop agronomic practices allowing for ensuring the level of mycotoxins in compliance with the EU regulation (2023/915), and also, for the comprehensive understanding of the FHB disease process, including mycotoxin accumulation in wheat crops. Through this postulate, the works realized in this thesis tried to go beyond a reductionist approach and the paradigm “one species = one disease” for studying FHB-associated species. By hypothesizing that the observed mycotoxin contaminations might result from the global metabolism of a blend of Fusarium species in interaction, we aimed to develop a new approach via a first proof of concept of a Fusarium Synthetic community (SynCom), named Meta-Fusarium, composed of the most common Fusarium species in Europe (F. graminearum, F. culmorum, F. poae, F. avenaceum, F. sporotrichioides, F. langsethiae, F. tricinctum), with one strain per species. The Meta-Fusarium SynCom was analyzed using a combination of quantitative PCR (qPCR) to track strain compositional dynamics and targeted/untargeted metabolomics to characterize metabolic outputs, including mycotoxins. Results obtained during this thesis highlight that the sum of individual isolates does not predict the collective behavior of the FHB disease complex. It revealed strong competitive interactions within the Meta-Fusarium SynCom, where F. culmorum predominates, followed by F. graminearum, and to a lesser extent F. poae. Furthermore, the metabolomic outcomes demonstrated that strain-specific metabolites, including mycotoxins, were differentially modulated by interspecies interactions. Notably, metabolite production did not simply reflect additive single-strain behavior, with certain metabolomic features suppressed or enhanced, or transformed in response to the presence of other Fusarium strains. External factors, including temperature, oxidative stress, and antifungal molecule (ferulic acid) exposure, further reshaped Fusarium interactions, modulating the composition between F. culmorum and F. graminearum, where a shift was observed and consequently the mycotoxin outcomes. Lastly, differential early developmental stage (germination rate) prior to inoculation critically influences competitive dynamics and metabolic outcomes. Overall, this thesis establishes the first proof of concept for using synthetic Fusarium communities to dissect the ecological complexity of FHB. The Meta-Fusarium SynCom provides a modular and reproducible framework to study interspecies interactions, bridging the gap between reductionist single-species approaches and the complex realities of FHB in the field. In perspectives, it aimed to provide new insights into FHB disease processes in order to predict future outbreaks and to develop new integrated practices to fight the mycotoxin contaminations.

• Thalamocortical mechanisms of adaptive learning: a specific role for the mediodorsal nucleus in secondary motor cortex computations

  by Margaux GIRAUDET (Institut Interdisciplinaire de Neurosciences)

  The defense will take place at 14h00 - Amphi BROCA 146 rue Léo Saignat 33076 Bordeaux

  in front of the jury composed of

  • Francisco CLASCA - Full professor - Autonoma University Medical School Madrid - Rapporteur
  • Audrey HAY - Chargée de recherche - Université Claude Bernard - Rapporteur
  • Shauna PARKES - Directrice de recherche - Université de Bordeaux - Examinateur
  • Fanny CAZETTES - Chargée de recherche - Université d'Aix Marseille - Examinateur

  Summary

  Adaptation to volatile environments is a fundamental, everyday challenge that relies predominantly on thalamocortical pathways. In particular, mediodorsal thalamus (MD) and ventroanterior-ventrolateral (VAVL)–to–secondary motor cortex (M2) projections are well positioned to shape action selection, yet their computations during adaptive choice remain unclear. During my thesis, I combined a self-initiated two-lever task in head-fixed mice, focal MD and VAVL lesions, longitudinal two-photon imaging of MD and VAVL to M2 axons, and a reinforcement-learning race model to determine how thalamocortical signals govern preparation, vigor, and action selection under uncertainty. Mice chose left (L1), right (L2), or near-simultaneous (L1&L2) presses; outcomes followed probabilistic (80/20) contingencies and double presses never yielded reward. A Rescorla–Wagner rule updated action values (Q1, Q2), which drove independent accumulators with a finite co-selection window (δt) that defines double-press events. The model captured choice fractions, reward rate, and reaction times (RTs), isolating two latent variables: ΔQ (value difference) that biases which action wins, and ΣQ (value sum) that controls initiation vigor. Targeted MD lesions selectively impaired de novo rule learning without disrupting reversal or gross motor output. Learning rates and the value-to-rate mapping were preserved, but δt widened specifically after MD lesion, producing more L1&L2 presses. Because double presses never yield reward, a wider δt mechanically reduced performance, compressed ΔQ, and diminished rate separation, thereby promoting ambiguous co-selection despite intact value updating. In contrast, VAVL lesions spared learning, indicating circuit specificity. Indeed, if both thalamic nuclei anatomically project to M2, our connectomic experiments demonstrated topographic segregation: MD preferentially targets medial M2, whereas VA/VL innervates M2 more broadly, supporting thalamocortical functional heterogeneity. To link these computations to thalamocortical signaling, we imaged MD and VAVL axons in M2 across learning in behaving mice. Clustering of MD bouton activities revealed a specific ensemble in M2 that remain stable over days and showed robust pre-press modulation. Critically, this pre-press MD-M2 signal was not a generic correlate of performance (it did not scale with success rate) but was tightly coupled to RT and to double-press control. Trials binned by RT showed that stronger pre-press activity preceded faster responses, explaining substantial RT variance. Its relationship to selection was context-dependent: under long pre-press delays, greater pre-press activity predicted fewer double presses (an anti-double-press bias), whereas under short delays its predictive power approached chance. This framework unifies behavior, computation, and circuit dynamics. The race model formalizes how ΔQ and ΣQ jointly shape selection and speed; δt widening after MD lesion pinpoints a temporal-coincidence gating failure that degrades performance by promoting co-activation; and imaging identifies a pre-press thalamocortical state that links preparation to both RT and double-press control. This mechanistic chain parsimoniously explains why MD lesions alter learning despite preserved motor abilities and value updating. Taken together, these findings indicate that MD-M2 circuit specifically conveys a control signal before the press that both accelerates RT and, under long delays, biases choices away from double-pressing toward the target lever. MD lesions selectively increase double-press, as expected from an altered pre-press signal that weakens the anti-co-selection bias at long delays. Overall, our data suggest that MD-M2 circuit controls cortical preparatory states and enforces temporally precise, single-action exclusive selection under uncertainty.

• Towards the genetic architecture of tolerance to Tobacco Mosaic Virus in Arabidopsis thaliana

  by Julien HENNEQUART (BFP - Biologie du Fruit et Pathologie)

  The defense will take place at 14h00 - Amphithéâtre Colette & Josy Bové (Bâtiment B2) Amphithéâtre Colette & Josy Bové, Bâtiment B2, 71 Av. Edouard Bourlaux, 33140 Villenave-d'Ornon

  in front of the jury composed of

  • Valérie SCHURDI-LEVRAUD - Professeure des universités - Université de Bordeaux - Directeur de these
  • Laurence ALBAR - Chargée de recherche - PHIM Plant Health Institute of Montpellier, University Montpellier, IRD, CIRAD, INRAE, Institut Agro, Montpellier, France - Rapporteur
  • Israel PAGAN - Professeur des universités - Universidad Politécnica de Madrid, Centro de Biotecnología y Genómica de Plantas UPM-INIA/CSIC - Rapporteur
  • Jérôme JOUBES - Professeur des universités - Université de Bordeaux, Laboratoire de biogénèse membranaire, UMR 5200 - Examinateur
  • Mélanie JUBAULT - Maîtresse de conférences - IGEPP Institut Agro, INRAE, Université de Rennes - Examinateur

  Summary

  Resistance to viruses is often effective only in the short term as the high rate of evolution of viruses allows them to bypass resistance mechanisms. Conversely, tolerance (i.e., the ability of plants to accumulate pathogens in their tissues without causing disease symptoms or loss of fitness) may be more durable as the viruses are subjected to less selective pressure. Tobacco mosaic virus (TMV), an RNA virus from the genus Tobamovirus, is able to infect a wide range of hosts, including Arabidopsis thaliana. Additionally, various Arabidopsis ecotypes exhibit diverse responses to TMV. For instance, Shahdara (Sha) is susceptible, exhibiting a high viral accumulation early on, with a distinct leaf curling symptom; Tsu-0 is resistant; and Col-0 is tolerant, with substantial viral accumulation without exhibiting the curly symptom. In this thesis project, our first goal was to characterize the tolerance phenotype across the Sha, Col-0, and Tsu-0 ecotypes. We measured symptoms, viral load and fitness traits (including seed set, biomass, size), and conducted transcriptomics and untargeted metabolomics analyses. Sha, Col-0 and Tsu-0 ecotypes had different metabolomic and transcriptomic responses to the virus, highlighting the specificities of the susceptible, tolerant and resistant responses. These specificities allowed us to identify tolerance markers, including a quantitative measure of tolerance called “range tolerance”. We used these markers to perform a genome-wide association study (GWAS) on 135 Arabidopsis thaliana ecotypes from a worldwide collection. These ecotypes displayed a wide range of tolerance levels, while only Shahdara had visible symptoms with the highest viral accumulation. The GWAS analysis allowed the identification of candidate genes for TMV resistance and a distinct set of genes for tolerance. These findings suggest that tolerance and resistance strategies rely on different mechanisms. Further experiments will be conducted to confirm the involvement of these candidate genes in tolerance.

ED Sciences Physiques et de l'Ingénieur

• Integration of biosignals in cyber-physical systems to assess the cognitive state of a user: contribution of neurovisceral and cneurovisceral dynamics

  by Alix BOUNI (Laboratoire de l'Intégration du Matériau au Système)

  The defense will take place at 14h00 - Petit Amphithéatre (O 108) Ecole Nationale Supérieure de Cognitique (ENSC), 109 Av. Roul, 33400 Talence

  in front of the jury composed of

  • Laurent ARSAC - Professeur des universités - Université de Bordeaux - Directeur de these
  • Michel AUDIFFREN - Professeur des universités - Université de Poitiers UFR Sciences du Sport - Rapporteur
  • Fabien SAUVET - Professeur agrégé - Institut de recherche biomédicale des armées (IRBA) - Rapporteur
  • Véronique DESCHODT-ARSAC - Professeure des universités - Université de Bordeaux - CoDirecteur de these
  • Pierre BOUNY - Maître de conférences - Université de Bordeaux - Examinateur
  • Gwenaelle CEYTE - Maîtresse de conférences - UFR STAPS, Université Paris Cité - Examinateur

  Summary

  The assessment of individual cognitive state, and its long-term evaluation, particularly in demanding professional situations requiring adaptive responses, is an increasingly important societal challenge. In this context, non-invasive physiological measurements offer a promising way to detect changes in cognitive state in real time, without disrupting the user's activity. This thesis takes a systemic approach to the human organism, exploring the possibility of inferring cognitive state from the combined analysis of motor and cardiac dynamics. While each of these signals have already prove value for detecting specific cognitive states, their simultaneous acquisition requires the development of reliable, synchronized, and adaptable measurement systems suitable for both experimental and real-world conditions. The work presented here aims to design a cyberphysics system (CPS) for multimodal data acquisition and to evaluate the relevance of various indicators derived from neurovisceral and cognitive-motor dynamics for detecting changes in the cognitive state. To achieve this objective, three complementary experimental protocols were implemented alongside the development of the CPS. These protocols examined how different indicators derived from movement and cardiac dynamics evolve under various conditions: (1) during cognitive-motor tasks of increasing complexity; (2) following an intervention designed to improve psychophysiological state; and (3) in settings close-to real work environments. The experimental results highlight the sensitivity of movement multifractality and non-linearity indicators to differences in experimental constraints. Cardiac indicators, particularly those from frequency-domain analyses, show pronounced sensitivity to variations in cognitive state. Moreover, the combined analysis of cardiac and motor dynamics offers a promising avenue for finely characterizing cognitive states according to individual response profiles. The thesis also explored a method for modulating cognitive state: cardiac coherence. A single five-minute session led to immediate and 24-hour improvements in participants' psychophysiological state, suggesting its potential utility in professional contexts. Finally, the last study demonstrated the integration of real-time cardiac analyses within the CPS, enabling the delivery of physiological indicators sensitive to cognitive state during task performance, which could enhance user comfort, safety, and performance. Overall, the work was based on a multidisciplinary approach, combining physiological signal acquisition and analysis with the development of an embedded electronic system for real-time processing. It offers practical routes towards the detection of cognitive state variations in real operational contexts.

• Quantum theory of photon emission in current-driven single-molecule tunnel junctions

  by Andrés Felipe BEJARANO SANCHEZ (Laboratoire Ondes et Matière d'Aquitaine)

  The defense will take place at 14h00 - NS Universidad del País Vasco (UPV/EHU) Donostia–San Sebastián, 20018, España

  in front of the jury composed of

  • Brahim LOUNIS - Professeur des universités - University de Bordeaux - Examinateur
  • Katharina KAISER - Junior professor - University of Göttingen - Rapporteur
  • Tomáš NEUMAN - Professeur - Institute of Physics of the Czech Academy of Sciences (FZU-CAS) - Rapporteur
  • Diego MARTIN-CANO - Chargé de recherche - Universidad Autónoma de Madrid - Examinateur
  • Nerea ZABALA - Professeure des universités - Universidad del País Vasco - Euskal Herriko Unibertsitatea - Examinateur

  Summary

  This thesis investigates light emission processes in molecular tunnel junctions, where a single molecule is placed between metallic electrodes and interacts with a confined plasmonic mode. In such systems, the injection of an electric current can induce photon emission, offering a unique platform to explore light–matter interactions at the nanoscale without external illumination. To describe these phenomena, we develop a theoretical framework based on quantum master equations. These models capture both charge transport and radiative dissipation, allowing us to calculate experimentally relevant observables such as emission spectra and photon correlation functions. We explore two complementary models. In the first, the molecule is treated as a two-level system with optically active transitions. Photon emission occurs through dipolar coupling to a plasmonic mode localized in the junction. Depending on system parameters, the interaction can be tuned from weak to strong coupling. To characterize this transition, we introduce the concept of cooperativity, borrowed from quantum optics, which quantifies the competition between coherent coupling and dissipative processes. We identify distinct spectral features for each regime. In particular, strong coupling gives rise to Rabi splitting and coherent oscillations in the photon statistics, which are signatures of quantum light–matter interaction. In the second model, the molecule is represented by a single electronic level where Coulomb repulsion prevents double occupancy. Here, photons are emitted not from internal transitions but from charge fluctuations driven by the current. This mechanism leads to remarkably rich photon statistics. By tuning the applied bias voltage, the system can act as a single-photon emitter (showing antibunching) or as a source of photon pairs (superbunching). This voltage-dependent behavior highlights the potential of such molecular devices for electrically controlled quantum light sources. Analytical results are complemented with numerical simulations, developed using a dedicated Python package. The code enables the efficient calculation of steady-state observables and correlation functions, providing insights into regimes where analytical solutions are not available. Altogether, this work demonstrates that simple molecular junctions can serve as versatile platforms for generating and studying non-classical light at the nanoscale. Our results contribute to the fundamental understanding of current-driven quantum emitters and pave the way for future applications in quantum communication, sensing, and nanoscale optoelectronics.

• Gray Molasses Cooling for Exploring Emergent Behaviors in Traveling-Wave Optical Cavities

  by Geovan SANTANA DE FIGUEIREDO (Laboratoire Photonique, Numérique & Nanosciences)

  The defense will take place at 14h00 - Amphitéatre Institut d'optique d'Aquitaine, Rue François Mitterrand, 33400 Talence

  in front of the jury composed of

  • Andrea BERTOLDI - Ingénieur de recherche - Laboratoire Photonique, Numérique et Nanosciences (LP2N), CNRS UMR 5298 - Directeur de these
  • Emmanuel ABRAHAM - Professeur des universités - Laboratoire Ondes et Matière d'Aquitaine (LOMA), CNRS UMR 5798 (Photonics and Materials Group) - Examinateur
  • Francesco MINARDI - Professeur associé - Università di Bologna - Rapporteur
  • Sylvain SWARTZ - Directeur de recherche - ONERA, Laboratoire Qtech. - Rapporteur

  Summary

  This thesis explores the cooling and organization of ultracold ({}^{87}mathrm{Rb}) atoms interacting with traveling-wave optical cavities. The experimental setup has been entirely rebuilt and optimized, including the implementation of a commercial fiber-based laser system frequency-doubled to (780,mathrm{nm}) and a fully reconfigured control architecture. These improvements enabled the realization of stable 2D and 3D magneto-optical traps, as well as the reliable loading of atoms into an optical dipole trap. We investigate sub-Doppler cooling based on the grey-molasses scheme, which relies on hyperfine dark states. The performance of the grey molasses in the presence of the dipole trap has been characterized in terms of temperature and capture efficiency. These results highlight the advantages of this cooling mechanism and pave the way toward the realization of a Bose–Einstein condensate without evaporative cooling. In parallel, a triangular monolithic ring cavity has been designed. This compact and rigid geometry provides a traveling-wave configuration suitable for studying light–matter crystallization in a regime where continuous translational symmetry can be broken. These developments aim to enable the exploration of light-induced collective phases such as self-organization and supersolidity, emerging from long-range interactions between the atoms and the intracavity field.

• Correlation between dielectric constant and thermal conductivity for building envelopes: experimental validation and in-situ assessment by GPR and GHP

  by Jinxia HU (I2M - Institut de Mécanique et d'Ingénierie de Bordeaux)

  The defense will take place at 14h00 - Amphi F Bâtiment A29, 351 cours de la Libération CS 10004，33400 Talence

  in front of the jury composed of

  • Mehdi SBARTAI - Professeur des universités - Université de Bordeaux - Directeur de these
  • Didier DEFER - Professeur des universités - Université de l'Artois - Rapporteur
  • Geraldine VILLAIN - Directrice de recherche - Université Gustave Eiffel - Rapporteur
  • Balayssac JEAN-PAUL - Professeur des universités - Université de Toulouse - Examinateur

  Summary

  Accurate assessment of thermal conductivity is crucial for evaluating the energy performance of building envelopes. However, in-situ measurement of thermal conductivity remains a significant challenge due to the complexity of wall structures, environmental variability, and the limitations of existing testing methods, which are often time-consuming, invasive, or unsuitable for practical conditions. Ground Penetrating Radar (GPR), as a non-destructive technique, offers potential for indirect thermal property estimation through its sensitivity to dielectric properties. This research explores the feasibility of using GPR to estimate thermal conductivity on-site, aiming to provide a practical solution for thermal performance diagnostics and quality control in building applications. The initial theoretical analysis and literature review revealed that both thermal conductivity and permittivity are positively correlated with material density, and are significantly influenced by fiber orientation and moisture content—especially in anisotropic materials like wood. Experimental results from wood species and insulation materials indicated a strong linear relationship between dielectric constant and thermal conductivity, with consistent anisotropic behavior present in both properties. These findings were further validated across a broader range of materials—including foam, cork, OSB, plaster, and concrete—demonstrating that the permittivity-thermal conductivity relationship remains valid across a wide density spectrum and in both single- and multilayer configurations. Importantly, the single layer model, multilayer and unified model were applied to real wall systems with and without internal defects. The results demonstrated that multilayer model offers the most suitable model for in-situ measurements. Besides, the experimentally measured dielectric constants in defective regions were slightly lower than theoretical predictions, while the corresponding thermal conductivities were significantly higher due to reduced thermal resistance. These deviations underscore the importance of accounting for defects in predictive models and further reinforce the diagnostic value of GPR in detecting hidden anomalies. Therefore, this study demonstrates that GPR is not only effective for detecting internal defects that compromise the performance of building envelopes, but also offers a novel method for indirect evaluation of thermal properties. By leveraging the strong correlation between dielectric properties and thermal conductivity, the proposed method allows for rapid, non-destructive prediction of heat transfer characteristics in building walls. Moreover, it would eliminate the reliance on ambient conditions required for steady-state predictions, and enable in-situ and real-time assessment, making it highly suitable for large-scale field applications. These findings suggest that GPR-based dielectric analysis has strong potential to serve as a practical tool for both thermal performance evaluation and structural health monitoring in real-world building applications.

ED Sociétés, Politique, Santé Publique

• Therapeutic research on COVID-19 in Sub-Saharan Africa: lessons learned from the implementation of an adaptive clinical trial; analysis of short-, medium-, and long-term morbidity in an at-risk population

  by Mwinmalo DA (Bordeaux Population Health Research Center)

  The defense will take place at 8h30 - Module 1.3 146 rue Léo Saignat, 33000 Bordeaux, France

  in front of the jury composed of

  • Alexandre DUVIGNAUD - Maître de conférences - praticien hospitalier - Université de Bordeaux - Directeur de these
  • Armel PODA - Full professor - Université Nazi Boni - CoDirecteur de these
  • Pierre TATTEVIN - Full professor - Université de Rennes - Examinateur
  • Rasmata OUEDRAOGO - Full professor - Université Joseph Ki Zerbo - Rapporteur
  • Hervé HIEN - Maître de recherche - Institut de Recherche en Science de la santé - Rapporteur
  • Valériane LEROY - Directrice de recherche - Institut national de la santé et de la recherche médicale - Examinateur
  • Didier EKOUEVI - Full professor - Université de Lomé - Examinateur

  Summary

  Title: Therapeutic Research on COVID-19 in Sub-Saharan Africa: Lessons from the Implementation of an Adaptive Clinical Trial; Analysis of Short-, Medium- and Long-Term Morbidity in a High-Risk Population Background - The COVID-19 pandemic exposed the structural vulnerabilities of health and research systems in sub-Saharan Africa (SSA), while simultaneously revealing their potential for innovation and adaptability. This doctoral work draws upon the experience of several therapeutic trials conducted mainly in SSA—COVERAGE Africa (CVGA), ANTICOV, and INTENS-CoV (ICOV)—to: (i) identify the enablers and barriers to implementing an adaptive clinical trial in an emergency context; (ii) determine the clinical severity factors among at-risk patients; and (iii) document the phenotypic expression, evolution, and determinants of post-infectious symptoms. Methods - The PEARLES conceptual framework was applied to identify operational enablers and barriers in the implementation of the ambulatory CVGA trial in Burkina Faso. Data from the multicountry platform trial ANTICOV, conducted in twelve African countries and Brazil, were analyzed to identify factors associated with progression to severe COVID-19 (defined as SpO₂ ≤ 93% or death by Day 21) among high-risk participants enrolled early with mild to moderate disease. Data from ICOV, CVGA, and the follow-up cohort EVOLCOV in Côte d'Ivoire and Burkina Faso will enable the analysis of phenotypic expression, trajectory, and determinants of clinical manifestations persisting beyond the acute phase. Results - The CVGA experience highlighted the facilitating role of agile institutional coordination, strong political commitment, and flexible financing mechanisms in the successful implementation of adaptive trials, as well as the barriers posed by complex administrative and regulatory processes. Our findings underscore the need for accelerated regulatory, administrative, and financial procedures, along with pre-approved adaptive protocols. In the ANTICOV platform trial, progression to severe COVID-19 was rare (1.1%), occurring mainly among participants aged >40 years or with comorbidities such as obesity (aOR 3.5, 95% CI 3.2–7.6), hypertension (aOR 3.7, 95% CI 2.6–11.4), and diabetes (aOR 4.5, 95% CI 2.8–15.7). Elevated CRP (≥20 mg/L), neutrophil-to-lymphocyte ratio (≥2), and hyperglycemia (≥6 mmol/L) were significantly associated with clinical deterioration. Analysis of long-term post-infectious manifestations is ongoing Conclusion - This research highlights the need for structural and organizational adjustments to enhance the efficiency of research during public health crises in SSA. It also demonstrates the prognostic value of simple indicators—age, metabolic comorbidities, and accessible biomarkers—that could be integrated into practical decision-support tools for managing emerging respiratory viral infections. Altogether, these findings provide an entry point for strengthening clinical research and health systems preparedness for future pandemics, through an agile, context-driven, and sovereign approach grounded in locally generated evidence. Keywords: COVID-19; Sub-Saharan Africa; adaptive trial; biomarkers; long COVID; pandemic preparedness

• PARTICIPATION OF THE POPULATION TO IMPROVE THE CONDITIONS FOR EFFECTIVENESS OF DIGITAL INTERVENTIONS IN PREVENTION AND HEALTH PROMOTION IN PRIMARY HEALTH CARE.

  by Melek AKTAS (Bordeaux Population Health Research Center)

  The defense will take place at 14h00 - Amphi Louis ISPED, Campus Carreire, 146 Rue Léo Saignat, 33000 Bordeaux

  in front of the jury composed of

  • Christine COHIDON - Maîtresse d'enseignement et de recherche - Unisanté Lausanne - Examinateur
  • Véronique REGNIER - Docteure - Université Jean Monnet Saint-Étienne - Rapporteur
  • Emmanuel RUSCH - Professeur des universités - praticien hospitalier - Université de Tours - Rapporteur
  • François ALLA - Professeur des universités - praticien hospitalier - Université de Bordeaux - Examinateur

  Summary

  With the digitization of healthcare, digital interventions in prevention and health promotion are playing an increasingly important role in public policy and practice. Presented as innovative solutions through a preventive approach, these interventions occupy a central place in primary care. However, their effectiveness is limited. The literature identifies user participation as a condition for the effectiveness of these interventions. Yet, participation is not defined in a consensual and its contours remain vague, limiting the ability of designers and researchers to develop participatory interventions. Often reduced to quantitative indicators such as the number of connections or frequency of use, this vision obscures the complexity of interactions between the individual, the tool, and their environment. This research aims to clarify the concept of user participation in digital interventions for health prevention and promotion, identify the attributes and conditions for effectiveness, and propose an operational framework to improve the design, implementation, and evaluation of participation. A three-part sequential approach was used: (1) a literature review to identify and conceptualize the attributes of participation; (2) a case study conducted via theory-based evaluation of the MedPrev program, a hybrid primary care program combining digital tools and human support, to analyze the conditions under which participation influences effectiveness; (3) an e-Delphi process involving a panel of experts composed of designers, professionals, and users to validate the identified attributes and develop a participation grid. First, the literature review identified 26 elements organized into five categories: attributes, levels of participation, facilitators, difficulties, and methods. Then, the case study revealed 16 general conditions for participation in MedPrev, structured around three levels of influence: individual and social, interventional, and environmental. An in-depth analysis of the interventional component highlighted the individual-centered nature of participation, emphasizing engagement based on personal skills and capacities. Furthermore, the examination of differentiated role of the two intervention modalities showed that human support proved to be central to the activation of behavior change techniques, particularly those related to self-regulation, motivation, and self-efficacy reinforcement. On the other hand, the digital component, which is still limited in its personalization and use for monitoring, had a limited impact on user engagement. Finally, the e-Delphi method was used to validate by consensus a grid consisting of 30 elements promoting participation and covering the design, implementation, and evaluation of interventions. The results emphasize that participation is a multidimensional, dynamic, and contextual concept influenced by individual, interventional, and environmental factors. This research highlights the importance of evidence-based design and thorough theory-based evaluation in order to understand the complexity of participation. The grid developed from this research provides a structured framework for designing, implementing, and evaluating participation. Future prospects include creating and validating a guide in different contexts and assessing its impact. This thesis contributes to a renewed understanding of participation, not as an end in itself, but as a condition for effectiveness and equity, requiring a holistic perspective on digital interventions in prevention and health promotion.