HAL-Paris 13
Not a member yet
33893 research outputs found
Sort by
Évaluation de l'impact des algorithmes d'imputation multiple sur les performances des modèles pharmacocinétiques : une étude basée sur la simulation
No full textInternational audienceThis study compared multiple imputation (MI) algorithms in a one-compartment pharmacokinetic (PK) scenario with oral absorption. Four covariates (two continuous, two dichotomous) linked to PK parameters were randomly removed under a missing completely at random (MCAR) mechanism. The aim was to identify which algorithm best preserves covariate distributions and PK parameter estimates. The original dataset included 100 individuals, each with five sampling occasions. Missing data were introduced at 5%, 20%, 50%, and 75% for the four covariates under the MCAR assumption. Five MI algorithms (Mice, Amelia, missForest, rMIDAS, XGBoost) were tested. Absolute and relative errors and concordance metrics were used to assess performance. Population and individual parameter estimates were compared across imputed and original datasets using Monolix2024R1®. MissForest (MF) and Amelia yielded lower errors for continuous covariates whereas dichotomous variables were poorly imputed. Based on objective function values, Mice perform best at 5% and MF at 20% of missingness. Increasing missingness decreased covariate effects and increased the estimated inter-individual variances. Individual parameter estimation accurately captured individual-level variability across all imputed datasets. MI methods appear effective for covariate imputation in PK modeling, offering reliable results up to 20% missingness under an MCAR mechanism. Future research should explore refined strategies, including advanced modeling frameworks and Bayesian approaches for imputation. Enhancing our understanding of missing data processes will be crucial for robust PK analyses across diverse clinical settings
TactiDigit-ART project: accessible art-historical approach to reading and constructing tactile images: How does this device facilitate accessibility and empowerment for visually impaired adults?
No full textInternational audienceAccess to works of art for the visually impaired is still limited, and most often involves oral or written description. Some words remain meaningless for lack of sensitive experience (cf. works by Bris, 2008, Bernard, 2018, 2022, Gentaz et al., 2020, 2022, Hatwell, 2000; Kastrup and Sampaio, 2012; Lewi-Dumont, 1997, 2020). Both personal and associations initiatives have been set up to develop accessibility devices (Tactidigi-Art Project). Our study, included in the research component of this project, aims to examine its impact on access to complex images by visually impaired people. While studies have already examined the specific characteristics of textbook access modalities (Castillan, 2020) or illustrations in children's books for children (Valente, Mascle, Chennaz et al., 2024), little is known in the cultural and artistic field about adults. Our exploratory research is part of this framework, with the theme of the Paris 2024 Olympic and Paralympic Games. Our hypothesis is that an image-reading device, linking iconic, tactile and linguistic aspects, promotes accessibility to works of art. Our aim is to test its accessibility. Using Mayer's multimedia learning theory (2014 ; Mayer & Fiorella, 2014) adapted to the tactile modalities of images, we studied the effects of several workshop sequences conducted with a dozen participants, aged between 25 and 60. The technical and artistic innovation resource is Olymp' Touch, sport told through touch, a book withs embossed prints. We will present our observation methodology based on focus groups before and after image-reading activities, as well as semi-directive interviews with participants, facilitators and image designers. The results will be presented qualitatively in the form of verbatim from interactions between participants, facilitators and/or researchers, opening up new perspectives on the accessibility of complex images
Cross-Domain Approach for Automated Thyroid Classification Using Diff-Quick Images
No full textInternational audienceClassification of thyroid images based on the Bethesda category using Diff-Quick stained images can assist in diagnosing thyroid cancer. This paper proposes a cross-domain approach that modifies the original deep learning network designed to classify X-ray images to classify stained thyroid images. Since the Diff-Quick stained images have large and high-quality sizes with tiny cells with essential characteristics that can help a doctor diagnose, resizing images is required to maintain this characteristic, which is significant. Thus, in this paper, we also research and evaluate the performance of different interpolation methods, including linear and cubic interpolation. The experiment results evaluated on a private dataset present promising results in the thyroid image classification of the proposed approach
Dual Weighted Residual-driven adaptive mesh refinement to enhance biomechanical simulations
No full textInternational audienceThis chapter describes how a posteriori error estimates targeting a user-defined quantity of interest, using the Dual Weighted Residual (DWR) technique, can be easily applied for biomechanical simulations in current engineering practice. The proposed method considers a very general setting that encompasses complex geometries, model non-linearities (hyperelasticity, fluid-structure interaction) and multi-goal oriented techniques. The developments are substantiated with some numerical tests
Magnéto-élasticité géante dans FeRh, mise en évidence par le couplage fort phonon-magnon
No full textWe report a strong coupling between a Love surface acoustic wave (SAW) and a spin wave in epitaxial FeRh as analyzed from their anti-crossing dispersion relationships obtained by Brillouin light scattering. Reproducing the results using a numerical model that solves the coupled Landau-Lifshitz and Christoffel equations reveals that this effect stems from both the large magneto-elastic coupling coefficient in this material, |B 2 |=17 MJ/m 3 , and the strength of the Love-wave magneto-acoustic effective field when the magnetization vector M is parallel to the SAW wavevector k. We also observe a substantial, albeit weaker, coupling in the unusual configuration of the Sezawa 1 surface wave crossing the pseudo Damon-Eshbach mode, i.e. when the angle between M and k is π/4. These findings identify the optimal acoustic modes and magnetic geometries for achieving strong phonon-magnon hybridization, paving the way for coherent information transfer via elastic waves in magneto-acoustic nanodevices.</div
Spectroscopic probing of Rydberg atoms close to dielectric surfaces: Parasitic electric fields and Casimir-Polder interactions
No full textInternational audienceHighly excited (Rydberg) atoms have exaggerated properties making them extremely sensitive to external electromagnetic fields and interacting strongly with their environment. Atomic vapor cells represent an attractive platform for studying Rydberg atoms and fabricating quantum devices. For example, Rydberg atoms in vapor cells have been used as sensitive detectors of electric fields of frequencies ranging from DC up to the THz range but also as single photon sources for quantum technology applications exploiting collective phenomena due to the Rydberg blockade effect [1]. Rydberg atoms also find applications in fundamental physics, in particular for the measurement of dispersive interactions of the Casimir-Polder type (atom-surface interactions) [2]. One major advantage of Rydberg atoms is that they expose limitations in the traditional perturbative approach of Casimir-Polder (CP) theory [3, 4]. Indeed, in the extreme near-field, the dipole approximation breaks down and higher-order terms need to be considered. We report on extensive experimental measurements of the Rydberg-surface interaction using spectroscopy in cesium vapor nanocells of a thickness ranging roughly from 200-700nm, as well as selective reflection spectroscopy on a macroscopic cesium all-sapphire cell. Atoms are first excited to the Cs(6P) level with a 894nm pump laser and subsequently a green laser ≈ 510nm probes Rydberg nD or nS states, where the principal quantum number n ranges between 15-17. Our experiments evidence the dipole-dipole term of the Casimir-Polder interaction providing a measurement of the C3 coefficient for cesium Rydberg states. Furthermore, our experiment clearly evidences an additional interaction that induces shifts and broadens the linewidth of the probed transitions in the vicinity of the dielectric windows of our cells. We believe that this interaction is due to electric fields that are either generated by patch charges (trapped on the surface or induced by the excitation lasers), or by cesium adsorbants. We show that the different polarizability (of opposing sign) between S and D Rydberg states can be exploited to extract quantitative measurements of the strength and distance scaling (z-dependence) of such parasitic electrostatic interactions.Our experiments suggest that the sensitivity of Rydberg atoms to external electric fields could provide a unique tool for probing electrostatic interactions in the vicinity of surfaces. This could allow systematic error corrections in Casimir-Polder experiments with excited or even ground state atoms that aim at putting bounds on the existence of non-Newtonian gravity [5]. We are currently exploring the possibility of coating the internal window interfaces with conducting 2D material such as graphene to reduce the influence of parasitic charges. This could allow us to probe atoms closer to the surface, where quadrupole interactions could be experimentally attainable for the first time [4].References[1] H. Kubler, J. P. Shaffer, T. Baluktsian, R. Loew, T. Pfau, Nat. Photon., 4, 112–116 (2013).[2] V. Sandoghdar et al., Phys. Rev. Lett 68, 3432–3435 (1993).[3] J. A. Crosse et al., Phys. Rev. A 82, 3010901 (2010).[4] B. Dutta et al., Phys. Rev. Res. 6, L022035 (2024).[5] A. Laliotis, B-S. Lu, M. Ducloy, D. Wilkowski, AVS Quantum Sci. 3, 043501 (2021)
Hypertensive acute heart failure: a critical perspective on definition, epidemiology, pathophysiology, and prognosis—a narrative review: a joint session with the Romanian Society of Cardiology (part II)
No full textInternational audienceHypertensive acute heart failure (HT-AHF) has historically been recognized as a distinct clinical phenotype of AHF, characterized by acute pulmonary congestion in the context of elevated systolic blood pressure (SBP), typically > 140 mmHg. However, emerging evidence has begun to challenge the diagnostic accuracy, clinical utility, and relevance of this category. A main criticism of HT-AHF is its considerable overlap with other AHF clinical profiles, including acute decompensated heart failure (ADHF) and acute pulmonary oedema (APO). Clinical features such as dyspnea and pulmonary congestion are not unique to HT-AHF. Additionally, some HT-AHF patients concurrently fulfill diagnostic criteria for the ADHF phenotype, including a history of HF or signs of volume overload, leading to ambiguity in diagnosis. HT-AHF is associated with very low in-hospital mortality (0–2%) compared to other AHF phenotypes. Notably, there is no robust evidence linking high SBP to poor short- or long-term outcomes, nor are there randomized clinical trials validating distinct management strategies for HT-AHF. Often associated with the management of HT-AHF, vasodilators have shown limited benefit across trials, contributing to a downgrade in guideline recommendations. The relatively favorable short-term prognosis and the lack of a standardized, evidence-based treatment approach weaken the rationale for classifying HT-AHF as a standalone AHF category. Given the heterogeneity of clinical presentations, overlap with other AHF phenotypes, and lack of prognostic distinction or targeted therapy, the term “AHF with high SBP at presentation” offers a more flexible and clinically meaningful descriptor, encouraging a more nuanced approach to treatment
Réponse Biomécanique de la Peau Humaine : Une Approche Multi-Compartiment des Milieux Poreux.
No full textLiving tissues, such as skin, are particularly sensitive to mechanical and physiological alterations, which can cause lesions, inflammatory redness, or open wounds. These alterations, collectively referred to as tissue damage, encompass any degradation of the structure or function of the tissue. Damage may arise under normal physiological conditions or result from external mechanical stress, such as moderate but prolonged pressure or intense sudden loading. Such situations are common among hospitalised patients or individuals with prolonged immobility - bedridden or wheelchair-bound - and can lead to serious complications such as pressure ulcers. With an ageing population and an increase in chronic diseases and sedentary lifestyles, these injuries represent a major public health challenge. Prevention and management are based on a better understanding of the underlying mechanisms to anticipate their occurrence and develop personalised care strategies.Tissue lesions are the result of a complex interplay of mechanical, biological, and individual factors. Animal studies have demonstrated a link between mechanical loading, oxygen deprivation (ischaemia), and cell death, particularly when mechanical deformation exceeds a critical threshold, whether through compression or shear. Furthermore, mechanical and ischaemic effects have been shown to act synergistically in initiating and propagating tissue damage. However, current mathematical models often treat these aspects separately, failing to explicitly account for the dynamic interactions between mechanical stress, tissue deformation, micro-circulation, and biological responses. This compartmentalisation limits the predictive power and applicability of these models. In addition, most calibration data come from animal studies, which hampers their direct applicability to human skin and their adaptation to patient-specific variability.The central hypothesis of this work is that a poromechanical, multiphasic and multiscale modelling framework, one that integrates the structural, fluidic, and biological components of tissue, can better characterise the critical conditions leading to skin damage. This approach also provides a promising tool for identifying susceptibility biomarkers, thus contributing to the development of personalised preventive care.The main objective of this thesis is therefore to develop a hierarchical poromechanical model of human skin under physiological conditions, coupling mechanical, micro-circulatory, and biochemical responses, and to validate it against experimental in vivo data.This work combines mathematical modelling with experimental validation. A poromechanical formulation incorporating the solid phase of the skin, the interstitium (including interstitial fluid and mobile cells), and the vascular micro-circulation was implemented using open-source software FEniCSx. In vivo tensile tests on human skin provided data to validate the model’s ability to reproduce the time-dependent mechanical response. Then an experimental study using skin indentation coupled with laser Doppler perfusion measurements was conducted during the Ph.D. to characterise the micro-circulatory response under external loading. The model was extended to include the vascular network and incorporate biological processes such as oxygen diffusion.This work thus proposes an approach to studying the interactions between mechanical stress and skin damage mechanisms through coupled modelling of mechanical and biological processes. By providing a tool that accounts for the multiscale complexity of human tissue and inter-individual variability, this research could inform improved strategies for personalised prevention and treatment, particularly in the context of pressure ulcers, wounds, and device-related injuries.Les tissus vivants, tels que la peau, sont particulièrement sensibles aux altérations mécaniques et physiologiques, pouvant conduire à l’apparition de lésions, de rougeurs inflammatoires ou de plaies ouvertes. Ces altérations, regroupées sous le terme d’« endommagement tissulaire », désignent toute dégradation de la structure ou de la fonction du tissu. Elles peuvent survenir dans des conditions physiologiques normales ou résulter de contraintes mécaniques externes, telles qu’une pression modérée mais prolongée, ou au contraire, une charge intense appliquée brutalement. Ces situations sont fréquentes chez les patients hospitalisés, les personnes en situation d’immobilité prolongée, alitées ou en fauteuil roulant, et peuvent conduire à des complications graves. Ces lésions, accentuées par le vieillissement démographique et les habitudes sédentaires, constituent un enjeu majeur de santé publique. Leur prévention et leur prise en charge reposent sur une meilleure compréhension des mécanismes sous-jacents, afin d’anticiper leur survenue et de développer des stratégies personnalisées de soin.L’apparition de lésions tissulaires résulte d’un enchevêtrement de facteurs mécaniques, biologiques et individuels. Les études animales ont mis en évidence un lien entre le chargement mécanique des tissus, la privation d’oxygène (ischémie) et la mort cellulaire, cette dernière survenant notamment lorsque la déformation mécanique dépasse un seuil critique, en compression ou en cisaillement. De plus, plusieurs travaux ont montré que les effets mécaniques et ischémiques agissent de manière synergique pour initier et propager les lésions. Toutefois, les modèles mathématiques actuels traitent encore trop souvent ces dimensions de manière séparée, sans intégrer explicitement les interactions dynamiques entre contraintes mécaniques, déformation du tissu, microcirculation et réponse biologique. Ce cloisonnement constitue une limite importante à la prédictivité et à la transférabilité de ces modèles. Par ailleurs, la majorité des données utilisées pour calibrer ces modèles provient d’animaux, rendant difficile leur application directe à la peau humaine et, a fortiori, leur adaptation à la variabilité interindividuelle des patients.L’hypothèse centrale de ce travail est qu’un cadre de modélisation poromécanique, multi-phasique et multi-échelle, intégrant les composantes structurelles, fluidiques et biologiques du tissu, permettrait de mieux caractériser le couplage entre chargement mécanique et réponse microvasculaire.La stratégie mise en œuvre combine étroitement modélisation mathématique et validation expérimentale (en conditions physiologiques). Une formulation poromécanique intégrant la phase solide du tissu cutané, l’interstitium (incluant fluide interstitiel et cellules mobiles), ainsi que la microcirculation vasculaire, a été implémentée dans le logiciel open-source FEniCSx. Des données expérimentales d’essais d’extension extit{in vivo} sur peau humaine ont permis de valider la capacité du modèle à reproduire la réponse temporelle du tissu cutané soumis à une sollicitation mécanique. Puis une étude expérimentale réalisée au cours de cette thèse, reposant sur l’indentation de la peau couplée à une mesure de la perfusion par sonde laser Doppler, a été conduite pour caractériser la réponse microcirculatoire sous chargement externe. Le modèle a été enrichi pour intégrer le réseau vasculaire et la possibilité d’incorporer des processus biologiques tels que la diffusion d'oxygène.Ce travail propose ainsi une approche pour étudier l’interaction entre les mécanismes d’endommagement cutané sous contrainte, en s’appuyant sur une modélisation couplée des processus mécaniques et biologiques. En offrant un outil capable de prendre en compte la complexité multi-échelle du tissu humain et sa variabilité interindividuelle, cette recherche pourrait inspirer de meilleures stratégies de prévention et de traitement personnalisés
Développement d’un procédé micro-plasma micro-onde pour l’élaboration de nano-carbonés hybridés sp3
No full textThis work focuses on the use of a microwave plasma torch based on a coaxial transmission line (CTRL) for the synthesis of nanodiamonds under mbar pressure conditions using H₂/CH₄ (± Ar) mixtures. Optical diagnostics of the plasma, including optical emission spectroscopy (OES) and picosecond two-photon laser-induced fluorescence (ps-TALIF), reveal that the torch produces high electron densities (2×10¹³ cm⁻³), high atomic hydrogen densities (up to 10¹⁷ cm⁻³), and gas temperatures ranging from 1000 to 1700 K. These parameters favor nanodiamond nucleation by stabilizing the diamond sp³ phase. Deposition studies confirmed the formation of nanodiamonds with a crystal structure and size distribution of around 10 nm. Deposition quality, assessed by the sp³/sp² ratio, shows an optimum value at 4% methane at 100 mbar and 90 W. The addition of argon (up to 25 sccm) further enhanced nanostructure performance by increasing gas temperatures and promoting favorable radical densities. The nucleation process was found to occur mainly in the gas phase, driven by key radicals such as CH₃, CH, C, and C₂, which form via hydrogen dissociation and hydrocarbon reactions. High concentrations of acetylene, however, led to the formation of amorphous carbon via the HACA mechanismCe travail porte sur l'utilisation d'une torche à plasma micro-ondes basée sur une ligne de transmission coaxiale (CTRL) pour la synthèse de nanodiamants dans des conditions de pression mbar en utilisant des mélanges H₂/CH₄ (± Ar). Les diagnostics optiques du plasma, comprenant la spectroscopie d'émission optique (OES) et la fluorescence picoseconde induite par laser à deux photons (ps-TALIF), révèlent que la torche produit une grande densité d’électrons (2×10¹³ cm⁻³), de grandes densités d’hydrogène atomique (jusqu’à 10¹⁷ cm⁻³) et des températures du gaz allant de 1000 à 1700 K. Ces paramètres favorisent la nucléation de nanodiamants en stabilisant la phase sp³ du diamant. Les études de dépôt ont confirmé la formation de nanodiamants avec une structure cristalline et une distribution de tailles d’environ 10 nm. La qualité du dépôt, évaluée par le rapport sp³/sp², montre une valeur optimale à 4 % de méthane à 100 mbar et 90 W. L'ajout d'argon (jusqu'à 25 sccm) a encore amélioré le rendement de la nanostructure en augmentant les températures du gaz et en favorisant des densités de radicaux favorables. On a constaté que le processus de nucléation se produisait principalement en phase gazeuse, sous l'effet de radicaux clés tels que CH₃, CH, C et C₂, qui se forment par dissociation de l'hydrogène et par des réactions d'hydrocarbures. Des concentrations élevées d'acétylène ont toutefois conduit à la formation de carbone amorphe via le mécanisme HAC