HAL Clermont Université
Not a member yet
    90968 research outputs found

    L’intelligence artificielle en assistance médicale à la procréation (AIVFTM) permet-elle d'optimiser le choix de l'embryon ?

    No full text
    Introduction: selecting the embryo that will give the best chance of pregnancy during assisted reproductive technology (ART) is a major challenge to improve pregnancy rates. Microscopic evaluation of embryo morphology is the usual method used to select the embryo to be transferred into the uterus. However, this observational method has its limitations, as it can be a source of inter- and intra-observer variability due to its potential subjectivity. Despite the introduction of a new technology called time-lapse and related algorithms for assessing embryonic morphology and kinetics, pregnancy rates in ART are not yet optimized and can be further enhanced. The use of artificial intelligence (AI) is developing in the medical field and seems a promising solution for optimizing results in ART while avoiding multiple pregnancies, by optimizing the selection of the embryo chosen for transfer to increase the chances of pregnancy. EMA is a secure digital AI platform developed by AIVFTM. Its use seems to provide a standardized and reproducible assessment of embryo development up to day 5 of embryo culture in ART. Furthermore, through the application of an algorithm, this system would provide assistance in choosing the embryo for transfer with the highest chances of pregnancy.Objective: the main objective of this retrospective study is to evaluate the use of artificial intelligence EMA (AIVFTM) as an aid to embryo selection in ART in comparison with the routinely applied evaluation of morphological and morphokinetic criteria in the ART laboratory at Clermont Ferrand University Hospital.Materials and Methods: this retrospective study was carried out in the ART laboratory at Clermont-Ferrand University Hospital between January 2022 and December 2023. 619 fresh and frozen blastocyst-stage embryo transfers from 420 in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) cycles with known clinical outcomes were retrospectively analyzed. Embryos were cultured in time-lapse until day 5 (n=590) and day 6 (n=29). EMA scores were compared with standard morphological (Gardner classification) and morphokinetic (Geri score) assessments.Results: EMA scores correlated significantly with blastocyst morphology, including trophectoderm quality, internal cell mass, degree of expansion, and overall blastocyst quality. Higher EMA scores were associated with better implantation and live birth rates for both fresh and frozen transfers. A threshold EMA score of 7.85 was identified, above which results were significantly better (implantation: 36.75%; live birth: 32.48%). This threshold outperformed both conventional morphological selection criteria and selection criteria based on the Geri score in predicting clinical pregnancy, with an odds ratio of 2.07 (p=0.022). When the embryos with the best EMA score matched those selected by the biologist, the results in terms of clinical pregnancy and live birth rates were significantly higher.Conclusion: the EMA score provides a robust, standardized and reproducible basis for embryo selection, showing superior predictive power for live birth compared with conventional methods. This study is the first clinical evaluation of the AIVFTM system in France and has demonstrated the potential value of AI in optimizing results in ART. It could improve embryoselection and reduce the number of transfers needed to achieve pregnancy.INTRODUCTION : le choix de l’embryon qui donnera le plus de chance de grossesse lors d’une prise en charge en Assistance Médicale à la Procréation (AMP) est un enjeu majeur pour pouvoir améliorer les taux de grossesse. L’évaluation de la morphologie de l’embryon en microscopie est la méthode usuelle d’évaluation pour sélectionner l’embryon à transférer dans l’utérus. Cependant, cette méthode observationnelle présente des limites car elle peut être source d’une variabilité inter- et intra-observateur du fait de sa subjectivité potentielle. Malgré l’introduction d’une nouvelle technologie appelée time-lapse et des algorithmes attenants par l’évaluation de la morphologie et la cinétique embryonnaire, les taux de grossesse en AMP ne sont pas encore optimisés et peuvent être encore améliorés. L’utilisation de l’intelligence artificielle (IA) se développe dans le domaine médical et paraît une option prometteuse pour optimiser les résultats en AMP tout en évitant les grossesses multiples par une optimisation de la sélection de l’embryon choisi pour le transfert pour augmenter les chances de grossesse. EMA est une plateforme numérique d’IA sécurisée développée par la société AIVFTM. Son utilisation semble pouvoir apporter une évaluation standardisée et reproductible du développement de l’embryon jusqu’au 5ème jour de culture embryonnaire en AMP. De plus, par l’application d’un algorithme, ce système apporterait une aide dans le choix de l’embryon en vue d’un transfert apportant les chances de grossesse les plus élevées.OBJECTIF : l’objectif principal de cette étude rétrospective est d’évaluer l’utilisation de l’intelligence artificielle EMA (AIVFTM) comme aide au choix de l’embryon en AMP en comparaison à l’évaluation appliquée en routine des critères morphologiques et morphocinétiques dans le laboratoire d’AMP au CHU de Clermont-Ferrand.MATÉRIELS ET MÉTHODES : cette étude rétrospective a été menée sein du service AMP-CECOS du CHU de Clermont-Ferrand entre janvier 2022 et décembre 2023. 619 transferts d'embryons au stade de blastocystes frais et congelés issus de 420 tentatives de fécondation in vitro (FIV)/injection intracytoplasmique de spermatozoïde (ICSI) dont les issues cliniques étaient connues ont été analysés rétrospectivement. Les embryons ont été cultivés en time-lapse (TLS) jusqu’au 5ème jour de culture (n = 590) et 6ème jour (n = 29). Les scores EMA ont été comparés aux évaluations morphologiques standard (classification de Gardner) et morphocinétiques (score Geri).RÉSULTATS : les scores EMA attribués par le système d’IA sont positivement corrélés à la morphologie du blastocyste, que ce soit pour la qualité du trophectoderme, de la masse cellulaire interne, du degré d’expansion ainsi que la qualité globale du blastocyste. Les embryons frais ou congelés ayant un EMA élevé sont associés à de meilleurs taux d'implantation et de naissances vivantes. Nous avons mis en évidence un score seuil EMA de 7,85, au-delà duquel les résultats en termes de chance d’implantation (36.75 %) et de naissance vivante (32.48 %) sont significativement meilleurs. Ce seuil semble plus performant que les critères de sélection morphologiques conventionnels et les critères de sélection basés sur le score morphocinétique « Geri » pour prédire une grossesse clinique, avec un odd ratio de 2,07 (p = 0,022). Lorsqu’il existe une concordance entre les embryons avec le meilleur score EMA et ceux sélectionnés par le biologiste, les résultats en termes de taux de grossesse clinique et de naissance vivante sont significativement plus élevés.CONCLUSION : le score EMA apporte une base robuste, standardisée et reproductible pour la sélection d'embryons, montrant un pouvoir prédictif supérieur pour la naissance vivante par rapport aux méthodes conventionnelles. Cette étude est la première évaluation clinique du système AIVFTM en France et a démontré la valeur potentielle de l'IA dans l'optimisation des résultats en AMP. Elle permettrait d’améliorer la sélection des embryons et de réduire le nombre de transferts nécessaires pour obtenir une grossesse

    : REphon-IA : Preliminary research protocol for expert system calibration

    No full text
    International audienceAs part of the REphon-IA project, a preliminary research protocol was set up to provide a qualitative and quantitative calibration of the expert system intended to feed artificial intelligence with the aim of improving the pronunciation and oral communication of allophone speakers. The specific aim of the protocol was to identify possible phonetic errors in allophone speakers. The methodology employed was based on the cognitive processes of language processing, which show that all speakers develop phonological deafness concomitant with the development of their mother tongue expertise. Stimulation of the phonatory audio loop by a foreign-language repetition activity provides clues as to the phonetic characteristics of the mother tongue or the languages used. The process of identifying possible phonetic "errors" was based on a rehearsal activity involving utterances calibrated for rigorous control of the variables under study.The presentation will provide details on the selection of variables and the calibrations carried out in order to build an effective corpus for obtaining a qualitative analysis of possible pronunciation "errors" produced in real time. It will also present the characteristics taken into consideration when constructing utterances in order to limit uncontrolled effects such as those of fatigue, affectivity or access to meaning. It will then focus on the various phonological, rhythmic and melodic variables taken into consideration. Analysis of the typology of "errors" obtained in the search ecology provides us with valuable clues to take into consideration with a view to automated error processing and machine calibration for AI parameterization for phonetic integration activities.Dans le cadre du projet REphon-IA , un protocole de recherche préliminaire a été mis en place afin d’offrir un calibrage qualitatif et quantitatif du système expert voué à alimenter l’intelligence artificielle dans le but d'améliorer la prononciation et la communication orale des locuteurs allophones. Le but spécifique du protocole était d’identifier les possibles erreurs phonétiques des locuteurs allophones. La méthodologie employée s’est fondée sur les processus cognitifs de traitement du langage qui montrent que tout locuteur développe une surdité phonologique concomitante au développement de son expertise en langue maternelle. La stimulation de la boucle audio phonatoire par une activité de répétition en langue étrangère offre des indices sur les caractéristiques phonétiques maternelles ou des langues d’usage. Le processus d’identification des possibles « erreurs » phonétiques s’est basé sur une activité de répétition d’énoncés calibrés pour un contrôle rigoureux des variables étudiées. L’intervention apportera des précisions sur la sélection des variables et les calibrages opérés afin de construire un corpus efficace pour obtenir une analyse qualitative des possibles « erreurs » de prononciation produites en temps réel. Elle présentera ainsi les caractéristiques prises en considération lors de la construction des énoncés afin de limiter les effets incontrôlés tels que ceux de la fatigue, de l’affectivité ou de l’accès au sens. Elle se concentrera ensuite sur les différentes variables phonologiques, rythmiques et mélodiques prises en considération. L’analyse de la typologie des « erreurs » obtenues dans l’écologie de la recherche nous donne de précieux indices à prendre en considération en vue d’un traitement automatisé des erreurs et du calibrage de la machine en vue du paramétrage de l’IA pour des activités d’intégration phonétique

    La collaboration multimétiers conseiller·ère en prévention – ergonome dans le milieu agricole : questionnements à partir d’une intervention d’accompagnement d’un projet de conception

    No full text
    International audienceAt the border of their own communities, ergonomist and risk preventer work together to build common actions. This communication is about one of these collaborations, in the context of support interventions for conception projects in the agricultural field. Based on an example of intervention for a market gardening firm, we analyse the key-moments of this multiprofession collaboration. It has contributed to help evolve the structuration of this project and the conception choices to make them more conducive to health and efficiency. This common activity analysis between risk preventer and ergonomist allows us to discuss the conditions of success of such collaboration, in order to contribute to the development of primary prevention practices, in the accompaniment of projects in agricultural field.Aux frontières de leurs communautés respectives, ergonomes et conseiller·ère·s en prévention collaborent pour construire des actions communes. Cette communication s’appuie sur une de ces collaborations, dans le cadre des accompagnements de projets de conception en milieu agricole. À partir d’un exemple d’intervention menée auprès d’une entreprise maraîchère, nous analysons les moments clés de cette coopération multimétiers. Cette coopération a contribué à faire évoluer la structuration du projet et les choix de conception afin de les rendre plus favorables à la santé et à l’efficacité. Cette analyse de l’activité conjointe entre préventeur et ergonome, nous permet d’engager une discussion sur les conditions de réussite d’une telle collaboration dans la perspective du développement des pratiques de prévention primaire dans l’accompagnement des projets dans le milieu agricole

    Income Shocks, Bride Price, and Child Marriage in Turkey

    No full text
    International audienceAbstract Early marriage persists in many countries in spite of legal prohibition. The role of income shocks and bride price norms is investigated in the context of Turkey. Using data from the Turkey Demographic and Health Surveys 1998 to 2018, rainfall shocks are exploited as an exogenous source of variation of household income. To study the role of marital payments, measures of shocks are interacted with a province-level indicator based on past prevalence of bride price. Estimation results show that girls exposed to a large negative shock during adolescence have a higher probability of being married before 15, if living in provinces with a high prevalence of bride price. More arranged and forced marriages are found after severe shocks in those provinces. Results suggest that daughter marriage, as a response to mitigate negative income shocks, is not limited to the poorest countries, and still contributes to household strategies in Turkey

    Effet de la température ambiante sur l'évolution de la concentration en NaCl dans la sueur lors d'un exercice sous maximal

    No full text
    International audiencePurpose: The human body primarily regulates core temperature through evaporation sweat loss, especially during physical activity. Previous studies suggested that an increased sweat rate (SR) leads to higher sweat-related sodium chloride (NaCl) concentration due to reduced NaCl reabsorption in sweat glands. In the heat, a rise in body core temperature and loss of body water via sweating are consequences of prolonged exercise. However, current methods for assessing SR and NaCl concen-tration in sweat do not allow real-time measurement, limiting the ability to make immediate adjust-ments in hydration or electrolyte management. This study aimed to compare the impact of ambient temperature on the evolution of SR and NaCl concentration in sweat using the BeOne armband, a device that continuously measures sweating throughout exercise in real time.Methods: 45 well-trained male athletes 45 male athletes (22±5 y ; 1.77±0.06 m ; 71±13 kg ; VO2 max : 66±11 ml.min-1.kg-1) performed a 45-minute cycling exercise at 50% of power output at which VO₂max was reached under two environmental conditions: a thermos-neutral climate (N: 24 ± 1°C, 30.0 ± 3% RH) and a hot climate (H: 35 ± 1°C, 30 ± 3% RH). During sessions NaCl concentration was continuously monitored using the BeOne armband, while whole-body SR was assessed via body mass changes. Core temperature (Tc), skin temperature (Ts), heart rate (HR), and heat production (indirect calorimetry) were also measured.Results: An increase in ambient temperature led to a shorter sudation delay from the armband sensor and a higher sweat rate (T-test). NaCl concentrations were time-normalized relative to the individual sweat measurement duration. ANOVA revealed a significant increase in sweat NaCl concentration starting from 30% of the sweat measurement period. In addition, core temperature, skin temperature, heart rate, RPE and thermal sensations increased with hot temperature.Conclusion: Higher ambient temperature led to faster sweating, increased sweat rate, and elevated NaCl concentration in sweat. Real-time, continuous measurement of NaCl concentration suggests that sweat rate increases progressively during the session. This study confirms the relationship between sweat rate and NaCl concentration, suggesting that higher sweat rate is associated with reduced NaCl reabsorption. Continuous monitoring of these variables in real time may provide more precise insights into athletes’ thermoregulatory capacity

    Advection, diffusion and linear transport in a single path-sampling Monte-Carlo algorithm: Getting insensitive to geometrical refinement

    No full text
    International audienceWe address the question of numerically simulating the coupling of diffusion, advection and one-speed linear transport, with a specific focus on managing geometrical complexity. We base our work on recent advances from the computer graphics community, which has developed Monte Carlo algorithms simulating linear radiation transport in physically realistic scenes, with numerical costs that remain unaffected by geometrical refinement: adding more details to the scene description does not impact the computation time. The resulting benefits in terms of engineering flexibility are already fully integrated into the cinema industry and are gradually being adopted by the video game industry. Here we demonstrate that the same insensitivity to the geometric complexity can be achieved when considering not only one-speed linear transport, but also its coupling with diffusion and advection. In this case, pure linear-transport paths are replaced with advection-diffusion/linear-transport paths, which are composed of subpaths. Each subpath represents one of the three physical phenomena, and coupling is handled by switching from one subpath (i.e. phenomenon) to another. This approach is illustrated using a porous medium involving up to 10,000 pores, with the computation time being strictly independent of the number of pores, showing its ability to facilitate engineering calculations in complex geometries

    Microbiology Galaxy Lab: The first community-driven gateway for reproducible and FAIR analysis of microbial data

    No full text
    The explosion of microbial omics data has outpaced the ability of many researchers to analyze it, with complex tools and limited computational resources creating barriers to discovery. To address this gap, we present the Microbiology Galaxy Lab: a free, globally accessible, community-supported platform that combines state-of-the-art analytical power with user-friendly accessibility. Supported by the Galaxy and global microbiology communities, this platform integrates over 315 tool suites and 115 curated workflows, enabling comprehensive metabarcoding, (meta)genomic, (meta)transcriptomic, and (meta)proteomic data analysis within a FAIR-aligned environment. It also supports research in the health and infectious disease sectors, as well as in environmental microbiology. The platform's utility is exemplified through various use cases, including antimicrobial resistance tracking, biomarker prediction, microbiome classification, and functional annotation of key microbes. Built on reproducibility and community engagement, it supports creation, sharing, and updating of best-practice workflows. Over 35 tutorials and learning paths empower scientists, fostering an ecosystem that keeps resources at the forefront of microbial science. The Microbiology Galaxy Lab enables collective analysis, democratising research, thereby accelerating discovery across the global microbiology community (microbiology.usegalaxy.org, .eu, .org.au, .fr)

    Just Another Craft? The Visual Arts in Jan van der Straet (1523-1605)’s Nova Reperta

    No full text
    International audienc

    On the identification of BiOF, Bi7O5F11 and β-BiO0.5F2 bismuth oxyfluorides

    No full text
    Heterogeneous photocatalysis is a promising strategy for environmental remediation and sustainable energy production. Within this field, bismuth-based photocatalysts stand out thanks to their distinctive optical, electronic, and chemical properties. Bismuth oxyfluorides (BiOxF3-2x) are particularly notable, forming a structural and compositional continuum between BiF3 and Bi2O3. However, the precise identification of oxyfluoride phases remains difficult, often leading to misassignments in the literature. This document addresses that challenge by proposing reliable analytical approaches for identifying three representative compounds: BiOF, Bi7O5F11, and β-BiO0.5F2. The materials are comprehensively characterized using XRD, FTIR, Raman spectroscopy, XPS, and SEM-EDX

    (Dé)mobilisations en Bourgogne : les conflictualités sociales et politiques contemporaines en question

    No full text
    International audienceCompte-rendu de la journée « (Dé)mobilisations en Bourgogne : les conflictualités sociales et politiques contemporaines en question », organisée par Romain Castellesi (Université du Littoral Côte-d’Opale), François Jarrige (Université de Bourgogne) et Adeline Blaszkiewicz-Maison (Université Paris 1), qui s’est tenue le 17 octobre 2024 à Dijon

    5,201

    full texts

    90,968

    metadata records
    Updated in last 30 days.
    HAL Clermont Université
    Access Repository Dashboard
    Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇