HAL - Université de Franche-Comté
Not a member yet
51444 research outputs found
Sort by
Automated classification of subsurface impact damage in thermoplastic composites using depth-resolved terahertz imaging and deep learning
No full textInternational audienceReliable detection of barely visible impact damage is critical to ensure the structural integrity of composite components in service, particularly in safetycritical applications such as pressure vessels and transportation systems. This study presents a solution for detecting such damage in woven glass fiberreinforced thermoplastic composites using terahertz (THz) time-of-flight tomography and convolutional neural networks. THz provides non-contact, non-ionizing, high-axial-resolution imaging of subsurface and back-surface damage, addressing key limitations of surface-based inspection methods. While THz imaging alone may not always permit conclusive damage identification, we bridge this gap by training neural network classifiers on depthresolved THz B-scan images using ground truth from co-located X-ray microcomputed tomography. Among several pretrained architectures tested via transfer learning, DenseNet-121 exhibits the highest accuracy. The model remains robust even when trained on truncated B-scans excluding surface indentation features, confirming its ability to detect structural anomalies located internally or on the back surface. This is particularly relevant for applications where back-side access is not feasible. Experimental validation is performed on impacted glass-fiber-reinforced thermoplastic coupons prepared in accordance with ASTM D7136, with damage severity quantified through force-displacement data and micro-tomographic analysis. Labeling for supervised learning conforms to acceptance criteria from industrial standards for composite pressure vessels (ASME BPVC Section X, CGA C-6.2), ensuring regulatory alignment and enabling deployment in quality control workflows. The proposed method minimizes the need for expert interpretation or secondary validation and offers direct applicability to in-service inspection and manufacturing quality control
Stratigraphic reconstruction from terahertz and ultrasonic signals by deconvolution: A review
No full textInternational audienc
Association Between Metabolic Syndrome, Obesity, and Cognitive Performances in Individuals With Bipolar Disorders: Cross‐Sectional and Longitudinal Analyses in the FACE ‐ BD Cohort
No full textInternational audienceIntroduction: Metabolic syndrome (MetS) has been suggested to be associated with cognitive impairments in bipolar disorder (BD); however, studies are limited by small sample sizes or cross-sectional design. Our objective is to evaluate the cross-sectional and longitudinal associations between MetS and cognitive performances in a large cohort of individuals with BD.Methods: 1175 individuals with a DSM-IV diagnosis of BD were included from the FACE-BD cohort, assessed with a standardized battery of clinical and neuropsychological tests and followed up with a cognitive retest at 2 years for a subsample (n = 367). A global cognitive index was created by using a Principal Component Analysis. Associations between MetS and cognitive performances at baseline were explored using multiple analyses of covariance and linear mixed models were used for longitudinal data.Results: The prevalence of MetS was 21.5% in this sample. Multivariable analyses identified associations between MetS and poorer cognitive performance in the cross-sectional analysis, independently of age, gender, education level, psychotropic treatments, and comorbidities. Specifically, individuals with MetS showed poorer results (global cognitive index, cognitive flexibility, inhibition, and verbal memory). After adjustment, the longitudinal analysis showed no change in the global cognitive index at year 2 and no time × metabolic syndrome interaction.Conclusions: Our results suggest that MetS is cross-sectionally, but not longitudinally, associated with poorer cognitive performances in BD. This study highlights the importance of systematically and accurately screening for metabolic abnormalities in individuals with BD, and screening for cognitive deficit especially in individuals with MetS. Our results suggest that MetS is not a risk factor for cognitive decline during the follow-up, but further longitudinal studies are required
Braneworld Baryogenesis and QCD-Era Magnetogenesis: A Predictive Link
No full textInternational audienceWe demonstrate that primordial magnetic fields (PMF) play a decisive role in the braneworld baryogenesis scenario of [Phys. Rev. D , 023520 (2024)], where C/CP violation arises from the coupling of visible and hidden matter-antimatter sectors through a pseudo-scalar field. Although this mechanism generates baryon number efficiently only after the quark-hadron transition, by incorporating a realistic stochastic PMF within a semi-analytical framework, we find that matching the observed baryon-antibaryon asymmetry robustly requires PMF strengths of order T right after the transition, in agreement with causal QCD-era magnetogenesis. We further reveal that magnetic fluctuations drive the baryon-density spectrum to white noise on large scales, yielding an isocurvature component compatible with Cosmic Microwave Background (CMB) bounds. This establishes a predictive link between the braneworld baryogenesis model and realistic early-Universe magnetic fields
Size doesn’t always matter: Greenspace connectivity can offset insufficient habitat patch size to improve urban tits breeding success
No full textInternational audienceUrban landscapes are often highly fragmented, constraining animals to live in and exploit a multitude of habitat patches (e.g., greenspaces) of varying size and isolation. Small greenspaces may not contain enough resources for species to maintain viable populations. Yet, appropriate spatial configuration of the habitat network (i.e., high greenspace connectivity) could theoretically alleviate or even compensate the local food resource limitations by allowing access to additional foraging grounds. Surprisingly, this effect has never been tested to explain the reproductive performances of urban fauna. We hypothesised that higher greenspace connectivity would improve the breeding outputs of two insectivorous bird species (tits), especially with decreasing nesting greenspace area (i.e., the habitat patch where the nest is located).For four years, we monitored the survival and mass of nestlings of Parus major and Cyanistes caeruleus using 240 nestboxes located along a multivariate urban gradient in Dijon (France), and analysed their variations with GLMMs while controlling for various confounding factors (e.g., urbanness, noise and light pollution, microclimate, vegetation management). Functional connectivity was measured through graph-based modelling.Greenspace connectivity was important to explain both nestling survival and mass while the nesting greenspace area was not a good predictor of breeding success. Furthermore, the positive effect of connectivity on nestling survival significantly increased with decreasing area of their nesting patches.Urban tits can maintain successful reproduction dynamics in small greenspaces provided those are well connected to the network of preferred foraging habitats. Promoting the connectivity of interstitial urban greenspaces could thus partly compensate for their frequently insufficient sizes in cities
Neurocognitive Correlates of Food and Alcohol Disturbances: An Integrated Neuropsychological Investigation
No full textInternational audienceFood and Alcohol Disturbance (FAD) represents a functional relationship between alcohol use and eating behaviors, in which individuals engage in disordered eating to enhance alcohol intoxication and/or compensate for alcohol-related caloric intake. FAD is highly prevalent among young adults, particularly University students. While the biopsychosocial correlates of FAD are documented, its specific neurocognitive correlates remain unexplored, despite extensive literature describing the distinct neurocognitive correlates of alcohol consumption and eating disorders. We therefore investigated whether FAD is associated with neurocognitive correlates in university students and examined whether different FAD subdimensions relate to distinct cognitive profiles. We assessed FAD in 130 French university students using the CEBRACS scale and administered an extensive neuropsychological battery measuring visuospatial abilities, episodic memory, and executive functions. We compared cognitive performance between individuals who do and do not engage in FAD and then conducted exploratory multivariate regression analyses to identify variations in cognitive profiles across the CEBRACS subscales. The general comparison between individuals who do and do not engage in FAD did not reveal significant differences. Conversely, analyses of the CEBRACS subscales identified specific patterns: (1) dietary restraint was associated with poorer visuospatial abilities and verbal episodic memory; (2) purging behaviors were associated with lower executive functioning but improved visual episodic memory; (3) extreme fasting and self-induced vomiting were associated with poorer visual episodic memory performance but higher executive functioning. These findings suggest that FAD is an umbrella term encompassing various cognitive profiles according to the distinct eating behaviors involved and highlight the importance of considering the subcomponents of FAD when exploring its neurocognitive correlates
Experimental determination of organ-specific fast neutron equivalent dose from modern radiotherapy LINACs
No full textInternational audienc
Nivolumab in Metastatic Clear-cell Renal Cell Carcinoma: An Integrative Biomarker Analysis from the NIVOREN GETUG-AFU 26 Phase 2 Study
No full textInternational audienceNivolumab improved survival in patients with refractory metastatic clear-cell renal cell carcinoma (ccRCC), but no reliable biomarker of activity has been identified. We conducted a real-world phase 2 trial of nivolumab in patients progressing after one or more vascular endothelial growth factor (VEGF) receptor-directed therapies, which included an integrated translational programme. Candidate tissue and circulating biomarkers were assessed using immunoassays and gene expression profiling. Overall, 720 patients were treated, with activity and safety in line with pivotal trial data. Exploration of tissue architecture showed that the presence of tertiary lymphoid structures, CD8+ lymphocytes, and CD163+ macrophage infiltration at the invasive margin were all marginally associated with longer progression-free survival, similarly to PD-1 expression on immune cells. Expression of hypoxia-related marker VEGF on tumour cells was however strongly associated with shorter progression-free and overall survival. Recapitulation of microenvironment composition based on gene expression signatures showed that patients harbouring a high tumour lymphocyte infiltration, concomitantly to low infiltration of neutrophil and non-immune stromal cells, had improved response to nivolumab. Conversely, circulating cytokines related to protumoral inflammation interleukin (IL)-6 and IL-8 were independently associated with shorter progression-free and overall survival. Overall, immune and angiogenic features helped inform outcomes to nivolumab. Circulating factors were best potential predictors for immunotherapy activity in ccRCC
Phi-FEM-FNO: a new approach to train a Neural Operator as a fast PDE solver for variable geometries
No full textInternational audienceIn this paper, we propose a way to solve partial differential equations (PDEs) by combining machine learning techniques and the finite element method called phi-FEM. For that, we use the Fourier Neural Operator (FNO), a learning mapping operator. The purpose of this paper is to provide numerical evidence to show the effectiveness of this technique. We will focus here on the resolution of two equations: the Poisson-Dirichlet equation and the non-linear elasticity equations. The key idea of our method is to address the challenging scenario of varying domains, where each problem is solved on a different geometry. The considered domains are defined by level-set functions due to the use of the phi-FEM approach. We will first recall the idea of -FEM and of the Fourier Neural Operator. Then, we will explain how to combine these two methods. We will finally illustrate the efficiency of this combination with some numerical results on three test cases. In addition, in the last test case, we propose a new numerical scheme for hyperelastic materials following the phi-FEM paradigm