Université de Toulouse

The Congolese upwelling system (CoUS), located along the West African coast north of the Congo River, is one of the most productive and least studied systems in the Gulf of Guinea. The minimum sea surface temperature in the CoUS occurs in austral winter, when the winds are weak and not particularly favorable to coastal upwelling. Here, for the first time, we use a high‐resolution regional ocean model to identify the key atmospheric and oceanic processes that control the seasonal evolution of the mixed layer temperature in a 1°‐wide coastal band from 6°S to 4°S. The model is in good agreement with observations on seasonal timescales, and in particular, it realistically reproduces the signature of the surface upwelling during the austral winter, the shallow mixed layer due to salinity stratification, and the signature of coastal wave propagation. The analysis of the mixed layer heat budget for the year 2016 reveals a competition between warming by air‐sea fluxes, dominated by the incoming shortwave radiation throughout the year, and cooling by vertical mixing at the base of the mixed layer, as other tendency terms remain weak. The seasonal cooling is induced by vertical mixing, where local wind‐driven dynamics play a secondary role compared to subsurface processes. A subsurface analysis shows that remotely forced coastal‐trapped waves raise the thermocline from April to August, which strengthens the vertical temperature gradient at the base of the mixed layer and leads to the mixing‐induced seasonal cooling in the Congolese upwelling system.

Photochemistry has provided a powerful alternative to generate reactive intermediates under mild reaction conditions and photobiocatalysis is currently widely applied in organic synthesis. While enzyme‐based photocatalysis has attracted a lot of attention, light‐driven DNA‐based biohybrid systems are also increasingly being applied to a large variety of applications from molecular imaging to enantioselective catalysis, and other fields such as DNA‐encoded libraries (DELs). This review gives insights into these research fields by classifying the existing light‐driven DNA‐based biohybrid systems developed to date.

Protective immune responses require close interactions between conventional (Tconv) and regulatory T cells (Treg). The extracellular mediators and signaling events that regulate the crosstalk between these CD4⁺ T cell subsets have been extensively characterized. However, how Tconv translate Treg-dependent suppressive signals at the chromatin level remains largely unknown. Here we show, using a murine bone marrow allograft model in which graft rejection is coordinated by CD4⁺ T cells and can be inhibited by Treg, that Treg-mediated T cell suppression involves Heterochromatin Protein 1 α (HP1α)-dependent gene silencing. Unexpectedly, our screen also reveals that T cells deficient for HP1γ or the methyltransferase SUV39H1 are better repressed by Treg than their wild-type counterparts. Mechanistically, our transcriptional and epigenetic profiling identifies HP1γ as a negative regulator of a gene network functionally associated with T-cell exhaustion, including those encoding the inhibitory receptors PD-1 and LAG-3. In conclusion, we identify HP1 variants as rheostats that finely tune the balance between tolerance and immunity. While HP1α converts immunosuppressive signals into heterochromatin-dependent gene silencing mechanisms, HP1γ adjusts Tconv sensitivity to inhibitory environmental signals.

Background and Purpose Working memory, a primary cognitive domain, is often impaired in pediatric brain tumor survivors, affecting their attention and processing speed. This study investigated the long‐term effects of treatments, including surgery, radiotherapy (RT), and chemotherapy (CT), on working memory tracts in children with posterior fossa tumors (PFTs) using resting‐state functional MRI (rs‐fMRI) and diffusion MRI tractography. Methods This study included 16 medulloblastoma (MB) survivors treated with postoperative RT and CT, 14 pilocytic astrocytoma (PA) survivors treated with surgery alone, and 16 healthy controls from the Imaging Memory after Pediatric Cancer in children, adolescents, and young adults study (NCT04324450). Working memory tracts were identified by combining seed masks from rs‐fMRI maps and whole‐brain tractography from diffusion MRI. Connectivity alterations were assessed qualitatively and quantitatively, alongside neuropsychological evaluations and correlations with behavioral outcomes and mean supratentorial dose. Results Compared to controls, MB survivors exhibited significant impairments in the working memory network, including reductions in tract volume (TV), fiber density, fiber cross‐section (FC), mean streamline length (MLS), and fractional anisotropy (FA) (all p = 0.04). Lower working memory scores were correlated with reduced TV and FA in MB survivors. Higher mean supratentorial doses were associated with lower TV, FC, and FA values across multiple tracts, particularly in the arcuate and superior longitudinal fasciculi. Conclusions Tractography‐derived features highlighted white matter damage as a biomarker of treatment‐related neurotoxicity in PFTs survivors. These findings underscore the detrimental impact of RT and CT on working memory networks and emphasize the importance of preserving cognitive function during treatment planning.

Background The Amyloid Imaging to Prevent Alzheimer’s Disease (AMYPAD) Prognostic & Natural History Study (PNHS) is a prospective longitudinal PET cohort of over 1,500 non‐demented individuals from 10 parent cohorts across Europe. We provide an overview of ongoing efforts to curate and integrate magnetic resonance imaging (MRI) multimodal images across sites and to extract biologically meaningful information (i.e., image‐derived phenotypes; IDPs) in this early AD population. Data will be made available on the ADDI platform (fair.addi.ad‐datainitiative.org). Method Imaging protocols included core (T1w, T2w, T2*, FLAIR) and advanced (rs‐fMRI, SWI, DWI, and ASL) MRI sequences. Figure 1 provides an overview of the different acquisitions, their corresponding pipelines, and the obtained IDPs. For T1w images, we computed regional volumes and thickness values using FreeSurfer v7.1.1 (surfer.nmr.mgh.harvard.edu/). Morphometric similarity networks were computed using MIND (github.com/isebenius/MIND). White matter hyperintensity volumes were obtained from FLAIR images using Bayesian Model Selection (BaMoS). With rs‐fMRI, mean functional connectivity of canonical networks were extracted using FSL MELODIC and dual‐regression analyses after preprocessing with fMRIPrep v23.0.1 (fmriprep.org). Diffusion MRI scans were processed with QSIprep v0.16 (qsiprep.readthedocs.io) to compute tract‐based spatial statistics and tractograms to build structural connectivity matrices. Finally, cerebral blood flow and the spatial coefficient of variation were computed from ASL images using ExploreASL (github.com/ExploreASL). Statistical harmonization was performed using RELIEF (github.com/junjypark/RELIEF). Result Baseline and follow‐up characteristics of the 1537 subjects with available T1w are described in Table 1. Raw and statistically harmonized variants of 380 core and 298 advanced IDPs were calculated. Figure 2 shows the distribution of hippocampal volumes per site before and after statistical harmonization, and their relation with participants' age (R2raw=0.179, R2harmonized=0.422). Considering recent developments in the field, the AMYPAD PNHS aims to include additional IDPs associated with vascular health, such as perivascular spaces and other glymphatic system‐related markers. Conclusion We present the pipeline built for MRI harmonization and feature extraction of the AMYPAD PNHS dataset. The extracted IDPs can help identify novel imaging outcomes, support the development of disease progression models for the preclinical stages of AD, and provide a reference point for future studies to promote replicability and robustness of findings.

Background The Amyloid Imaging to Prevent Alzheimer’s Disease (AMYPAD) Prognostic & Natural History Study (PNHS) is a prospective longitudinal PET cohort of over 1,500 non‐demented individuals from 10 parent cohorts across Europe. We provide an overview of ongoing efforts to curate and integrate magnetic resonance imaging (MRI) multimodal images across sites and to extract biologically meaningful information (i.e., image‐derived phenotypes; IDPs) in this early AD population. Data will be made available on the ADDI platform (fair.addi.ad‐datainitiative.org). Method Imaging protocols included core (T1w, T2w, T2*, FLAIR) and advanced (rs‐fMRI, SWI, DWI, and ASL) MRI sequences. Figure 1 provides an overview of the different acquisitions, their corresponding pipelines, and the obtained IDPs. For T1w images, we computed regional volumes and thickness values using FreeSurfer v7.1.1 (surfer.nmr.mgh.harvard.edu/). Morphometric similarity networks were computed using MIND (github.com/isebenius/MIND). White matter hyperintensity volumes were obtained from FLAIR images using Bayesian Model Selection (BaMoS). With rs‐fMRI, mean functional connectivity of canonical networks were extracted using FSL MELODIC and dual‐regression analyses after preprocessing with fMRIPrep v23.0.1 (fmriprep.org). Diffusion MRI scans were processed with QSIprep v0.16 (qsiprep.readthedocs.io) to compute tract‐based spatial statistics and tractograms to build structural connectivity matrices. Finally, cerebral blood flow and the spatial coefficient of variation were computed from ASL images using ExploreASL (github.com/ExploreASL). Statistical harmonization was performed using RELIEF (github.com/junjypark/RELIEF). Result Baseline and follow‐up characteristics of the 1537 subjects with available T1w are described in Table 1. Raw and statistically harmonized variants of 380 core and 298 advanced IDPs were calculated. Figure 2 shows the distribution of hippocampal volumes per site before and after statistical harmonization, and their relation with participants' age (R² raw=0.179, R² harmonized=0.422). Considering recent developments in the field, the AMYPAD PNHS aims to include additional IDPs associated with vascular health, such as perivascular spaces and other glymphatic system‐related markers. Conclusion We present the pipeline built for MRI harmonization and feature extraction of the AMYPAD PNHS dataset. The extracted IDPs can help identify novel imaging outcomes, support the development of disease progression models for the preclinical stages of AD, and provide a reference point for future studies to promote replicability and robustness of findings.

Background The early mortality of trauma patients, mainly from hemorrhagic shock, raises interest in detecting the presence of non-exteriorized bleeding. Intra-hospital EFAST (Extended Focused Assessment with Sonography for Trauma) has demonstrated its utility in the assessment and management of severe trauma patients (STP). However, there is a lack of data regarding the diagnostic performance of prehospital EFAST (pEFAST). The main objective of our study was to evaluate the pEFAST performance to predict a positive CT scan in STP. Methods This was a retrospective, multicenter, database-driven study. All severe trauma patients managed by a prehospital medical team were included. The results of pEFAST were compared with the admission CT scan. Results Data from 495 patients were included. The pEFAST had sensitivity of 27% (95% CI 22; 32) and specificity of 94% (95% CI 90; 97) for predicting the presence of a lesion on CT scan at hospital admission. The area under the curve (AUC) was 0.66 (95% CI 0.57; 0.63), the positive predictive value 84% (95% CI 75; 87), the negative predictive value was 51% (95% CI 44; 66), the positive likelihood ratio was 4.24 (95% CI 2.46; 7.3) and the negative likelihood ratio 0.78 (95% CI 0.72; 0.85). Conclusion Prehospital EFAST has an excellent specificity but a poor sensitivity for predicting a positive CT scan on hospital admission. We do not know whether this low sensitivity is secondary to the delay between the two examinations or to the poor performance of pEFAST. Therefore, a negative pEFAST should not be reassuring. A positive pEFAST is highly informative, as it predicts a lesion and enables hospital management to be prepared accordingly.

Surrogate models of turbulent diffusive flames could play a strategic role in the design of liquid rocket engine combustion chambers. The present article introduces a method to obtain data-driven surrogate models for coaxial injectors, by leveraging an inductive transfer learning strategy over a U-Net with available multifidelity Large Eddy Simulations (LES) data. The resulting models preserve reasonable accuracy while reducing the offline computational cost of data-generation. First, a database of about 100 low-fidelity LES simulations of shear-coaxial injectors, operating with gaseous oxygen and gaseous methane as propellants, has been created. The design of experiments explores three variables: the chamber radius, the recess-length of the oxidizer post, and the mixture ratio. Subsequently, U-Nets were trained upon this dataset to provide reasonable approximations of the temporal-averaged two-dimensional flow field. Despite the fact that neural networks are efficient non-linear data emulators, in purely data-driven approaches their quality is directly impacted by the precision of the data they are trained upon. Thus, a high-fidelity (HF) dataset has been created, made of about 10 simulations, to a much greater cost per sample. The amalgamation of low and HF data during the the transfer-learning process enables the improvement of the surrogate model’s fidelity without excessive additional cost.

Ruthenium (Ru) has been identified as a durable and relevant substitute to copper (Cu) to answer the access resistance lowering of the back‐end‐of‐line (BEOL) metal levels, which is a high‐priority concern for future devices. Herein, the nonequilibrium and local properties of pulsed scanning laser annealing (SLA) technique are used to enhance the structural and electrical properties of thin polycrystalline Ru layers (<30 nm). For the best annealing conditions, transmission electron microscopy observations show a substantial grain size enlargement, with large grains (≈80 nm) occupying the whole layer height. It goes with a 53% resistivity reduction, measured by 4‐point probe, confirming the strong grain boundary scattering reduction. A Mayadas–Shatzkes model incorporating temperature‐dependent resistivity measurements allows the extraction of promising reflectivity and specularity coefficients of around 0.58 and 0.98, respectively. Beyond the best conditions, failure modes for devices integration are observed, such as surface wrinkling and local buckling. Given the studied system, a semiquantitative analysis of these phenomena is given and simulations based on the finite element method are used to find further optimal annealing conditions. This study confirms the potential of Ru as a promising BEOL material, but also SLA as a convincing technique for future 3D architectures.

Background Thalamic strokes produce neurological, cognitive, and behavioral symptoms depending on the thalamic nuclei involved. While traditionally associated with severe cognitive deficits, recent studies suggest more modest impairments. This study aims to identify the factors that influence the severity of cognitive impairment following thalamic stroke. Methods We recruited 40 patients (mean age 51.1) with chronic isolated thalamic stroke and 45 healthy subjects (mean age 48.5) who underwent neuroimaging and neuropsychological assessment. Cluster and principal component analyses were used to discriminate patients from healthy subjects based on cognitive tasks. Disconnectome maps and cortical thickness were analyzed to understand the distant impact of thalamic strokes. Results Two cognitive profiles emerged from the cluster analysis. Cluster 1 included mostly healthy subjects (n = 43) and patients with no or minor deficits (n = 20). Cluster 2 included patients (n = 19) and two healthy subjects with severe deficits in verbal memory, executive functions, and attention. Cluster 1 encompassed all patients with right thalamic stroke, while Cluster 2 included all patients with bilateral stroke or mammillothalamic tract interruption. Patients with left-sided stroke were equally divided between clusters. Significant differences between clusters included age, education, interthalamic adhesion disruption, lesion volume, and location. Patients with left-sided stroke in Cluster 2 had more lateral thalamic lesions and greater disruption of the anterior thalamic projection. Conclusions Contrary to common expectations, our findings suggest that many patients with thalamic stroke have relatively good cognitive outcomes. In contrast, we identified the factors behind poor outcomes that will help clinicians.

Background To reduce the number of deaths caused by exsanguination, the initial management of severe trauma aims to prevent, if not limit, the lethal triad, which consists of acidosis, coagulopathy, and hypothermia. Recently, several studies have suggested adding hypocalcemia to the lethal triad to form the lethal diamond, but the evidence supporting this change is limited. Therefore, the aim of this study was to compare the lethal triad and lethal diamond for their respective associations with 24-h mortality in severe trauma patients receiving transfusion. Methods We performed a multicenter retrospective analysis of patients in TraumaBase®, a French database (2011–2023). The patients included in this study were all trauma patients who had received transfusions of at least 1 unit of red blood cells (RBCs) within the first 6 h of hospital admission and for whom ionized calcium measurements were available. Hypocalcemia was defined as an ionized calcium level < 1.1 mmol/L. Results A total of 2141 severe trauma patients were included (median age: 39, interquartile range [IQR]: 26–57; median injury severity score: 27, IQR: 17–41). Patients primarily presented with blunt trauma (81.7%), and a 24-h mortality rate of 16.1% was observed. Receiver operating characteristic curve analysis revealed no significant difference in the association with 24-h mortality between the lethal diamond (area under the curve [AUC]: 0.71) and the lethal triad (AUC: 0.72) (p = 0.26). The strength of the association with 24-h mortality was similar between the lethal triad and the lethal diamond, with Cramer’s V values of 0.29 and 0.28, respectively. Conclusions This study revealed no significant difference between the lethal triad and the lethal diamond in terms of their respective associations with 24-h mortality in severe trauma patients requiring transfusion. These results raise questions about the independent role of hypocalcemia in early mortality.

Background Postneonatal cerebral palsy (PNCP) is rare and requires large databases to be studied over time. Objectives To study the time trend of prevalence of PNCP overall and by cause, and to describe the clinical characteristics of children with PNCP according to cause and compared with children with pre/peri/neonatal CP (PPNCP). Methods The Surveillance of Cerebral Palsy in Europe (SCPE) database was used. Primary events (the first known chronological event in the causal chain) were classified according to the SCPE classification (six main and 19 sub‐categories). Prevalence trends for children born during 1976–2012 were modelled using multilevel generalised linear models. The clinical characteristics of PNCP and PPNCP cases born after 1998 were reported as proportions. Results The prevalence rates of PNCP were 1.76 (95% confidence interval (CI) 1.37, 2.23) and 0.82 per 10,000 live births (95% CI 0.73, 0.92) in children born during 1976–1980 and 2006–2012, respectively. The models showed a 2% annual decline in overall prevalence (prevalence rate multiplied by 0.98 each year) and a 10% decline for infectious causes for every 5‐year change. The prevalence rate in children born during 2006–2012 was 0.26 per 10,000 (95% CI 0.21, 0.32) for infectious causes, which remained the most frequent. No trend emerged for other causes. Unilateral spastic CP, associated impairments and severe gross motor dysfunction were more frequent in PNCP than in PPNCP, and PNCP showed predominantly grey matter injury (55.6%). Seventeen percent were born preterm. PNCP differed by cause, with cerebrovascular accidents presenting the least severe and hypoxic causes the most severe forms. Conclusion Our study confirms the decrease in the prevalence of PNCP in children born up to 2012, particularly for CP, due to infectious causes, which remain the most frequent. Children with PNCP had more severe presentation overall than those with PPNCP, with severity depending on the cause.

The bio‐inspired dinuclear iron(II) complex [FeII(L)]₂ (L²⁻=2,2’‐(2,2’‐bipyridine‐6,6’‐diyl)bis(1,1’‐diphenylethanethiolate)), featuring alkyl thiolate ligands, was found to react with hydrazine (N₂H₄) to form an unprecedented alkyl thiolate‐supported FeII−N₂H₄ adduct, [LFeII(N₂H₄)]. This complex was characterized structurally by X‐ray crystallography, spectroscopically (¹H NMR, UV‐vis, IR), and electrochemically. Unexpectedly, in CH3CN solution, [LFeII(N₂H₄)] gradually evolved to yield an FeII‐acetohydrazonohydrazide complex, [LFeII(H₂N₂C(CH₃)N₂H₃)], the C−N bond formation resulting from the reaction between N₂H₄ and CH₃CN mediated by the FeII ion. Its structure, confirmed by X‐ray diffraction on single crystals, reveals a η¹‐coordination of the acetohydrazonohydrazide ligand. The proposed mechanism is based on the electrophilic activation of a CH3CN molecule by FeII, followed by the nucleophilic attack of a free hydrazine. This work expands the reactivity landscape of iron‐hydrazine complexes and provides new insights into potential pathways for N−N bond functionalization in nitrogen fixation processes.

In a gold deposit near Nassara, southern Burkina Faso, gold occurs closely associated with pyrite within a network of veins hosted by metavolcanic and metasedimentary rocks. Using SEM and LA-ICP-MS analyses, we identified three generations of pyrite with distinct roles in gold mineralization. Pyrite 1 (Py1) formed early during mineralization, replacing alteration minerals like ankerite in metabasalt. Pyrite 2 (Py2) developed around Py1 in pressure shadows caused by localized micro-shear zone reactivation during successive micro-seismic events. Pyrite 2 is enriched in As and Au, unlike Py1. Pyrite 3 (Py3), unrelated to mineralization, formed at a later stage. Gold occurs in pyrite as micro-inclusions (in Py1 and Py2), fracture-fillings (mainly in Py2), and within the pyrite structure as invisible gold, including nanoparticles (predominantly in Py2). Combining electron backscatter diffraction (EBSD) and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) analysis reveals that deformation-induced misorientation of pyrite facilitated the remobilization of invisible gold, which subsequently re-precipitated as colloidal particles along sub-grain boundaries and within fractures, mimicking visible inclusions. These findings demonstrate that gold perceived as inclusions (visible or invisible) often precipitates within micro/nano-fissures and sub-grain boundaries during remobilization. This highlights the critical importance of thorough ore characterization for accurately determining gold deportment. Such insights advance our understanding of mineralization processes and support the development of more efficient recovery strategies.

The background oriented schlieren (BOS) technique provides quantitative measurements of integrated light deflection caused by gradients of refractive index in an optical medium. For an axisymmetric medium and parallel rays orthogonal to the axis of symmetry, Abel inversion enables fast, direct, and non-iterative reconstruction of the refractive index distribution. In this work, we relax the parallel ray hypothesis, demonstrating that Abel inversion can be effectively applied to observations obtained by central projection with a standard entocentric lens, with a simple correction for measured deflections. This correction derives from an original second-order approximation of ray deflection under the paraxial approximation. The accuracy of the method is demonstrated using a synthetic low-density jet and a cone of vision aperture of approximately 40°. Its practical relevance is further illustrated by reconstructing the temporal mean density fields of two experimental cases: a heated high-subsonic dual-stream jet and an under-expanded supersonic jet characterized by significant shock-induced discontinuities.