Search

Physics-based deep learning experienced a major breakthrough a few years ago with the advent of neural operators. Beyond the traditional use of deep neural networks to predict the solution to a fixed Partial Differential Equation (PDE), these novel methods are able to learn the operator solution to a class of PDEs.Comparisons and analyses of popular neural operators such as Fourier Neural Operator and DeepONet have been conducted for numerical case studies. However, they are still lacking for more realistic problems in complex settings.In this study, we compare several neural operators to predict the propagation of seismic waves in heterogeneous media. Our database is composed of more than 12 million ground motion timeseries generated from 50,000 media. We quantify the accuracy of the neural operators, their memory requirements, and their dependence towards both the initial condition and the PDE parameters. We also propose insights on their possible extension to 3 dimensions.

In the last decades, geophysicists have developed numerical simulators to predict earthquakes and other natural catastrophes. However, the more precise the model is, the higher the computational burden and the time to results. In addition, even if we could reproduce the phenomenon with more complex and more representative models, the underlying uncertainty would remain significantly high, affecting the reliability of the final prediction. In response to this challenge, we adopted a hybrid strategy, consisting into mixing physics-based numerical simulations and machine-learning. The goal is to transform synthetic earthquake ground motion, obtained via physics-based simulation, accurate up to a frequency of 5 Hz, into a broader-band prediction that mimics the recorded seismographs. In doing so, we factorize the latent representation of the seismic signal, by forcing an encoding that splits features into two parts: a low frequency one (0-1 Hz) and a high frequency one (1-20 Hz). In the following, we train a convolutional U-Net neural network and apply two different signal-to-signal translation techniques: pix2pix and BiCycleGAN. The latter strategies are compared with the prior work of Gatti et al., 2020, on the Stanford Earthquake Dataset (STEAD) showing their capability of mimicking recorded seismographs. We finally tested the two strategies on the synthetic time-histories obtained for the 2019 Le Teil earthquake (France).

Hodgkin lymphoma (HL) is today one of the most curable pediatric cancers. Despite survival rates now exceeding 90%, survivors of pediatric HL are still at higher risk to develop late effects of cancer therapy. Premature aging has been proposed as a paradigm to explain the onset of long-term complications in these subjects. High levels of advanced glycation end products (AGEs), together with chronic inflammation and oxidative unbalance, have been shown to be among the main factors contributing to aging. The present study aims to evaluate glycoxydation, inflammatory status, and oxidative stress in plasma and peripheral blood mononuclear cells (PBMC) obtained from 20 adult survivors of pediatric HL and 40 age- and sex-matched healthy controls. After the isolation of PBMC and the collection of plasma, we performed the analyses of gene expression by qRT-PCR and measured inflammatory and oxidative-stress markers. AGEs plasma levels, expressed as Nϵ-carboxymethyl-lysine and methylglyoxal hydroimidazolone, were markedly higher in HL survivors than in healthy subjects. HL survivors also showed a condition of higher oxidative stress, as demonstrated by an increased expression of NADPH oxidase on PBMC. Antioxidant defenses, evaluated in terms of alpha-tocopherol, GSSG/GSH ratio and catalase plasma levels, were strongly impaired in survivors. This pro-oxidative condition led to the over-expression of both NLRP3 and NFkB genes in PBMC and, consequently, to increased plasma levels of interleukin(IL)-1β and IL-6. Finally, the expression of the receptors for AGEs in PBMC confirmed the dysregulated AGE pathways. Data show AGEs accumulation in survivors of pediatric HL. The consequent activation of the receptor for AGEs leads to the persistent activation of intracellular signaling toward inflammation. These results suggest that the co-existence of AGEs accumulation, unbalanced oxidative status, and inflammation could play a role in the onset of late complications in HL survivors.

International audience

The prediction of the seismic response of critical structures is highly sensitive to many aspects, among which the earthquake source and the geological setting are prominent. The related uncertainty issues must be taken into account in seismic risk mitigation studies, for example through the evaluation of several realizations of a future earthquake scenario. This aspect is crucial when addressing vulnerability studies at a regional scale. When opting for physical-based numerical simulations (PBS), however, the computational burden increases along with the expected degree of fidelity, making it difficult to evaluate more than a few dozens of alternatives. To cope with this disadvantage, in this work an alternative method is proposed, which exploits a rather low number of synthetic earthquake simulations and combines them with the empirical Green function (EGF) method, to finally generate thousands of alternative yet realistic seismic response of the site of interest. This hybrid strong motion predictions benefit of both (i) PBS high fidelity and (ii) data assimilation of strong ground motion records in the seismic area of interest, via EGF method, producing broadband synthetics at a relative cheap computational price. The power of the hybrid method is tested on a real case scenario, embodied by the ground-shaking prediction at the nuclear site of Cadarache, in the surroundings of the fault of Middle Durance, in South-Eastern France. Thousands of broadband (0–15 Hz) hybrid synthetic seismic response are generated, associated with different fault parameters (EGF method) and based upon a few key physics-based simulations, accurate up to 5 Hz.

The seismic risk in France, a region of low to moderate seismicity, is of paramount importance given the large number of industrial and nuclear installations. However, the large uncertainties on the geology and the poor knowledge of active faults make the seismic hazard estimation a challenging task. Despite being a promising tool to explore the underlying uncertainties, numerical simulations must be duly calibrated by reproducing specific events. In this work, we considered the 2019 Mw4.9 earthquake that occurred at Le Teil village in southern France. This event was recorded by 17 stations of 3-component accelerometers, within an area of 50 km around the epicenter (French Accelerometric Network). We used these records to calibrate the numerical simulation. The seismological P- and S-wave speed profiles used result from a 3D weighted average model for Metropolitan France. In addition, the topography was included in the spatial discretization. The uncertainties on dip, strike, and rake angles were explored in order to calibrate the far-field synthetic ground motion model by determining the eigenquakes that efficiently span a large diversity of sources.A good agreement between synthetic and recorded time histories was found, despite the simplicity of the geological and source model.

Hirata’s disease (HD) is a rare autoimmune cause of hypoglycemia. Patients suffering from this condition have a genetic predisposition, determined by HLA DR4, with some differences in the worldwide population. In Caucasians HLA DRB1*0403 is the most frequent susceptibility background on which some drugs play as triggers. We reported the case of a woman with several hypoglycemic episodes, characterized by high insulin and c-peptide levels. Biochemical and morphological exams excluded a neuroendocrine tumor. HD was diagnosed according to insulin autoantibodies positivity and patient’s history, particularly about drugs taken. The HLA analysis revealed DRB1*0415 allele. We found a potential new predisposing factor for HD, HLA DRB1*0415 allele, never described before as genetic background to insulin autoimmune syndrome in Caucasians.

International audience; Civil structures, infrastructures and lifelines are constantly threatened by natural 1 hazards and climate change. Structural Health Monitoring (SHM) has therefore become an 2 active field of research in view of online structural damage detection and long term maintenance 3 planning. In this work we propose a new SHM approach leveraging a deep Generative Adversarial 4 Network (GAN), trained on synthetic time histories representing the structural responses of both 5 damaged and undamaged multistory building to earthquake ground motion. In the prediction 6 phase, the GAN generates plausible signals for different damage states, based only on undamaged 7 recorded or simulated structural responses, thus without the need to rely upon real recordings 8 linked to damaged conditions.

“Mappa Italia” is a WebGIS service, issued by ANIA Servizi (Associazione Nazionale fra le Imprese Assicuratrici), conceived to support the business activity of insurance companies. It returns a “Risk Index” for the entire Italian territory by analyzing national institutional geo-hazard maps and the physical exposure of the assets located in each terrain unit used in this study. The analysis has been performed by processing data from census tracts, land use map and flood and landslide hazard maps. Census tracts provides data about population, buildings (residential, commercial and industrial) and firms (number of companies and employees). Data about land use (urban and rural areas), have been extracted from the Corine LandCover. Data about flood and landslide hazard have been obtained from institutional maps published by ISPRA (Italian Institute for Environmental Protection and Research). The assets have been classified into four classes (D1, D2, D3, D4) considering their type and amount in each terrain unit using a k-means clustering technique.Then, four different risk matrixes were applied to get a Risk Index one for each asset class (D1, D2, D3, D4), by plotting the classified asset classes against the probability of occurrence of a given flood/landslide event. Then, a total Risk Index was derived for each terrain unit, by summing up the Risk Index of each asset class.Insurance companies may access results by a WebApp and a Web Service by authentication Service providing login and password.This work has to be considered a starting point targeted to assess and make available information on exposure, hazard and of the vulnerable elements.

Cancer therapy-induced bone loss (CTIBL), occurring especially in hormone-treated breast and prostate cancer patients, is a noteworthy long-term consequence of cancer treatments. Because of its negative impact on the quality life of cancer survivors, it deserves much attention. We here summarize the pathophysiology of CTIBL in breast and prostate cancer, its clinical presentation, management, and treatment.

Purpose Postoperative assessment of acromegaly activity is typically performed at least 3 months after neurosurgery (NS). Few studies have evaluated the use of early postoperative growth hormone (GH) levels as a test to predict short- and long-term remission of acromegaly. Our objective was to evaluate the diagnostic performance of serum random GH on a postoperative day one (D1-rGH) and two (D2-rGH), particularly in predicting long-term disease persistence. Materials and methods Forty-one subjects with acromegaly who were undergoing NS were enrolled (mean age ± SD 47.4 ± 13.1 years at diagnosis; women 54%; macroadenomas 71%). The final assessment of disease activity was performed one year after NS. ROC curves were used to evaluate the diagnostic performance of D1-rGH and D2-rGH. Results After a 1-year follow-up, the overall remission rate was 55%. ROC analysis identified an optimal D1-rGH cut-off value of 2.1 ng/mL for diagnosing long-term disease persistence (55.6% SE; 90.9% SP). The cut-off point became 2.5 ng/mL after maximizing specificity for disease persistence (yielding a 100% positive predictive value) and 0.3 ng/mL after maximizing sensitivity for disease remission. The optimal D2-rGH cut-off value was 0.6 ng/mL (81.8% SE; 50% SP); the cut-off point became 2.9 ng/mL after maximizing specificity and 0.1 ng/mL after maximizing sensitivity, with no clinical utility. Conclusions D1-rGH could be a highly specific test for the early diagnosis of long-term acromegaly persistence, which is predicted by a value > 2.5 ng/mL with a great degree of certainty. The diagnostic performance of D2-rGH was insufficient. Further research is required to validate these preliminary results prior to modifying the postoperative management of acromegaly.

Cancer therapy-induced bone loss (CTIBL), occurring especially in hormone-treated breast and prostate cancer patients, is a noteworthy long-term consequence of cancer treatments. Because of its negative impact on the quality life of cancer survivors, it deserves much attention. We here summarize the pathophysiology of CTIBL in breast and prostate cancer, its clinical presentation, management, and treatment.

This study investigates the sensitivity of physics-based earthquake prediction to the choice of the regional geological model. The generally poor information on the mechanical properties of the Earth’s crust induces large uncertainty margins on the synthetic seismic response. Therefore, uncertainty quantification must be associated with high-fidelity numerical simulation, especially when verging on broad-band simulations (i.e., numerical theoretical accuracy higher than the conventional 1 Hz limit). In this paper, the synthetic seismic response of the Niigata region (Japan), during the Ch u etsu offshore earthquake (MW6.6) serves as representative case study. Three plausible three-dimensional (3-D) geological structures are compared for two point-wise aftershock numerical simulations, duly validated. The framework of comparison targets the Ground Motion Prediction Equation (GMPE) for pseudo-spectral acceleration at the surface. The impact of the geology on the variability of between-event and the within-event residuals is inferred from synthetic simulations and related to common parameters representing site-specific conditions.

Introduction: Central diabetes insipidus (CDI) is a frequent complication of pituitary surgery, but its diagnosis lacks standardized criteria. Copeptin, a surrogate marker of arginine vasopressin release, is triggered by psycho-physical stresses such as pituitary surgery. Low postoperative copeptin could predict CDI onset. The aims of this study were the validation of copeptin as a predictor of post-neurosurgical CDI and the identification of the optimal timing for its determination. Methods: Sixty-six consecutive patients operated for a hypothalamic-pituitary lesion were evaluated. Copeptin was determined preoperatively and at 1, 6, 12, 24 and 48 h post-extubation. Fifty-eight patients were reassessed after 3–6 months post-surgery to confirm transient (3 cases) or permanent CDI (5 cases) diagnosis. Results: A marked copeptin peak was identified at 1 h after extubation, when a value below or equal to 12.8 pmol/L had a good accuracy in identifying CDI cases (AUC 0.866, 95% CI 0.751–0.941). Moreover, a copeptin peak above 4.2 pmol/L excluded permanent forms (AUC 1, 95% CI 0.629–1). Regression analysis identified copeptin as the only significant predictor of CDI (OR 0.86, 95% CI 0.75–0.98, p = 0.02). A copeptin T1/T0 ratio below or equal to 1.47 identified patients at risk of isolated biochemical alterations even in the absence of an overt CDI. Conclusions: A prompt increase of copeptin is expected at 1 h after extubation. The absence of this peak is a reliable predictor of post-neurosurgical CDI.

This paper presents a new strategy to blend the outcome of physics-based numerical simulations with massive but poorly-labelled experimental databases such as in-situ data routinely recorded for monitoring purposes. The proposed approach relies on a set of adversarial learning techniques with a twofold purpose: (1) finding two reduced-dimensional non-linear representations of both synthetic and experimental data; (2) training a stochastic generator of fake experimental responses conditioned by the physics-based simulation results. This methodology is applied to earthquake ground motion prediction. Indeed, regional three-dimensional high-fidelity numerical models accounting for both extended sources and complex geology are still limited to a low-frequency range. Moreover, they are prone to significant uncertainties induced by a lack of data on small scale geological structures and rupture processes. Databases of broadband seismic signals recorded worldwide at seismological networks are used to retrieve some pieces of information on these small scale data to generate realistic broadband signals from synthetic ones. Outstanding performances in encoding seismic signals are demonstrated, together with efficient generation capabilities, provided that the physics-based results carry enough information to properly condition the stochastic generator. In addition, this paper shows that the proposed method, fed only with raw data from both databases and numerical models, outperforms other random signal generators based on pre-existing expertise such as prescribed spectra and more or less complex phenomenological models.

In this article, a novel strategy to generate broadband earthquake ground motions from the results of 3D physics-based numerical simulations (PBSs) is presented. Physics-based simulated ground motions embody a rigorous seismic-wave propagation model (i.e., including source, path, and site effects), which is however reliable only in the long-period range (typically above 0.75–1 s), owing to the limitations posed both by computational constraints and by insufficient knowledge of the medium at short wavelengths. To cope with these limitations, the proposed approach makes use of Artificial Neural Networks (ANNs), trained on a set of strong-motion records, to predict the response spectral ordinates at short periods. The essence of the procedure is, first, to use the trained ANN to estimate the short-period response spectral ordinates using as input the long-period ones obtained by the PBS, and, then, to enrich the PBS time histories at short periods by scaling iteratively their Fourier spectrum, with no phase change, until their response spectrum matches the ANN target spectrum. After several validation checks of the accuracy of the ANN predictions, the case study of the 29 May 2012 Mw 6.0 Po Plain earthquake is illustrated as a comprehensive example of application of the proposed procedure. The capability of the proposed approach to reproduce in a realistic way the engineering features of earthquake ground motion, including the peak values and their spatial correlation structure, is successfully proven.

International audience; The scope of this paper is to give an insight into the advantages of a new, all-embracing, modeling approach of a strong ground motion scenario, by carrying out a source-to-structure analysis at regional scale, accounting explicitly for the uncertainties related to the databases and the models. To this end, a suitable case-study is represented by the 2007 Mw6.6 Niigata-Ken Ch¯ uetsu-Oki seismic sequence (west Japan), that damaged the Kashiwazaki Kariwa Nuclear Power Plant. This study describes the effect of the wave propagation path within the Earth's crust on the seismic response of nuclear reactor buildings located nearby a seismogenic source. The multiscale problem is de-coupled into three steps: (1) a parallel simulation of seismic-wave propagation throughout the Earth's crust at regional scale (≈ 60 km wide, major 3-D geological interfaces found below the nuclear site), reliable up to 5.0 Hz; (2) a mid hybridization step consisting in enriching the synthetic wave-field at high frequency (up to 30 Hz), employing an Artificial Neural Network to predict the short-period (SP) spectral ordinates; (3) a high-resolution structural dynamic analysis, introducing the hybrid broad-band synthetics as input wave-motion. A simplified stress-test is performed, ✩ Fully documented templates are available in the elsarticle package on CTAN. by simulating two small point-wise aftershocks at different source-site position. The impact of the underground 3-D geology on the structural components is finally quantified, by injecting the obtained broad-band time-histories in a Soil-Structure Interaction (SSI) model of the nuclear reactor building. The good fit obtained in terms of amplification factor at different recording stations assures the high-fidelity of the holistic philosophy endorsed.

International audience; The aim of this paper is to clarify the extent of near-source effects and non-linear soil response that took place at the Japanese nuclear site of Kashiwazaki-Kariwa,during the MW 6.6 Niigata-Ken Chūetsu-Oki earthquake (July, 16th 2007). The interest ofthis case study stems from the relative small source-to-site distance and shallowhypocenter depth, coupled with the consistent seismic record database available from twodown-hole arrays of strong motion accelerometers, installed within the nuclear site.Records were processed first to highlight those features related to near-source conditions,such as directivity and polarization, and, subsequently, to identify dependence of siteresponse on the direction of motion and on the non-linear soil behaviour. Moreover,borehole interferometry was used to check the two available velocity profiles (estimated byin situ PS logging) used for 1D linear equivalent numerical simulations. In one case, animproved agreement with the empirical amplification functions was obtained using asmoothed profile. The main non-linear features highlighted by signal processing wereglobally reproduced by means of linear-equivalent analyses, that pointed out the stronginfluence of the input motion direction. Among the principal results of this research, it wasfound that the hanging-wall and directivity effects along with the non-linear site-effectsdominate the ground response during the main shock, and that the site amplificationfeatures are strongly dependent on the rotation angle of the horizontal components.

International audience

International audience; The physics-based numerical simulation is a reliable tool to build-up source-to-site strong ground motion earthquake scenarios. In this paper, we enhance those forward simulations by including (1) the hysteretic non-linear behaviour of geomaterials and (2) the heterogeneous nature of soils. Scattering-induced damping, ground motion spatial incoherence and de-amplification may therefore be reproduced. A classical case study is considered herein: a seismic wave generated by a localized double-couple earthquake propagating towards the surface through layered elastic homogeneous bedrock, underlying a heterogeneous non-linear soil stratum. 3D numerical simulations were performed by means of a Spectral Element Method (SEM) code. The latter was coupled with a highly scalable random field generator to reproduce the media heterogeneity. Moreover, an extension to J 2 elasto-plasticity numerical integration was introduced. Time-histories and lagged-coherency curves at the surface show the ground motion de-amplification (mainly related to soil non-linearity) and loss of coherency (due to the wave-scattering caused by soil heterogeneity).

Minisymposium: The SINAPS@ French research project on integrated seismic risk assessment for nuclear safety; International audience; This study presents a preliminary calibration of a large-scale seismological model of the M W 6.6 Niigata-Ken Chuetsu-Oki (Japan) earthquake of July, 16 th 2007. The strong ground motion affected a wide area in the surrounding of the Kashiwazaki-Kariwa Nuclear Power Plant (KKNPP). Due to the relative small source-to-site distance and shallow hypocenter depth, this seismic scenario resulted very interesting and well documented (a consistent database of seismic recordings is available). In this context, the test-case represents a suitable benchmark for the work-packages of the SINAPS@ project. SINAPS@ is the first French research project with the objective to quantify the uncertainty of the procedures to estimate the seismic risk. In this context, an omni-comprehensive approach is followed, modelling the wave-propagation from the fault to the structural components. This study is intended to describe the steps to build up and calibrate a reliable seismic scenario, capable to provide a synthetic wave-field at the regional scale. This objective is pursued by (1) assessing a stratified geological model for the Niigata region, (2) testing the effect of the source parameters in a kinematic approach (e.g. the rise time and the shape of the Source Time Function) and (3) checking the topography effect. A series of 3D source-to-site numerical simulations of the M W 4.4 NCOEQ2007 aftershock was therefore carried out with this regard (performed by means of the Spectral Element Method, SEM). The seismic scenario was calibrated with the aid of the semi-analytical solution provided by the Wave-Number Integration Method (WNI) and by comparing the synthetics with the recordings at several locations (i.e. the KNET-Kik-Net stations nearby and the KKNPP site). Finally, the forward physics-based analysis ended up by providing broad-band synthetic time-histories to be exploited in further studies as a reliable regional incident wave-field for the engineering bedrock

The paper presents a numerical approach for the construction of seismic fragility curves for shallow metro tunnels considering the soil-structure-interaction (SSI) and the aging effects due to corrosion of the lining reinforcement. The tunnel response under ground shaking is evaluated through 2D non-linear dynamic analyses, for increasing levels of seismic intensity. An elasto-plastic model is used to simulate the soil non-linear behavior under ground shaking, while the effects of lining mechanical properties, soil conditions and ground motion characteristics are also accounted for. The effect of corrosion on the lining behavior is encountered through proper modification of the lining strength properties. Damage state thresholds are defined based on the exceedance of the lining capacity. The fragility curves are estimated in terms of peak ground acceleration at the ground free field conditions for different time periods considering the associated uncertainties. The proposed approach is applied for the fragility assessment of selected soil-tunnel configurations. The derived fragility functions are compared with existing empirical and analytical fragility models, highlighting the important role of soil conditions and corrosion effects in the vulnerability of tunnel structures. The proposed fragility models contribute towards an advanced vulnerability and risk assessment of transportation systems and infrastructures.

Introduction: Multiple studies tried to identify cortisol cut-offs after pituitary surgery that could accurately assess hypothalamic-pituitary-adrenal (HPA) axis function; however, there is no consensus nowadays. This study aimed to evaluate the accuracy of morning cortisol after transsphenoidal surgery in predicting long-term secondary adrenal insufficiency. Methods: In our tertiary center, we prospectively determined first- and second-day cortisol after transsphenoidal surgery in 92 patients without preoperative adrenal ­insufficiency and not treated with glucocorticoids perioperatively. Definitive diagnosis of secondary adrenal insufficiency was obtained with re-evaluation 3 months after trans­sphenoidal surgery and clinical follow-up of at least 1 year. Results: Ten patients (10.8%) developed long-term postoperative secondary adrenal insufficiency. The ROC curves demonstrated that first-day cortisol had a moderate diagnostic accuracy, while a second-day cortisol ≤9.3 µg/dL (257 nmol/L) showed the best performance in predicting adrenal insufficiency (sensitivity [Se] 88.9%, specificity [Sp] 86.9%, AUC 0.921). Moreover, a second-day cortisol ≤3.2 µg/dL (89 nmol/L) was able to diagnose adrenal insufficiency in 100% of cases (Se 22.2%, Sp 100%) and >14 µg/dL (386 nmol/L) was able to exclude ACTH deficiency (Se 100%, Sp 57.4%). Conclusions: Adrenal function can be carefully studied on the second day after pituitary surgery, using cut-off values that international guidelines suggested for non-stressed conditions. In fact, second-day cortisol levels ≤3.2 μg/dL (89 nmol/L) and >14 μg/dL (386 nmol/L) are diagnostic of secondary adrenal insufficiency and normal function, respectively. We also suggest performing a definitive re-evaluation with an HPA axis stimulation test when second-day cortisol values are between 3.3 and 14 μg/dL (90–386 nmol/L).