Browse Articles

Enhanced prediction of ventilator-associated pneumonia in patients with traumatic brain injury using advanced machine learning techniques

Abstract Ventilator-associated pneumonia significantly increases morbidity, mortality, and healthcare costs among patients with traumatic brain injury. Accurately predicting risk can facilitate earlier interventions and improve patient outcomes. This study leveraged the MIMIC III database, identifying traumatic brain injury cases through standardized clinical criteria. A rigorous data preprocessing workflow included missing value imputation, correlation checks, and expert-driven feature selection, reducing an initial set of features to a subset of critical predictors encompassing demographics, comorbidities, laboratory values, and clinical interventions. To address class imbalance, the Synthetic Minority Oversampling Technique (SMOTE) was applied within a five-fold cross-validation framework, ensuring a balanced training set while maintaining an unbiased validation process. Six machine learning models, including Support Vector Machine, Logistic Regression, Random Forest, XGBoost, Artificial Neural Network, and AdaBoost, were trained using extensive hyperparameter tuning. Comprehensive evaluations were conducted based on multiple metrics, including Area Under the Curve (AUC), accuracy, F1 score, sensitivity, specificity, Positive Predictive Value, and Negative Predictive Value. XGBoost emerged as the top performing algorithm, achieving an AUC of 0.94 and an accuracy of 0.875 on the test set, marking substantial improvements over previously reported best results. An ablation study validated the necessity of each retained feature, indicating that any feature removal led to a decline in model performance. Furthermore, SHAP analysis underscored ICU length of stay, hospital length of stay, serum potassium, and blood urea nitrogen as key contributors to ventilator associated pneumonia risk. Overall, the results demonstrate that advanced ensemble learning, meticulous feature selection, and effective class imbalance handling can significantly enhance early detection in traumatic brain injury cases. These findings have meaningful clinical implications, offering a framework for more timely interventions, optimized resource allocation, and improved patient care in critical settings.

Oxygen level alters energy metabolism in bovine preimplantation embryos

Abstract Mammalian preimplantation embryo development is a complex sequence of events. This period of development is sensitive to oxygen (O 2 ) levels that can affect various cellular processes. We compared the influence of O 2 tension by culturing embryos either in normoxic (20% O 2 ) or physiological hypoxic (6% O 2 ) conditions, or sequential low O 2 concentration starting with 6% O 2 until 16-cell stage and then switching to ultrahypoxic conditions (2% O 2 ). Due to ethical concerns, we used bovine as an animal model with a good similarity of embryogenesis to human. We found that the cleavage rate was not affected by O 2 levels but there was a clear difference in blastocyst formation rate. In hypoxia, 36% of embryos reached blastocyst stage while in normoxia only 13%. In ultrahypoxia conditions only 4.6% of embryos developed up to blastocyst stage. Transcriptomic profiles showed that normoxic conditions slowed down oocyte transcript degradation which is a prerequisite for reprogramming of the embryonic cell lineages. There were also clear differences in the expression of key metabolic enzymes between hypoxic and normoxic conditions at the blastocyst stage. Both hypoxic and ultrahypoxic conditions seemed to induce appropriate energy production by upregulating genes involved in glycolysis and lipid metabolism typical to in vivo embryos. In contrast, normoxic conditions failed to upregulate glycolysis genes and only depended on oxidative phosphorylation metabolism. We conclude that constant hypoxia culture of in vitro embryos provided the highest blastocyst formation rate and appropriate energy metabolism. Normoxia altered the energy metabolism and decreased the blastocyst formation rate. Even though ultrahypoxia at blastocyst stage resulted in the lowest blastocyst formation, the transcriptional profile of surviving embryos was normal.

Optimizing the multi-model ensemble of CMIP6 GCMs for climate simulation over Bangladesh

Central retinal artery catheterization for retinal artery occlusion with balanced salt solution

Interpretable machine learning model for early prediction of disseminated intravascular coagulation in critically ill children

Plasticulture detection at the country scale by combining multispectral and SAR satellite data

Abstract The use of plastic films has been growing in agriculture, benefiting consumers and producers. However, concerns have been raised about the environmental impact of plastic film use, with mulching films posing a greater threat than greenhouse films. This calls for large-scale monitoring of different plastic film uses. We used cloud computing, freely available optical and radar satellite images, and machine learning to map plastic-mulched farmland (PMF) and plastic cover above vegetation (PCV) (e.g., greenhouse, tunnel) across Germany. The algorithm detected 103 103 ha of PMF and 37 103 ha of PCV in 2020, while a combination of agricultural statistics and surveys estimated a smaller plasticulture cover of around 100 103 ha in 2019. Based on ground observations, the overall accuracy of the classification is 85.3%. Optical and radar features had similar importance scores, and a distinct backscatter of PCV was related to metal frames underneath the plastic films. Overall, the algorithm achieved great results in the distinction between PCV and PMF. This study maps different plastic film uses at a country scale for the first time and sheds light on the high potential of freely available satellite data for continental monitoring.

Orbital angular momentum detection of vortex beams by #-type lines

Comparative study on bivariate statistical characteristics of drought in Shandong using SPI and SPEI

Analysis of combining ability and heterosis based on controlled pollination populations of eucalypt

A fiber optic approach for cement placement and hydration assessment of deep geothermal boreholes

Abstract Achieving well integrity is mandatory for a geothermal well’s safe and sustainable operation. One of the most critical steps is the success of the primary cementing. Conventional monitoring only shows discrete snapshots after completion of the cement job. However, optical fiber sensors enable monitoring of the entire cementing process. Here, we investigate the cement placement and early hydration for a surface casing at a geothermal site in Munich, Germany. We show that distributed dynamic strain rate sensing (DDSS or DAS) allows for tracking rising fluid interfaces, determining the setting time of cement, and assessing the cement job’s success at each depth. We used DDSS and DTS (distributed temperature sensing) with a fiber optic cable permanently deployed behind the casing and combined the results with operational data, a model for the rise of fluids in the borehole, and laboratory experiments to estimate the cement setting phase. Our approach enables monitoring all phases of primary cementing, which can increase the success rate of achieving well integrity. Furthermore, it can reduce costs and improve society’s acceptance of deep geothermal wells in urban areas.

Evaluating regulatory influences on coal mine accidents in China using a fuzzy multi-criteria decision-making approach

Prediction of $$\pi$$-electronic energy and physical properties of benzenoid hydrocarbons using domination degree based entropies

Abstract This study introduces a novel approach to calculating graph entropies using topological indices, inspired by Shannon’s entropy concept. These entropies, as information-theoretic measures, are applied to evaluate the structural properties of chemical graphs. Graph theory is utilized to examine correlations between specific chemical properties and graph entropy measures. Within this framework, several physicochemical and quantum properties, including boiling point, enthalpy, molecular weight, and $$\pi$$ -electronic energy are analyzed. Certain new graph entropy measures, termed domination entropies, are introduced based on domination topological indices and computed for 29 benzenoid hydrocarbons. Additionally, a QSPR analysis is conducted to investigate the linear and multilinear relationships between these entropies and the physicochemical properties, as well as the $$\pi$$ -electronic energy of the hydrocarbons. The predictive accuracy of these new domination entropies is confirmed through various statistical tools.

Latent class analysis identifies risk groups to model the expected benefits of SARS-CoV-2 interventions among university students

Tracing the heliospheric magnetic field via anisotropic radio-wave scattering

Abstract Astrophysical radio sources are embedded in turbulent magnetised environments. In the 1 MHz sky, solar radio bursts are the brightest sources, produced by electrons travelling along magnetic field lines from the Sun through the heliosphere. We demonstrate that the magnetic field not only guides the emitting electrons, but also directs radio waves via anisotropic scattering from density irregularities in the magnetised plasma. Using multi-vantage-point type III solar radio burst observations and anisotropic radio wave propagation simulations, we show that the interplanetary field structure is encoded in the observed radio emission directivity, and that large-scale turbulent channelling of radio waves is present over large distances, even for relatively weak anisotropy in the embedded density fluctuations. Tracing the radio emission at many frequencies (distances), the effects of anisotropic scattering can be disentangled from the electron motion along the interplanetary magnetic field, and the emission source locations are unveiled. Our analysis suggests that magnetic field structures within turbulent media could be reconstructed using radio observations and is found consistent with the Parker field, offering a novel method for remotely diagnosing the large-scale field structure in the heliosphere and other astrophysical plasmas.

Genetic landscape and phenotypic spectrum of osteogenesis imperfecta in the Kazakhstani pediatric population

A functional digital model of the Dingo thermal neutron imaging beamline

Impact of age, sex, body constitution, and the COVID-19 pandemic on the physical fitness of 38,084 German primary school children

Abstract Physical fitness (PF) is a vital health indicator, but many children do not meet the WHO physical activity guidelines. Low PF in children raises the risk of non-communicable diseases (NCDs) and negatively impacts their quality of life, a situation amplified during the COVID-19 pandemic. This study uses data from 38,084 German third-graders (7.8 to 9.4 years) across seven cohorts (school years 2017/18–2023/24) who participated in a statewide fitness evaluation program in the Federal State of Thuringia. We aimed to examine age, sex, and pandemic effects on endurance (6-minute run), coordination (star run), speed (20-meter linear sprint), lower limb power (powerLOW; standing long jump), upper limb power (powerUP; ball-push test), and static balance (one-legged stance with eyes closed). Data on height and mass was incorporated to clarify the task-specific impact of the body mass index (BMI) on performance. Our results corroborated reported age and sex effects, demonstrating linear development across the age range for all tests, and in static balance only, girls scored higher than boys. The relationship between BMI and task-specific performance revealed an inverted U-shaped function in weight-bearing tests (first four). Functions were steeper for boys than girls, implying a stronger BMI impact on performance for boys. BMI and age interaction suggest that being overweight may limit age-related performance benefits. Negative pandemic declines in endurance, coordination, powerLOW, and powerUP were more pronounced in “fitter” schools. In conclusion, task-specific performances should be interpreted considering age, sex, and body constitution. Pandemic-related performance declines emphasize the role of access to physical activity resources for all children.

Tokay geckos adjust their behaviour based on handler familiarity but according to context

Content-based image retrieval assists radiologists in diagnosing eye and orbital mass lesions in MRI

Abstract Diagnosing eye and orbit pathologies through radiological imaging presents considerable challenges due to their low prevalence, the extensive range of possible conditions, and their variable presentations, necessitating substantial domain-specific expertise. This study evaluates whether a ML-based content-based image retrieval (CBIR) tool, combined with a curated database of orbital MRI cases with verified diagnoses, can enhance diagnostic accuracy and reduce reading time for radiologists diagnosing eye and orbital pathologies. It explores whether this tool alone, or in combination with status quo reference tools (e.g. Radiopaedia.org, StatDx) provides these benefits. In a multi-reader, multi-case study involving 36 radiologists and 48 retrospective orbital MRI cases, participants diagnosed eight cases: four using status quo reference tools and four with the addition of the CBIR tool. Analysis using linear mixed-effects models revealed significant improvements in diagnostic accuracy when using the CBIR tool alone (55.88% vs. 70.59%, p  = 0.03, odds ratio = 2.07) and an even greater improvement when used alongside status quo tools (55.88% vs. 83.33%, p  = 0.02, odds ratio = 3.65). Reading time decreased when using the CBIR tool alone (334 s vs. 236 s, p  < 0.001) but increased when used in conjunction with status quo tools (334 s vs. 396 s, p  < 0.001). These findings indicate that CBIR tools can significantly enhance diagnostic accuracy for eye and orbit diagnostics, though their impact on reading time varies.

Characterization and optimization of azo dyes degrading microbes isolated from textile effluent