Browse Articles

Censoring-adjusted tree-based policy learning for estimating dynamic treatment regimes with censored outcomes

Disruption of microtubules with low intensity ultrasound rescues hair follicle damage by paclitaxel in mouse models

Enhancing medical Q&A systems with multimodal knowledge graphs and dual-layer attention mechanisms

Medical intelligent question-answering (QA) systems have become important tools for improving the efficiency of healthcare services, and recent research has increasingly emphasized performance optimization and multimodal integration. However, existing systems still face several challenges in intent recognition, entity extraction, and multimodal knowledge fusion, particularly reduced accuracy in multi-label classification, heavy reliance on large-scale annotated data, and limited support for cross-modal retrieval. To address these issues, this study proposes a medical intelligent QA framework that integrates a dual-layer attention mechanism, a large language model, and a multimodal medical knowledge graph to improve system understanding and response generation in complex clinical scenarios. Specifically, we develop a text-based intent recognition model with a dual-layer attention architecture, in which a global contextual attention module is introduced to capture long-range semantic dependencies and improve multi-label classification performance. In addition, an instruction-tuned large language model is employed for zero-shot medical entity recognition, thereby reducing dependence on manually annotated datasets. Building on this foundation, we construct a multimodal medical knowledge graph comprising more than 15,000 associated medical images and develop a visualization-oriented retrieval interface using Flask and ECharts. Experimental results show that the proposed intent recognition model achieves a peak Micro-F1 of 94.42% on multiple benchmark datasets, outperforming several baseline methods. The LLM-based entity recognition module achieved competitive recall in medical entity extraction, demonstrating strong capability in identifying medical entities. User evaluation results further indicate that the system is effective and practical across a variety of medical query types. This study provides a feasible framework for advancing medical QA systems through improved intent recognition, low-resource entity extraction, and multimodal knowledge integration.

Entrepreneurship education shapes creativity among Egyptian hospitality students through entrepreneurial attitudes and intentions

Abstract This study developed and tested a psychological model explaining associations between entrepreneurship education (EE) and student creativity and entrepreneurial behavior, drawing on the theory of planned behavior (TPB) and the componential theory of creativity (CTC). Using a cross-sectional survey of 335 hospitality students and analyzing the data with partial least squares structural equation modeling (PLS-SEM), the study examined the relationships of EE with individual creativity, university-level creativity, entrepreneurial attitudes, and entrepreneurial intentions. The results show that EE is positively associated with students’ entrepreneurial intentions, attitudes, and individual creativity. Entrepreneurial attitudes and intentions appear to act as key psychological mediators in several pathways, suggesting an interconnected nature of creativity and entrepreneurial behavior. However, entrepreneurial intentions do not mediate the relationship between EE and university-level creativity. The findings provide contextual support and possible extensions to TPB and CTC by highlighting how educational interventions relate to both creative capacity and entrepreneurial tendencies in resource-constrained higher education contexts. Practically, the study provides preliminary guidance for enhancing entrepreneurship curricula and designing supportive ecosystems that promote creativity, experiential learning, mentorship, and entrepreneurial behavior among hospitality students.

Harnessing local chemical order in high-entropy ceramics for broadband electromagnetic absorption

Evaluation of plasma neurofilament light chain and glial fibrillary acidic protein in myasthenia gravis: A controlled cohort study

Aims To evaluate plasma neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) as candidate biomarkers in myasthenia gravis (MG). Methods Ninety MG patients and 40 healthy controls were recruited. Disease severity was assessed by the Myasthenia Gravis Foundation of America (MGFA) classification, Myasthenia Gravis Composite (MGC) score, and Myasthenia Gravis Activities of Daily Living (MG-ADL) scale. Plasma NfL and GFAP were quantified using Single Molecule Array (Simoa) assays. Results NfL and GFAP plasma concentration did not differ between MG and controls ( p  > 0.05). Neither biomarker correlated with MG-ADL or MGC, and no differences were observed across MGFA classes (p > 0.05). Biomarker levels were unrelated to myasthenic crisis history or treatment exposure. Conclusion Plasma NfL and GFAP, although informative in other neuroimmunological and neurodegenerative conditions, do not distinguish MG from healthy controls and show no association with disease severity. This study adds to the emerging literature on NfL in MG and represents one of the larger controlled analyses incorporating both NfL and GFAP biomarkers in this disease. The findings argue against adopting NfL or GFAP for MG monitoring and highlight the need for MG-specific biomarker strategies.

Physiological and behavioral response to weaning stress in primiparous and multiparous ewes

Mechanically locking enzymes in covalent organic frameworks via light-responsive nanohands for stable biocatalysis

Correction: Salivary biomarkers of tactical athlete readiness: A systematic review

Antimicrobial susceptibility profile of environmental Legionella pneumophila isolates from shower water in Shandong Province, China

Formation of iron oxide-apatite deposits triggered by magmatic assimilation of evaporitic sulfate

Abstract The geological origins of iron oxide-apatite (IOA) rocks, important resources for iron and rare-earth elements, are intensely debated. Using triple oxygen isotope data, we here show that magnetite from IOA deposits near Kiruna, northern Sweden, and related igneous rocks contain high concentrations of oxygen derived from evaporitic sulfate. To explain these observations, we propose that the Kiruna IOA assemblage formed in response to massive assimilation of evaporites by silicate magmas. Sulfate from the evaporites would have oxidised ferrous iron in these magmas, facilitating the formation of immiscible ferric iron-rich melts and/or magnetite, which then separated from the magmas to form ore deposits. Ferric iron-bearing fluids with low Δ′ 17 O values, exsolved from the silicate magmas or the ore-forming melts, would have crystallised additional magnetite. An inventory study reveals that Proterozoic and Cambrian IOA deposits have lower Δ′ 17 O values than post-Cambrian IOA deposits. This shows that the Δ′ 17 O values of global IOA deposits reflect the changing isotope composition of atmospheric O 2 incorporated by evaporitic sulfate over time, and demonstrates that oxygen released from evaporitic sulfate is a common component in IOA deposits.

A unified framework for interpretable elevator fault diagnosis and predictive maintenance via style-aware CoT fine-tuning

Although Large Language Models (LLMs) have shown potential in industrial applications, they encounter significant hurdles in vertical scenarios like elevator maintenance, including hallucinations, lack of domain specificity, and an inability to interpret numerical physical states. To bridge this semantic-physical gap, this paper proposes a Unified Style-Aware Chain-of-Thought (SA-CoT) framework tailored for Small Language Models (SLMs). The novelty of our approach lies in two aspects: first, we construct a robust instruction dataset using a style-aware augmentation strategy to simulate diverse real-world user behaviors and noise; second, we innovate by textualizing raw sensor data, enabling the fine-tuned 4B-parameter SLM to generate high-dimensional embeddings for downstream numerical analysis. Experiments demonstrate a dual breakthrough: in generative diagnosis, the SA-CoT framework consistently outperforms general models, achieving a 5.6-fold improvement in BLEU-4 scores compared to GPT-4o. Furthermore, its embeddings capture physical features more effectively than traditional baselines, yielding highly competitive accuracy in Alarm Type Classification and Vibration Magnitude Regression. These results suggest that domain-aligned SLMs offer a robust and cost-effective framework for autonomous predictive maintenance, indicating that knowledge density plays a more critical role than parameter scale in specialized industrial applications.

Day‐Long Persistent Luminescence in Intrinsically Integrated Donor‐Acceptor Carbon Dots Enabled by Defect‐Mediated Charge Trapping

ABSTRACT Long‐persistent luminescence (LPL) materials capable of storing and releasing optical energy over extended timescales are highly desirable for next‐generation photonic technologies, yet structurally stable integrated donor‐acceptor (D‐A) systems capable of day‐scale, color‐tunable LPL remain elusive. Here, we report intrinsically integrated D‐A carbon dots exhibiting continuously tunable LPL from deep blue to yellow‐green, featuring day‐scale persistence of up to 36 h and a naked‐eye visible afterglow exceeding 4 h. Structurally, nitrogen‐doped carbon‐core donor domains are covalently coupled with arylboronic acid derived acceptor moieties through B─N linkages, forming an integrated D‐A framework enriched with intrinsic defect‐related trap states. Upon photoexcitation, intraparticle charge transfer (CT) generates long‐lived charge‐separated states, some of which are stabilized by intrinsic traps. Subsequent thermally activated detrapping releases the stored carriers and drives charge recombination, ultimately giving rise to day‐scale LPL. Furthermore, modulation of the acceptor electronic structure regulates the energy of the emissive CT state, enabling rationally tunable multicolor LPL. Benefiting from its day‐scale persistence and continuously tunable multicolor emission, this system enables potential applications in high‐resolution displays, dynamic anti‐counterfeiting, and intelligent information encryption. More importantly, this work establishes an intrinsically integrated D‐A‐trap design principle for ultralong LPL and a rational acceptor‐engineering strategy for multicolor persistent luminescence.

Ketodarolutamide may interact with the receptor-binding domain of the SARS-CoV-2 spike glycoprotein: the importance of halogenated benzonitrile

Causal effects of wildfire PM2.5 on hospital costs and length of stay in Brazil

UAMP: Consistent video object segmentation with uncertainty-aware memory propagation

The Segment Anything Model 2 (SAM 2) has emerged as a robust foundational model for video object segmentation. Nevertheless, it exhibits notable limitations in crowded scenarios, particularly those involving fast-moving objects or self-occlusion. Furthermore, its greedy-selection memory architecture is afflicted by “error accumulation,” which collectively impairs its performance in consistent object segmentation tasks. To address these critical drawbacks and enhance the consistency and robustness of SAM 2 in complex video scenarios, an enhanced variant of SAM 2, termed UAMP, is proposed, which integrates memory propagation based on explicit appearance and motion uncertainty modeling, coupled with a dual mechanism encompassing long-term memory updating and short-term memory selection. By fusing uncertainty-aware representations with these dual memory mechanisms, UAMP effectively accommodates dynamic variations in object appearance and motion, further refines the object memory bank, and thereby realizes consistent video object segmentation. Quantitative and qualitative evaluations conducted on diverse benchmark datasets demonstrate that UAMP achieves superior performance, particularly in scenarios involving occlusions and object reappearances. Specifically, UAMP yields consistent performance improvements over state-of-the-art (SOTA) methods across five video object segmentation (VOS) benchmarks, with a maximum enhancement of 5.6 points in the J&F metric. These findings underscore the robustness and effectiveness of the proposed UAMP method in addressing complex tracking scenarios, thus providing a valuable enhancement to SAM 2 for practical video object segmentation applications.

Quantum Spin‐1/2 Rings Built From [2]Triangulene Molecular Units

ABSTRACT Quantum spin rings represent fundamental model systems that exhibit distinctive quantum phenomena arising from their periodic boundary conditions and enhanced quantum fluctuations. Here, we report the on‐surface synthesis and atomic‐scale characterization of antiferromagnetic S = 1/2 quantum spin rings composed of pristine [2]triangulene units on Au(111). Using stepwise on‐surface synthesis followed by scanning tunneling microscopy tip‐induced dehydrogenation, we precisely constructed cyclic five‐ and six‐membered spin rings and investigated their spin states via scanning probe microscopy and multireference calculations. Bond‐resolved noncontact atomic force microscopy imaging reveals that the six‐membered ring retains a planar geometry, whereas the five‐membered ring exhibits pronounced structural distortion. The six‐membered ring hosts a uniform excitation gap that can be accurately described by a Heisenberg spin model and multireference CASCI calculations. In contrast, although an ideal C 5 ‐symmetric pentamer is theoretically expected to host a degenerate, frustrated ground state, the experimentally realized five‐membered ring is structurally distorted, which lifts this degeneracy and produces asymmetric spatial distributions of the spin ground state. Our findings establish a versatile molecular platform for exploring correlated magnetism and quantum spin phenomena in cyclic organic magnetic architectures.

Lipid A modification is associated with colistin susceptibility in Serratia nevei clinical isolates

Endothelial cannabinoid CB1 receptor deficiency reduces shear stress-induced arterial inflammation and lipid uptake

Abstract Peripheral cannabinoid CB1 receptor antagonists that lack central nervous system effects are emerging as promising therapies for metabolic disease, yet the role of endothelial CB1 signaling in atherosclerosis remains unclear. Here, we show that endothelial CB1 is expressed in human atherosclerotic plaques, is induced by oscillatory shear stress in atheroprone flow regions, and promotes vascular inflammation, permeability and lipid uptake. Endothelial-specific Cnr1 deletion or peripheral CB1 antagonism in mice attenuates atherosclerosis, reduces endothelial caveolae–dependent low-density lipoprotein uptake by downregulating caveolin-1 and ALK1 expression, and improves metabolic parameters in brown and white adipose tissue and the liver. The anti-atherogenic and metabolic effects are more pronounced in females, which is possibly linked to estrogen signaling. These findings identify endothelial CB1 as a proatherogenic, sex-biased regulator of vascular lipid transport and plaque development and associated metabolic dysfunction.

Bond strength of debonded orthodontic brackets after different reconditioning and priming protocols: An in vitro study

Introduction Although various bracket recycling methods have been studied, the combined influence of bracket reconditioning methods and primer protocols on the bonding performance of debonded orthodontic brackets remains unclear. This study aimed to evaluate the effects of three reconditioning methods and three primer protocols on the bond strength of rebonded orthodontic brackets to enamel. Materials and methods One hundred stainless steel brackets were bonded to extracted premolars and divided into 10 groups (n = 10). Ninety brackets were debonded, reconditioned using OneGloss bur, flaming, or sandblasting, and rebonded with Mani Bond, Denu Bond, or no primer. Ten new brackets served as controls. Bracket bases were examined using scanning electron microscopy (SEM). Bond strength, adhesive remnant index (ARI), and bonding reliability were evaluated using SBS testing, ARI scoring, and Weibull analysis. Results Bond strength was significantly influenced by the interaction between reconditioning method and primer protocol (p < 0.001). Sandblasting combined with Mani Bond produced the highest bond strength (9.33 ± 3.02 MPa), whereas OneGloss and sandblasting combined with Denu Bond achieved bond strengths comparable to new brackets and higher reliability, reflected by greater Weibull moduli (4.74 and 4.72). SEM demonstrated more effective mesh cleaning after sandblasting than after OneGloss bur treatment or flaming. ARI distributions differed significantly among groups (p = 0.01), with failures occurring predominantly at the bracket–adhesive interface. Conclusions The bonding performance of rebonded orthodontic brackets is influenced by the interaction between reconditioning and priming protocols. Appropriate combinations of these procedures may improve the rebonding outcome.