Browse Articles

Optimization of attention mechanisms in choral conducting education using EEG and eye movement tracking with transfer learning algorithms

Expression of Concern: BlockTicket: A framework for electronic tickets based on smart contract

Cellular water-potential sensing through biomolecular condensation

Distribution characteristics and influencing and controlling mechanisms of sediment/pore water metals in the intermountain basin of the Northwest Pacific Ocean

Coral restoration alters reef soundscapes but machine learning and manual analyses suggest different recovery rates

Coral restoration is recognised as a critical tool to mitigate pantropical degradation of reef ecosystems. Robust monitoring of restoration progress is crucial for projects to evaluate their success, improve practice, and share knowledge. However, traditional visual surveys often fail to capture the full impact of coral restoration on reef function. Therefore, we employed Passive Acoustic Monitoring (PAM) to assess whether the soundscape of a coral restoration site in the Seychelles differs from adjacent healthy and degraded reference reefs. We applied two methods of soundscape analysis: manual detection of unidentified fish sounds; and machine learning-based Uniform Manifold Approximation and Projection analysis. Results were approach-specific: the manual approach highlighted similarities in fish calls between the restoration site and the healthy reference reef, while the machine learning approach extracted broader soundscape patterns, clustering the restoration site alongside the degraded reference reef. Although this is a single-site study, these findings suggest that a) coral restoration alters reef soundscapes, though recovery time may be taxon-specific, and b) multiple metrics are needed to bridge single-taxon and broad soundscape scales. This study contributes to the evolving field of soundscape ecology in coral reef ecosystems, highlighting the utility of PAM in monitoring changes to reef function through coral restoration.

Expression of Concern: Development of innovative alkali activated paste reinforced with polyethylene fibers for concrete crack repair

Could a pill prevent the world’s deadliest cancer?

Chemical composition, seasonal variability and source identification of PM2.5 in an urban background site in Medellín—Colombia

This study assessed the chemical characterization of PM 2.5 at an urban background site in Medellín—Colombia, a city situated in a topographically constrained valley exposed to air pollution accumulation influenced by its complex Andean topography and a combination of local and regional emission sources. A total of 112 samples of PM 2.5 were collected between March 2019 and March 2020, samples went through comprehensive sampling and chemical characterization (ICP-MS for metals, ion chromatography for anions, thermal/optical analysis for carbonaceous species). Daily PM 2.5 concentrations ranged from 8.15 to 37.86 μ g/m 3 (mean: 21.73 ± 6.75 μ g/m 3 ). Principal Component Analysis (PCA) identified five major pollution sources, with mineral and resuspended dust (33.6% of explained variance) and secondary aerosols (17.0%) being the most prominent. The study area experiences clearly defined dry and wet periods, marked by distinct precipitation regimes. During the year these atmospheric conditions influence the concentration levels of pollutants. The integration of NOAA HYSPLIT back-trajectories and NASA FIRMS (VIIRS J2) fire hotspot data revealed long-range transport from the Magdalena, Orinoco and Amazon basins, deteriorating local air quality specially during dry periods. While local traffic and industrial emissions constitute a constant baseline, regional biomass burning and unfavorable meteorological conditions are the primary drivers of episodic high-pollution events. The study underscores the need for targeted strategies addressing both persistent sources as traffic or industrial emissions and episodic events to mitigate health and environmental impacts.

Effects of social media sponsorship disclosure on adolescents’ advertising literacy and purchase intention

Introduction In the rapidly evolving digital age, adolescents have become a particularly vulnerable group impacted by social media marketing. This study aims to examine how different types of sponsorship disclosure (full, partial, and none) in social media affect adolescents’ ability to recognize advertising, their advertising literacy, and their purchase intentions. Materials and methods A self-administered questionnaire was used to assess the effects of three types of sponsorship disclosure in social media posts on adolescents. In 2022, a total of 3,149 high school students were recruited from 36 schools across Taiwan using a probability proportional to size sampling method. Adolescents were randomly assigned to one of three experimental groups and exposed to social media posts featuring varying levels of sponsorship disclosure. Results The study found that adolescents exposed to full sponsorship disclosure had significantly higher rates of advertisement recognition and greater conceptual advertising literacy compared to those in the partial or no disclosure conditions. The multivariate analysis revealed that advertisement recognition was associated with higher levels of both conceptual and attitudinal advertising literacy. Adolescents who could recognize advertisements demonstrated higher attitudinal advertising literacy, which was associated with lower purchase intentions. Conclusions Clear sponsorship disclosure is crucial for raising advertising awareness among adolescents, enhancing their advertising literacy, and potentially protecting them from increased susceptibility to purchase intentions.

Human blood stem cells remember previous inflammation

HIV/AIDS burden, attributable risk factors, and projections among reproductive-age adults in China, 1990–2035: A GBD 2023 analysis

Background HIV/AIDS remains a leading infectious disease burden globally. While highly active antiretroviral therapy has transformed HIV into a manageable chronic condition, contemporary trends and attributable risk factors among reproductive-age adults in China—a critical demographic for epidemic control—remain poorly characterized. This study provides an updated comprehensive assessment of HIV/AIDS burden in this population using Global Burden of Disease Study 2023 data, which incorporates methodological refinements and post-pandemic data not available in earlier versions. Methods We extracted age-standardised incidence, prevalence, mortality and disability-adjusted life-years (DALYs) for Chinese adults aged 15–49 years from GBD 2023. Joinpoint regression estimated annual percent change and identified significant trend inflection points. Population attributable fractions were analysed for unsafe sex, drug use, and intimate-partner violence. ARIMA models forecasted burden trends to 2035 with 95% uncertainty intervals. Results Between 1990 and 2023, age-standardised rates increased substantially: incidence +301% (from 0.82 to 3.29 per 100,000), prevalence +774% (4.84 to 42.29), mortality +1,623% (0.13 to 2.24), and DALYs + 1,439% (7.96 to 122.48). Males showed consistently steeper increases than females across all metrics (all p < 0.05). Unsafe sex accounted for 70% of HIV/AIDS mortality in 2023. ARIMA projections indicate that while male incidence will plateau at ~5.0 per 100,000 by 2035, female metrics will continue rising (incidence +21%, prevalence +30%), widening the sex disparity. Conclusion China’s HIV epidemic has transitioned to a chronic disease burden. Notably, this burden shows pronounced sex differences. Mortality is rising among people living with HIV, particularly among women. These trends highlight urgent needs for gender-differentiated prevention strategies. Our findings provide measurable targets for China’s National HIV/AIDS Action Plan (2024–2030) and SDG 3.3 monitoring. They further demonstrate how standardized burden estimation supports precision public health.

Fault diagnosis of aero-engines using transfer dispersion entropy and dispersion patterns

Dispersion entropy (DE) can effectively detect the chaotic features in ordered sequences. However, DE is only defined by the probability distribution of static dispersion patterns, ignoring the dynamic transitions between pairwise dispersion patterns. Therefore, in this paper, by introducing the transition probability between pairwise dispersion patterns, we propose the transfer dispersion entropy (TDE) to detect the chaotic characteristics of the system. Additionally, based on both the difference in the number of each dispersion patterns and the difference in the transfer probability matrix, we define a dissimilarity measure for different sequences, namely transfer dissimilarity based on dispersion patterns (TDDP). Then, the proposed methods are verified on numerical simulation data and NASA-CMAPSS aero-engine simulation data.Compared with the existing entropy algorithms, the results show that TDE can not only capture more detailed chaotic changes, but also identify the early degradation of gas path components by detecting abnormal shifts in pattern transition behaviors, enabling more timely fault warnings.Finally, the combination of TDDP and multidimensional scaling can be used for similarity classification of aero-engine simulation time series.

Correction: Megakaryocytic Leukemia 1 (MKL1) Regulates Hypoxia Induced Pulmonary Hypertension in Rats

mRNA splicing in turkey muscle satellite cells is dynamically altered upon thermal challenge

Regulation of gene expression at the transcriptional and post-transcriptional levels is essential for proper development and growth, with tightly coordinated cellular processes supporting key biological functions. While transcription determines the available mRNA pool, post-transcriptional modifications such as alternative splicing (AS) increase transcriptome complexity and enable the production of diverse protein isoforms. In muscle, AS is critical in generating muscle-specific proteins required for normal development and function and may be particularly susceptible to disruption by thermal stress. This study examines how thermal challenges—both cold and heat—affect muscle biology by analyzing AS events during the proliferation and differentiation of skeletal muscle satellite cells (SCs). Isoform identification and AS analyses were performed on RNA-seq data from a prior study of skeletal muscle SCs derived from commercial turkeys and exposed to three temperature conditions (33°C, 38°C, or 43°C) during proliferation or differentiation. Analyses revealed 61,266 predicted splicing events across 5,202 annotated genes. Significant differential splicing was observed in all temperature comparisons, and between proliferating and differentiating cells at each temperature. Additionally, there was a strong association between differentially spliced genes (DASs) and differentially expressed genes (DEGs). This study provides a comprehensive catalog of splice isoforms for future functional analyses, many of which are likely to result in protein variants that influence SC proliferation, differentiation, and ultimately, muscle development and performance.

A taxonomy for detecting and preventing temporal data leakage in machine learning-based build prediction: A dual-platform empirical validation

Modern software development relies on automated build systems that compile and test code whenever developers make changes. Predicting whether these builds will succeed or fail before execution could save computational resources and developer time. However, many machine learning models for build prediction suffer from temporal data leakage, a methodological flaw where the model inadvertently uses information that would only be available after the build completes, producing artificially inflated accuracy that fails in real-world deployment. This study develops a three-type taxonomy to systematically identify and prevent such leakage: (1) Direct Outcome Encoding (using the build result itself as a feature), (2) Execution-Dependent Metrics (information generated during build execution), and (3) Future Information Leakage (using data from chronologically later builds). Applying this taxonomy reveals that prior studies reporting 95–99% accuracy likely used contaminated features, while realistic accuracy is substantially lower. The methodology is validated on 175,706 builds from two open-source CI/CD platforms spanning 10 years: TravisTorrent (100,000 builds, 2013–2017) and GHALogs (75,706 workflows, 2023). Removing leaky features reduces accuracy by 15.07 percentage points on TravisTorrent (97.8% to 82.73%) but only 0.48 points on GHALogs (83.77% to 83.30%), revealing that modern GitHub Actions’ tight integration with repositories enables accurate prediction from static project metadata alone. Using only legitimately available pre-build features, Random Forest classifiers achieve 82.73% (TravisTorrent) and 83.30% (GHALogs) accuracy, sufficient for practical deployment. Surprisingly, project maturity and build history prove more predictive than code complexity metrics, suggesting organizational factors outweigh code quality. The models generalize across programming languages (Java, Ruby, Python, JavaScript) with minimal performance variation. Open-source tools for detecting temporal leakage in any software prediction task are provided.

Comparison of scenario reduction approaches for reservoir inflow timeseries generated by a Bayesian Neural Network

Dealing with uncertainty in predicted inflows presents a major challenge in optimal reservoir flood control. Scenario-based stochastic control approaches address this by generating multiple inflow time series from probabilistic models, each representing a possible future with associated likelihoods. However, using too many scenarios increases computational complexity, while too few may compromise representativeness. Although the two critical steps of scenario generation and reduction have been extensively explored in other fields, their application to reservoir inflow dynamics remains limited. This study develops and applies a probabilistic data-driven model, specifically, a Bayesian Neural Network (BNN), for scenario generation. While the model exhibits limitations in predicting peak inflows due to data scarcity, it effectively captures temporal dependencies in inflow time series and achieves high short-term accuracy, as measured by the Nash–Sutcliffe Efficiency Coefficient (NSE) and Root Mean Squared Error (RMSE), though performance declines over longer horizons. For scenario reduction, four distance measures widely used in other domains, i.e., the Manhattan, Euclidean, Wasserstein, and energy distances, are evaluated. Experimental results show that the energy distance best preserves the statistical properties of the full scenario set, followed by the Manhattan and Euclidean distances. However, in terms of retaining extreme inflow scenarios, which are critical for flood control, the Manhattan and Euclidean distances outperform others based on a custom index measuring the envelope size of the original scenario set using the l 1 -norm. In terms of computational efficiency of scenario reduction approaches, the energy distance is the most expensive (quadratic in m , the number of reduced scenarios), while the Wasserstein scales linearly. In the examples used, reduced sets are shown to adequately capture extremes when the number of scenarios m  ≥ 30. Considering the trade-off between preserving extremes and computational cost, the Manhattan and Euclidean distances with m  = 30 are recommended as a practical choice for reservoir inflow scenario reduction.

Modeling the microbial contribution to human energy balance using the Digestion, Absorption, and Microbial Metabolism (DAMM) model

Colonic microorganisms have been linked to human health and disease, specifically metabolic disease states such as obesity, but causal relationships remain to be established. Previous work demonstrated that interactions between the host’s diet and intestinal microbiome were associated with human energy balance by affecting the human’s energy absorption, quantified by metabolizable energy. We developed the Digestion, Absorption and Microbial Metabolism (DAMM) model, which explicitly accounts for the energy contributions of the colonic microbial community in five steps. 1) The DAMM model breaks down the diet composition into the gross energy of the individual macronutrients. 2) It calculates direct absorption in the upper gastrointestinal tract. 3) It uses microbial stoichiometry to estimate the consumption of the remaining unabsorbed nutrients by microbes in the large intestine. 4) It quantitatively predicts microbial production of short-chain fatty acids (SCFA) and methane in the colon. 5) The DAMM model estimates absorption from the colonic tract to the host, including SCFAs. When used to predict the results from a clinical study that compared two distinctly different diets, the DAMM model captured the directionality and magnitude of change in measured metabolizable chemical oxygen demand (which can be converted to metabolizable energy), estimated substrate availability within the colon, and predicted rate of production of microbially derived short-chain fatty acids. It improved on the accuracy of metabolizable chemical oxygen demand predictions compared to the Atwater factors, increasing the fit from R 2  = 88% (Atwater) to R 2  = 96% (DAMM). The model reduced systematic bias on one of the diets and decreased the mean difference between measurement and predictions from −22.3 gCOD d -1 to −2.5 gCOD d -1 . The DAMM model now can be linked to existing human models that predict changes in body energy stores to extend our understanding of how microbial metabolic processes affect macronutrient absorption and metabolizable energy.

Molecular Engineering for Nonlinear Fluorescence: En Route to Three-Photon Absorption via Sequential One-Photon Excitation

Hyperelongate ornamental tail feathers in a new early Cretaceous enantiornithine bird

Bird diversity is reflected in the abundance and variety of extraordinary plumages. Some of these include elongate, ornamental tail feathers that are typically attributed to either intraspecific communication in monomorphic species or sexual selection in sexually-dimorphic ones. Enantiornithines (Aves: Ornithothoraces) were the most diverse group of birds during the Cretaceous. Importantly, some enantiornithine fossils preserve soft tissues, most often in the form of feathers surrounding the body. Unlike any living bird, many enantiornithine specimens lack tail feathers (rectrices) all together, with the tail region consisting entirely of contour feathers. However, when present, enantiornithine rectrices typically consist of a pair of elongate, ornamental feathers with unusually wide rachises, referred to as rachis-dominated feathers. Here we describe Plumadraco bankoorum gen. et sp. nov., a new bohaiornithid enantiornithine with a pair of exceptionally long rectrices. These tail feathers measure twice the individual’s body length, ending in proportionally small pennaceous rackets, thus adding to the growing diversity of these unusual feathers. The fine preservation of these tail feathers, in comparison to other enantiornithine rectrices, reveals previously unrecognized structural variation that hints at their potential function in courtship displays. Although ornamental feathers in enantiornithines are widely considered sexually dimorphic, determining the selection pressures that shaped them is difficult due primarily to limited soft tissue data. Enantiornithine rectrices are likely the result of an interplay between both sexual and naturally selective pressures, similar to the processes which produce analogous structures in birds today.

Adaptive and multi-scale feature fusion for Chinese news headline classification

The rapid growth of online news has led to an explosion of short Chinese headlines, which often suffer from sparse features, limited context, and high ambiguity—posing significant challenges for accurate classification. To address these issues, this paper proposes two tailored deep learning models: ERNIE-AAFF-SECNN for large-scale datasets, which enhances semantic representation via adaptive fusion of multi-layer ERNIE features and improves local feature extraction with SE-empowered CNN; and ERNIE-MSSE-DSCNN for small-scale datasets, which integrates multi-scale SE attention, depthwise separable convolutions, and adversarial training to boost robustness under data scarcity. A large number of experiments have shown that both of these models have achieved the most advanced performance. It is worth noting that the accuracy of ERNIE-AAFF-SECNN on the THUCNews and Toutiao datasets is 1.28% and 0.55% higher, respectively, than that of the lightweight SOTA model TinyBERT. The accuracy of ERNIE-MSSE-DSCNN on a 10% training dataset is 2.62% and 3.76% higher than that of the lightweight SOTA model TinyBERT, respectively. It demonstrates outstanding effectiveness under both standard and low-resource Settings. These results demonstrate that targeted architectural enhancements—such as adaptive feature fusion and multi-scale attention with adversarial training—can significantly improve the accuracy and robustness of short-text classification in practical Chinese news applications.