Browse Articles

Prevalence of insufficient sleep and its associated factors among children and adolescents in Wuxi, China: a cross-sectional survey

Correction: Impact of trace metals in fish waste-based organic fertilizer on growth promotion and nutritional components of spinach plant (Spinacia oleracea L.)

Enhanced neutrosophic group decision-making framework for risk evaluation of engineering project empowered by new quality productive forces: a case study for photovoltaic power generation projects

HIV dominates T-cell immune responses in HIV–HBV co-infection and is associated with increased HIV reservoir transcription

Human umbilical cord mesenchymal stem cell exosomes promote liver regeneration by inducing CD206+ macrophage polarization and suppressing ZBP1 PANoptosis

Development and psychometric evaluation of a symptom assessment scale for adjuvant radiotherapy in breast cancer patients based on the symptom experience model

Bioprocessing of monomethyl ether from Alternaria alternata, a multi-target antiproliferative compound as emphasized by in silico analysis

Construction and validation of an interpretable machine learning model for predicting diabetes risk in COPD patients

A paper-based fluorometric sensor using L-Cys-CdTe quantum dots coupled with smartphone readout for automated pretreatment and detection of Cu(II)

An X-linked long non-coding RNA, PTCHD1-AS, and the core features of autism

Hybrid noise-resistant technique for malware classification

Abstract Visualization-based malware detection has recently gained significant attention for binary and multiclass malware classification using machine learning and deep learning techniques. However, existing visualization-based frameworks still face several important limitations, including insufficient robustness evaluation, limited cross-dataset validation, restricted malware diversity, and difficulty distinguishing visually similar and noise-sensitive malware families. In many cases, the visual similarity between malware classes and the presence of perturbations negatively affect feature extraction quality, leading to degraded classification performance and reduced generalization capability. To address these challenges, this study proposes a novel hybrid malware representation framework that integrates Convolutional Autoencoder (CAE)-based latent structural learning with Local Binary Pattern (LBP)-based texture feature extraction for robust malware classification. To the best of our knowledge, this study represents one of the first comprehensive investigations of hybrid latent-texture representation learning within a memory-forensics malware visualization setting while jointly addressing robustness, perturbation resilience, scalability, and cross-dataset generalization through a unified evaluation framework. The proposed framework combines global hierarchical representations learned through CAE with fine-grained local texture descriptors extracted using LBP to improve the discrimination of visually similar malware families and enhance robustness against noisy visualization conditions. The extracted features are subsequently evaluated using multiple machine learning classifiers, where XGBoost achieved the highest performance with an accuracy of 99.90%, precision of 99.79%, recall of 99.92%, and F1-score of 99.85%. To comprehensively evaluate the proposed framework, extensive experiments are conducted using both a memory-forensics malware dataset and the large-scale BODMAS dataset containing 134,435 PE malware samples spanning 581 malware families. The experimental evaluation incorporates cross-validation, ablation analysis, robustness assessment under multiple perturbation conditions, and feature-space visualization analysis. The results demonstrate that the proposed CAE+LBP framework consistently outperforms standalone feature extraction approaches and conventional end-to-end CNN models while maintaining strong robustness and cross-dataset generalization capability across diverse malware distributions and noisy conditions.

Lineage and organ signals sequentially build organ intrinsic nervous systems

The impact of housing cost burden on self-rated health among urban renters in China

A multimodal artificial intelligence system integrating facial expression voice emotion and behavioral movement analysis for assessing social-emotional competence in preschool children

A neuro-synthesis of psychological constructs relevant to posttraumatic stress disorder

Abstract Post-traumatic stress disorder (PTSD) is a heterogeneous neuropsychiatric disorder defined by a constellation of symptoms and psychological processes following trauma exposure. PTSD research spans diverse constructs, but heterogeneous terminology and task-specific findings complicate efforts to link these constructs to neural systems. While neuroimaging meta-analyses have revealed structural, functional, and network abnormalities in PTSD, few studies have examined how PTSD-relevant psychological constructs are organized according to broader meta-analytic activation patterns. To address this gap, we combined Neurosynth and PubMed text mining to (1) identify psychological construct terms that are represented in PTSD research discourse and have sufficient neuroimaging meta-analytic support, and (2) determine whether these constructs cluster into neuroimaging-informed domains that align with canonical brain networks. Using graph network and community detection clustering, we identified four stable clusters that were labeled “Cognition,” “Emotion,” “Memory Function,” and “Decision-Making.” The cognition cluster converged on middle frontal gryus, amygdala, and dorsal anterior insula. The emotion cluster mapped onto the amygdala, subgenual anterior cingulate (sgACC), and ventromedial prefrontal cortex (vmPFC). The memory function cluster involved hippocampus, amygdala, and vmPFC. The decision-making learning cluster encompassed vmPFC, striatum, and sgACC. By integrating Neurosynth term-based meta-analyses with PubMed filtering, network-based clustering, our approach bridges semantic, mechanistic, and neural organization, reducing conceptual fragmentation providing a neuroimaging informed conceptual framework of PTSD-related discourse.

New phenyl γ-butyrolactone glucosides derived from γ-irradiation of esculin in methanol exhibit significant α-glucosidase inhibitory activity

Safe causal-graph primal-dual multi-agent scheduling for energy- and latency-constrained edge-assisted cognitive radio networks

Enantioselective hydrogen atom relay via non-covalent catalyst assembly

Abstract Most biological functions are regulated by chiral molecules 1 that contain at least one tertiary stereogenic carbon, that is, a carbon with one C( sp 3 )–H bond. Hydrogen atom transfer (HAT) 2 is a straightforward strategy that can be used to either edit 3 or introduce tertiary stereocentres in multiple synthetically useful transformations 4 , especially when coupled with photoredox catalysis 5,6 . However, traditional de novo design of chiral HAT catalysts that provide sufficient enantiocontrol over short-lived open-shell intermediates 7 has represented a major hurdle in the development of enantioselective HAT reactions. Here we describe a distinct approach in which chiral HAT catalysts are obtained in situ by non-covalent self-assembly of privileged chiral phosphoric acids and commercial 2-mercaptopyridines. The phosphoric acid serves as a modular interchangeable chiral element that renders the achiral thiol effectively chiral, thereby allowing access to a previously inaccessible combinatorial space of chiral HAT catalysts. This platform enabled the photochemical deracemization of 2-aryl pyrrolidines, which are prevalent scaffolds in active pharmaceutical ingredients. Optical enrichment occurs by means of enantioselective hydrogen atom relay, in which a single chiral assembly orchestrates hydrogen atom abstraction and delivery. This conceptual approach of relaying chiral information through non-covalent assembly paves the way for discovery of numerous asymmetric radical transformations.

Modelling continuous-time fault contagion in power grids with a graph neural Hawkes process

Bar adsorptive microextraction for the qualitative detection of cocaine and its metabolites in urine for anti-doping control

Abstract Anti-Doping laboratories face ongoing challenges due to the continuous emergence of new drugs. To meet the stringent technical requirements of the World Anti-Doping Agency (WADA), laboratories must enhance their operational procedures. The Bar Adsorptive MicroExtraction (BAμE), a modern technique aligned with the green analytical chemistry and successfully applied to various matrices, represents a promising alternative in the anti-doping context. The BAμE was applied for the qualitative identification of Cocaine and three metabolites (BenzoylecGonine, ecGonine Methyl Ester and Cocaethylene) in human urine, followed by derivatization and gas chromatography–mass spectrometry analysis. Using optimized conditions, the method provided limits of identification of 1.0–10.0 ng/mL, linear calibration curves of 1.0–200.0 ng/mL, r 2  > 0.9950, extraction recoveries of 39.4%–71.1%, with high selectivity, robustness, accuracy and precision. Five urine samples were successfully assayed, with unequivocal detection of several target compounds. The BAμE-based methodology was successfully applied for the first time to anti-doping control of cocaine in human urine, providing high selectivity, sensitivity, and reduced interferences with minimal toxic solvent use. The BAμE(P3)-μLD/GC–MS(SIM) approach is cost-effective, user/eco-friendly, and fit-for-purpose as it complies with all WADA technical requirements.