Browse Articles

Benchmarking machine learning architectures for menstrual recovery prediction using physiologically informed synthetic wearable data

A temporal-spectral dual-stream anti-noise bearing fault diagnosis model based on adaptive mode decomposition

Abstract In recent years, deep learning-based bearing fault diagnosis models have achieved remarkable performance under ideal experimental conditions. However, the accuracy and reliability of these models degrade significantly in real industrial scenarios where the acquired signals are often contaminated by strong background noise. Although a variety of anti-noise bearing fault diagnosis approaches have been developed, several limitations still remain. Most models directly take raw noisy signals as inputs, lacking an effective front-end noise suppression mechanism, which makes it difficult to sufficiently highlight fault-related information. Moreover, most models predominantly focus on the extraction and analysis of temporal features, neglecting the complementary fault characterization information embedded in other feature domains, thereby limiting the richness and discriminative power of feature representation. To address these issues, this paper proposes a temporal-spectral dual-stream anti-noise bearing fault diagnosis model based on adaptive mode decomposition. First, an adaptive mode decomposition module is designed to process raw noisy signals, aiming to suppress irrelevant noise components and enhance fault-related information, thereby providing a cleaner signal representation for subsequent diagnosis. Second, a temporal-spectral dual-stream bearing fault diagnosis framework is constructed to extract fault information from the same signal under different perspectives, aiming to enhance the richness and discriminative power of feature representation, thus the proposed model’s diagnostic performance under noise interference. Finally, extensive experimental results on two real-world cases demonstrate that, compared to current mainstream anti-noise bearing fault diagnosis models, the proposed approach achieves higher diagnostic accuracy under various noise intensities, sufficiently validating its effectiveness.

Hsuan‐Hung Liao

Effect of smoking on vitamin D status in the adults of Henan, China: the role of HDL-C and LDL-C

The mediating role of physical fitness in the relationship between 24-h movement guidelines and arithmetic fluency in adolescents: the MOVESCHOOL study

Evaluation of serum neurofilament light chain, GFAP, and peripherin as biomarkers in hereditary transthyretin amyloidosis

Abstract Early diagnosis and accurate monitoring of disease progression are crucial for timely therapeutic intervention in hereditary transthyretin amyloidosis (ATTRv). Neurofilament light chain (NfL) has emerged as a sensitive biomarker of neuroaxonal injury across neurodegenerative disorders. This study aimed to investigate serum levels of NfL, glial fibrillary acidic protein (GFAP), and peripherin (PRPH) in patients with ATTRV30M amyloidosis and pre-symptomatic gene carriers, compared with controls. Serum samples from 34 ATTRV30M patients, 17 pre-symptomatic ATTRV30M carriers, and 35 controls were analysed using conventional commercially available ELISA platforms to quantify NfL, GFAP, and PRPH concentrations. Serum NfL (sNfL) levels were significantly elevated in ATTRV30M patients compared with controls (threefold, p  = 0.0005) and were 1.6-fold higher than in pre-symptomatic ATTRV30M carriers. No significant differences were observed between pre-symptomatic carriers and controls. sNfL concentrations were higher in patients with polyneuropathy disability (PND) score > I compared with PND I ( p  = 0.0007). Serum GFAP and PRPH levels did not differ significantly among the study groups. Serum NfL represents a promising non-invasive biomarker for assessment of early symptomatic disease and monitoring of disease progression in ATTRV30M amyloidosis. In contrast, sGFAP and sPRPH appear to have limited diagnostic utility in this context.

Risk factors of osteoporosis in post-gastrectomy gastric cancer patients: a systematic review and meta-analysis

Machine learning-driven optimization of alkali-activated sustainable pavement concrete incorporating agro-industrial wastes

Abstract The integration of agro-industrial wastes into alkali-activated concrete (AAC) offers a sustainable alternative to traditional concrete pavement material, yet the synergistic effects of waste foundry sand (WFS) as a fine aggregate replacement and rice husk ash (RHA) as a partial binder replacement in ground granulated blast furnace slag (GGBS)-based AAC designed for rigid pavement applications remain under-explored. This study investigates the machine learning (ML)-driven optimization of AAC, addressing a critical research gap in material synergy and predictive modeling. Experimental results indicate that increasing WFS (0–30%) and RHA (0–20%) reduces compressive strength (CS) to some extent; however, an optimized mix containing 20% WFS and 15% RHA achieved a 28-day CS satisfying pavement-quality concrete requirements while maintaining workability within acceptable limits. A comparative ML framework comprising six algorithms; Multiple Linear Regression, Decision Tree, Random Forest, AdaBoost, Support Vector Regression, and Gradient Boosting was developed and benchmarked to predict the CS of 330 experimental samples. The Random Forest model achieved the highest predictive accuracy (R 2  = 0.9176, RMSE = 2.99 MPa, MAE = 2.62 MPa), with performance statistically comparable to Gradient Boosting and AdaBoost. Feature importance and interpretability analysis revealed that GGBS and fine aggregate content significantly influence CS, while higher WFS and RHA levels adversely affect strength. This research provides a scalable framework for designing low-carbon pavement materials by bridging the gap between experimental mix design and advanced computational optimization that substantially reduce dependence on resource-intensive experimental trials for AAC incorporating agro-industrial wastes

High daily fructose intake and late-evening fruit consumption are associated with ultrasonographically assessed hepatic steatosis in family medicine outpatients

The genetic etiology of spontaneous abortion: insights from chromosomal microarray analysis and whole-exome sequencing

Climatic transferability and regional validation of a PAT-based mechanistic model for apple scab ascospore maturation under temperate Himalayan conditions

Relative HU profiling on clinical CT estimates the subchondral bone plate boundary: an ex vivo cadaveric mechanical validation study

Abstract The subchondral bone plate (SBP) is crucial for joint biomechanics and is implicated in degenerative joint disease. Absolute CT Hounsfield units (HUs) vary with scanner settings, limiting generalisability. We evaluated whether CT-derived relative HU profiling can estimate the location and thickness of the subchondral mineralized interface by comparison with mechanical indentation-derived measurements. Eighteen distal radii from nine formalin-fixed cadaver donors underwent low-dose CT and indentation testing. Post-test scans were registered to pre-test scans so CT- and indentation-based measures were sampled at identical locations. The SBP starting point and thickness were defined from HU attenuation transitions and load–displacement curves, respectively. Association between methods was evaluated with linear mixed-effects models using donor identity as a random intercept and summarised with marginal/conditional R²; agreement was descriptively assessed using Bland–Altman analysis. Inter-observer reliability was assessed using intraclass correlation coefficients. HU-based measures were strongly associated with indentation-derived measures, with fixed-effect slopes of 0.953 for the starting point and 0.988 for thickness, and R²m/R²c values of 0.916/0.918 and 0.912/0.912, respectively. Donor-level clustering effects were minimal. Reliability was good for HU-based measurements and good-to-excellent for indentation-based measurements. Bland–Altman analysis showed smaller bias and narrower limits of agreement for thickness than for the starting point. Relative HU profiling on standard clinical CT provided reproducible estimates of subchondral mineralized boundary features in this ex vivo cadaveric model. These findings support the feasibility of using relative HU transitions as a radiologic surrogate, while further histological and in vivo validation is required before clinical application.

Circumstellar Origin of Chrysene (C ₁₈ H ₁₂ ) via Self‐Recombination of Resonantly‐Stabilized 1‐Indenyl Radicals and Implications to the Aromaticity of the Carbonaceous Asteroid Ryugu

ABSTRACT Polycyclic aromatic hydrocarbons (PAHs) represent key molecular building blocks of carbonaceous nanoparticles and have been identified in cold molecular clouds (TMC‐1), meteorites (Murchison, Allende), and asteroids (Bennu, Ryugu). However, the understanding of their formation and molecular mass growth processes in these extreme environments has remained elusive, especially with the emergence of resonantly‐stabilized free radical (RSFR)‐mediated pathways. Here, we report a combined experimental and computational study on the self‐recombination of the aromatic and resonantly‐stabilized 1‐indenyl radical (C 9 H 7 • ) isomer‐selectively forming the 18π‐Hückel PAH chrysene (C 18 H 12 ) in an overall endoergic reaction. These features account for the circumstellar origin of chrysene in pristine samples of the carbonaceous asteroid Ryugu and of the Murchison meteorite. At the microscopic level, the ring expansion involving cyclopentadienyl moieties to two six‐membered aromatic rings via pivotal spiroaromatic intermediates affords a versatile, RSFR‐initiated mass growth channel essentially leading to graphene‐type PAHs and eventually two‐dimensional nanostructures in high temperature circumstellar envelopes and in combustion processes.

MicroRNA-125a-5p promotes NLRP3/caspase-1/GSDMD pathway-mediated pyroptosis in endothelial cells during Kawasaki disease

Imaging of <i>Staphylococcus aureus</i> Infections and Biofilms Using a Selective Covalent Probe for the Unique Serine Hydrolase FphE

ABSTRACT Staphylococcus aureus is the leading cause of soft tissue infections which can often be treated with antibiotics. However, it can also cause significant mortality and morbidity from systemic infections and infections of surgical implants. Implant infections typically require invasive surgery, and treatment often necessitates removal of the implant because S. aureus biofilms are extremely difficult to eradicate with antibiotic treatment alone. Therefore, there is a significant need for improved diagnostic tools for rapid, noninvasive confirmation of S. aureus infections. We recently developed an activity‐based probe containing an oxadiazolone electrophile that selectively and covalently labels the S. aureus ‐specific serine hydrolase, FphE. Here we describe a Cy5‐labeled version of the probe, JJ‐OX‐012, and its characterization as an imaging agent for detecting biofilms both in vitro and in vivo. The probe labeled S. aureus biofilms in vitro, with virtually no background labeling of bacteria that lack FphE expression, such as Escherichia coli . Furthermore, using a mouse surgical implant infection model, we demonstrate that JJ‐OX‐012 can be used for noninvasive fluorescent imaging to detect S. aureus biofilms in vivo. Overall, these findings support the potential for using covalent probes targeting FphE as imaging agents for rapid detection and diagnosis of staphylococcal infections in vivo.

Multiobjective blood pump impeller optimization with three response surface methods and prototype stator casting for an integrated motor pump

Dual Single‐Atom Catalysts Unveiling Active Sites for Enhanced Ethylene Methoxycarbonylation

ABSTRACT Designing heterogeneous catalysts with well‐defined structure remains crucial for identifying true active sites and achieving superior performance. Here, we report a series of Ru‐based dual single‐atom catalysts (DSACs) constructed on TiO 2 , in which catalytically inert second metals (Zn, Mn, Mo, In) delicately modulates the electronic structure of Ru. Among them, Ru 1 Zn 1 /TiO 2 with most electron‐deficient Ru site exhibits highest activity and durability for acid‐free ethylene methoxycarbonylation (EMC) with a turnover frequency of 1403 h −1— one to two orders of magnitude higher than previously reported Ru‐based catalysts, ranking among the most efficient heterogeneous EMC catalysts. Benefiting from well‐defined and fine‐tuned structure of DSACs, an unambiguous establishment of the structure‐activity relationship has been achieved, pinpointing electron‐deficient Ru δ+ species as the true active sites for the EMC reaction. Further DFT calculations reveal the electronic coupling in Ru‐Zn site optimizes the d‐band center near the Fermi level, and facilitates the adsorption and activation of reactants, thereby reducing the energy barrier for the C‐C coupling step. This work establishes a general strategy for constructing bimetallic single‐atom catalysts as precise model systems for elucidating active sites and guiding the rational design of efficient and stable catalysts for carbonylation and related transformations.

An adaptive oppositional grey wolf optimizer for complex engineering problems

Boosting Triplet Exciton Harvesting via Multi‐Channel High‐Lying Reverse Intersystem Crossing in a Hot Exciton Material Featuring Locally Excited‐State Emission

ABSTRACT Hot exciton (HE) materials possessing a locally excited (LE) S 1 state are ideal for achieving narrow emission and high exciton utilization efficiency (EUE) in organic light‐emitting diodes (OLEDs). However, constrained by a single high‐lying reverse intersystem crossing (hRISC) channel, the currently established donor–bridge–acceptor (D–B–A) design suffers from low EUE max (≤ 50%) and consequently low external quantum efficiency (EQE max : 2%). Herein, we introduce a “multi‐functional subunit” triad strategy to circumvent this limitation by engineering the luminescent core to maintain the LE– S 1 state while simultaneously participating in the formation of multiple near‐degenerate T n states that exhibit substantial spin–orbit coupling with the S 1 state, thus activating multiple efficient hRISC pathways. This concept is validated using a newly designed molecule, P‐Cz‐SO, which exhibits a well‐defined LE– S 1 state with narrow blue emission. Transient spectroscopy reveals two distinct delayed fluorescence (DF) components, providing the first direct experimental evidence for multi‐channel hRISC processes. The resulting OLED demonstrates near‐unity EUE max and a record EQE max of 15.5% among deep‐blue HE‐OLEDs (CIEy ≤ 0.1). Comparative studies with a reference compound, P‐Cz‐Ph, confirm the critical role of the multi‐channel hRISC design. This work provides a general paradigm for achieving highly efficient LE– S 1 HE emitters.

Robust image quality evaluation in optical coherence tomography of skin using global, region-independent metrics

Abstract Optical coherence tomography (OCT) is a powerful imaging modality for visualizing tissue microstructures, but its utility is often limited by artifacts such as speckle, intensity decay, and blurring. While numerous image enhancement algorithms have been proposed to address these issues, the field lacks robust, objective metrics to consistently quantify image quality, hindering fair comparison and development of such methods. Existing metrics, such as signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR), depend heavily on user-selected regions of interest, introducing substantial variability. In this work, we introduce a fully automated framework that segments OCT skin images into air, signal, and noise regions and defines three global image quality metrics: noise-free pixel ratio (NFPR), global SNR (gSNR), and global contrast (gCN). These metrics analyze the entire image without requiring ROI selection, enabling reproducible and user-independent evaluations that correlate with both acquisition parameters and human visual perception. Extensive validation on skin OCT datasets demonstrates that the proposed metrics provide more consistent and stable characterization of image quality compared to conventional ROI-based metrics under the tested conditions. These results suggest that the proposed framework can support objective evaluation of OCT image quality and facilitate the development and benchmarking of image enhancement methods.