Browse Articles

Optimizing YOLOv11 for automated classification of breast cancer in medical images

Retraction: Navigating board dynamics: Configuration analysis of corporate governance’s factors and their impact on bank performance

Silk-enabled conformal intraventricular interfaces for minimally invasive neural recordings

Abstract Flexible neural interfaces capable of monitoring subcortical neuronal activity facilitate the study of deep brain neural circuits and their interactions with the cortex. However, there exists a paucity of translational tools for interfacing subcortical nuclei surfaces within the intraventricular cerebrospinal fluid. Here, we developed a flexible and conformal intraventricular interface (IVI) featuring a deformable microelectrode array paired with a silk scaffold. The IVI can be minimally invasively implanted into the lateral ventricles with the assistance of commonly used clinical catheters, self-unfolding in the cerebrospinal fluid environment to conformally attach to the surfaces of periventricular neural structures, and capturing high-quality signals by virtue of the microelectrode’s in-plane shielding. In parkinsonian ewes, the IVI detects deep brain abnormalities and achieves stable, biocompatible in vivo recordings for four weeks. This platform enables chronic monitoring and circuit analysis of healthy and diseased deep brain regions, facilitating studies of neural circuits between periventricular surface neurons and distant brain areas.

Segmentation of gastroesophageal reflux events using a semi-U-Net architecture with 1D/2D CNNs

Lung cancer symptoms awareness among Ethiopian adults: A latent class analysis

Background There is limited evidence regarding lung cancer awareness in developing countries. In Ethiopia, 92.2% of lung cancer patients present at facilities with late stages, leading to poor treatment outcomes. This emphasizes the importance of early detection. Symptom awareness is crucial for reducing delays. This study aimed to identify latent classes of lung cancer symptom awareness and their predictors, guiding class-specific interventions. Methods A population-based cross-sectional survey was conducted from October to December 2023 among a randomly selected 2388 adults in Addis Ababa, Ethiopia. A face-to-face interview was conducted using the validated Lung Cancer Awareness Measure (Lung CAM). Latent class analysis and latent class multinomial logistic regression were used to identify classes and predictors of class membership. Results Three distinct classes of participants were identified: “poor awareness” (Class 1: 38%), “fair awareness” (Class 2: 37.5%), and “good awareness” (Class 3: 24.5%). The average symptom awareness score was 7.8 out of 14. The most commonly recognized symptom was coughing up blood (72%), while changes in the shape of fingers were the least recognized (20%). Being male, employed, having a higher education level, using out-of-pocket money for health expenses, and knowing someone with cancer significantly increased the odds of belonging to the “good awareness” class, with adjusted odds ratios ranging from 1.66 to 12.60. Conclusion Only one-fourth of participants were classified as class 3, denoted as “good awareness,” indicating a significant gap in symptom awareness. Respiratory symptoms were mostly well-known. Class membership varied across sociodemographic and related characteristics. Hence, there is a need for class-specific educational intervention and a focus on non-respiratory symptoms.

Modern sea-level rise breaks 4,000-year stability in southeastern China

Abstract Quantifying physical mechanisms driving sea-level change—including global mean sea level (GMSL) and regional-to-local components (that is, sea-level budget)—is essential for reliable future projections and effective coastal management 1,2 . Although previous research has attempted to resolve China’s sea-level budget from the 1950s 3,4 , these studies capture short timescales and lack the long-term context necessary to fully assess modern sea-level rise in southeastern China 5 —one of the world’s most densely populated regions with immense socioeconomic importance 6 . Here we show that GMSL followed three distinct stages from 11,700 years before present (BP) to the modern day: (1) rapid early Holocene rise driven by the deglacial melt of land ice; (2) 4,000 years of stability from around 4200 BP to the mid-nineteenth century when regional processes dominated sea-level change; and (3) accelerating rise from the mid-nineteenth century. Our results arise from spatiotemporal hierarchical modelling of geological sea-level proxies and tide gauge data to produce site-specific sea-level budget estimates with uncertainty quantification. It is extremely likely ( P  ≥ 0.95) that the GMSL rise rate since 1900 (1.51 ± 0.16 mm year −1 , 1 σ ) has exceeded any century over at least the past four millennia. Moreover, our analysis indicates that at least 94% of rapid modern urban subsidence is attributable to anthropogenic activities, with localized subsidence rates often exceeding GMSL rise. Such concurrent acceleration of global sea-level rise and rapid localized subsidence has not been observed in our Holocene geological record.

DNA computing function switching by programming base stacking interactions with minimal molecular architecture changes

Research on multimodal data enhanced SLAM algorithm for global mapping of underground coal mines

Contrastive learning-enhanced personalized interaction dual tower network for recommendation

Dual-tower retrieval models have become a prevalent solution in large-scale recommendation systems due to their scalability and deployment efficiency. However, they face critical limitations including insufficient modeling of user behavior sequences, lack of personalized inter-tower interactions, and poor representation learning for long-tail content. To address these issues, we propose a novel framework called Contrastive Learning-Enhanced Personalized Interaction Dual Tower Network (CL-EPIDTN). This model integrates a multi-layer Transformer to capture dynamic user preference shifts, and introduces a dual-path personalized enhancement mechanism to strengthen user–item feature dependencies. Additionally, a contrastive learning strategy is employed to enhance the representation learning of long-tail items and low-activity users under sparse data conditions. Extensive experiments on two public datasets (Amazon Books and TmallData) demonstrate the effectiveness of our method. CL-EPIDTN achieves the best performance across multiple metrics, with Hit Rate@10 of 0.0351 and Recall@50 of 0.1123 on Amazon Books, and Hit Rate@10 of 0.0901 and Recall@50 of 0.1599 on TmallData, outperforming six state-of-the-art baselines. These results highlight the potential of CL-EPIDTN for both academic research and practical deployment in real-world recommender systems, particularly in handling personalization and data sparsity challenges.

A US case-control study to estimate infant group B streptococcal disease serological thresholds of risk-reduction

Abstract Maternal vaccines to prevent infant Group B Streptococcus (GBS) disease have progressed through phase II development and may be licensed based on immunologic endpoints, which have yet to be approved by regulatory authorities. Here we present a multistate case control study to characterize the relationship between serotype-specific anti-capsular polysaccharide (CPS) immunoglobulin G concentrations near birth and infant GBS disease risk reduction. Antibody concentration distributions are significantly lower for cases (n = 643) than controls (n = 2801) and serologic thresholds varied by serotype and age at onset, with 80% serotype-specific protective thresholds ranging from 0.52 to 2.49 mcg/mL for early-onset disease (EOD; <7 days old) and 0.02 to 0.14 mcg/mL for late-onset disease (LOD; 7-89 days old). Our study provides the most robust data to date that protection thresholds vary by serotype and are notably lower for LOD than EOD, thereby informing potential serological endpoints for phase III trials evaluating CPS-based maternal GBS vaccine candidates.

Hallmark features of conventional BCS superconductivity in 2H-TaS2

Abstract Layered transition metal dichalcogenides (TMDs) are model systems to investigate the interplay between superconductivity and the charge density wave (CDW) order. Here, we use muon spin rotation and relaxation ( μ + SR) to probe the superconducting ground state of polycrystalline 2H-TaS 2 , which hosts a CDW transition at 76 K and superconductivity below 1 K. The μ + SR measurements, conducted down to 0.27 K, are consistent with a nodeless, BCS-like single-gap s -wave state. Fits to the temperature dependence of the depolarization rate and Knight shift measurements support spin-singlet pairing. Crucially, no evidence of time-reversal symmetry breaking (TRSB) is observed, distinguishing 2H-TaS 2 from polymorphs like 4Hb-TaS 2 , where TRSB and unconventional superconductivity have been reported. These findings establish 2H-TaS 2 as a canonical BCS superconductor and provide a reference point for understanding the diverse electronic ground states that emerge in structurally distinct TMD polymorphs.

Non-linear associations of a body shape index with diabetes among adults: A cross-sectional study

Objective This study aimed to explore the relationship between a body shape index (ABSI) and diabetes, and to assess the robustness of this relationship across different population subgroups. Methods Utilizing data from the National Health and Nutrition Examination Survey (NHANES) database, this study employed multivariate linear regression analysis to evaluate the association between ABSI and the likelihood of having diabetes. This research further explored non-linear issues in gender stratification through smooth curve fitting and two-part linear regression models, analyzing different subgroups including gender, race, hypertension, and stroke. Results A total of 34,693 participants were involved in the study, with a diabetes prevalence of 11.60%. The prevalence increased with higher tertiles of ABSI. After comprehensive adjustment, ABSI was positively correlated with diabetes (OR = 1.42, 95% CI: 1.33, 1.52). Participants in the highest quartile of ABSI had a 96% higher odds of having diabetes (OR = 1.96, 95% CI: 1.69, 2.26), than did those in the lowest quartile. Smooth curve fitting analysis revealed a non-linear, inverse L-shaped relationship between ABSI and diabetes, with a breakpoint at 9.54. Subgroup analyses indicated that the association between ABSI and diabetes remained stable across different populations, except for those with a history of stroke. Conclusion Our findings suggest that there is a positive correlation between diabetes and increased ABSI. ABSI may serve as an effective alternative indicator to other obesity indices, such as BMI.

Photocatalytic non-oxidative dehydrogenation of ethane to ethene with near unit selectivity

Identification and characterization of tissue- and stress-specific circular RNAs (circRNAs) of tea to generate the largest tea circRNAs data repository

The study on the response of erosion and deposition evolution in the main channel of the Tarim River to water and sediment conditions under different representative years

To investigate the impact of different water and sediment conditions on the morphological shaping of the middle reaches of the Tarim River, this study establishes an erosion-deposition evolution model for the Yingbazha to Wusiman River section under the 2018 shoreline conditions using MIKE21 software and conducts validation. Five working conditions were selected for typical years of high-flow, normal-flow, and low-flow, as well as years of extreme floods and extreme droughts, to simulate the river channel’s erosion and deposition evolution under varying water and sediment conditions. By analyzing metrics such as erosion and deposition volume, depth of scour and fill, changes in the channel’s planform morphology, thalweg elevation variations, and cross-sectional changes along the river reach, the patterns of erosion and deposition evolution in this segment were systematically examined. The results indicate that: (1) Under different representative year conditions, the river was always in a net deposition state. The sediment deposition was highest in the extreme flood year (1.885 × 10 7 tons, accounting for 34% of the incoming sediment) and lowest in the extreme drought year (1.109 × 10 6 tons, accounting for 83% of the incoming sediment). The unit runoff sediment transport efficiency increased by 54% with the increase in flow. However, when the runoff exceeded the critical threshold of 3.7 × 10 9 m 3 , the increase in scour volume (+440%) far outpaced the deposition volume (+143%), revealing a critical turning point in the erosion-deposition mechanism. (2) The erosion-deposition process follows a three-stage evolutionary pattern of “deposition-scouring-redeposition”: At low flow, insufficient sediment-carrying capacity leads to continuous deposition. After surpassing the critical flow, the sediment-carrying capacity dominates scouring. Under high flow conditions, the water and sediment volume increases sharply, restarting deposition. The deep pool elevation exhibits corresponding “rise-drop-rise” periodic fluctuations. (3) Through analysis of typical cross-sections, it is shown that in extreme flood years and typical wet years, the lateral swinging of the main channel causes scouring. However, the collapse of the bank and the widening of the river channel result in increased deposition in the main channel, leading to an overall elevation of the riverbed. In typical drought years and extreme drought years, due to lower flow, water levels, and flow velocities, the erosion-deposition process only occurs within the main river channel. The research results provide a better understanding of the erosion-deposition evolution trend of the meandering section of the Tarim River’s main channel, offering scientific guidance for the future development, management, and sustainable development of the middle reach of the Tarim River.

AI bots wrote and reviewed all papers at this conference

AgentMD: Empowering language agents for risk prediction with large-scale clinical tool learning

Abstract Clinical calculators play a vital role in healthcare, but their utilization is often hindered by usability and dissemination challenges. We introduce AgentMD, a novel language agent capable of curating and applying clinical calculators across various clinical contexts. As a tool builder, AgentMD first uses PubMed to curate a diverse set of 2,164 executable clinical calculators with over 85% accuracy for quality checks and over 90% pass rate for unit tests. As a tool user, AgentMD autonomously selects and applies the relevant clinical calculators. Our evaluations show that AgentMD significantly outperforms GPT-4 for risk prediction (87.7% vs. 40.9% in accuracy). Results on 698 real-world emergency department notes confirm that AgentMD accurately computes medical risks at an individual level. Moreover, AgentMD can provide population-level insights for institutional risk management. Our study illustrates the capabilities of language agents to curate and utilize clinical calculators for both individual patient care and at-scale healthcare analytics.

Investigation of SARS-CoV-2 in various environments and patients in dental units and dental clinics

Relationship between obesity indices and cognitive function in Japanese men: A cross-sectional study

We aimed to investigate the associations among various obesity indices, including visceral (VAT) and subcutaneous adipose tissue (SAT), and cognitive function in community-dwelling Japanese men. This population-based cross-sectional study used data of 853 men who participated in the follow-up examinations of the Shiga Epidemiological Study of Subclinical Atherosclerosis. Among them, we analyzed data of 776 men who completed the Cognitive Abilities Screening Instrument (CASI) and had abdominal VAT and SAT areas measured using computed tomography. The VAT-to-SAT ratio (VSR) was calculated; participants were categorized into VSR quartiles. Using analysis of covariance, we computed crude and adjusted means of the CASI total and domain scores across VSR quartiles, adjusting for potential confounders. No significant differences were observed in total CASI scores among body mass index, VAT, or SAT quartiles. However, in the multivariable-adjusted model, participants in the lowest VSR quartile (Q1) had significantly lower CASI total scores than those in the third quartile (Q3) (Q1: 89.5, Q3: 90.9). Low VSR was independently associated with lower cognitive function in a community-based sample of middle-aged and older Japanese men. In summary, VSR may be associated with cognitive function in Japanese men, highlighting the importance of fat distribution in cognitive health and highlighting VSR as a useful indicator.

Gigantic black holes did not have starring role in early cosmic transition