Browse Articles

LGMMFusion: A LiDAR-guided multi-modal fusion framework for enhanced 3D object detection

Multi-modal data fusion plays a critical role in enhancing the accuracy and robustness of perception systems for autonomous driving, especially for the detection of small objects. However, small object detection remains particularly challenging due to sparse LiDAR points and low-resolution image features, which often lead to missed or imprecise detections. Currently, many methods process LiDAR point clouds and visible-light camera images separately, and then fuse them in the detection head. However, these approaches often fail to fully exploit the advantages of multi-modal sensors and overlook the potential for enhancing the correlation between modalities before feature fusion. To address this, we propose a novel LiDAR-guided multi-modal fusion framework for object detection, called LGMMfusion. This framework leverages the depth information from LiDAR to guide the generation of image Bird’s Eye View (BEV) features. Specifically, LGMMfusion promotes spatial interaction between point clouds and pixels before the fusion of LiDAR BEV and image BEV features, enabling the generation of higher-quality image BEV features. To better align image and LiDAR features, we incorporate a multi-head multi-scale self-attention mechanism and a multi-head adaptive cross-attention mechanism, using the prior depth information from point clouds to generate image BEV features that better match the spatial positions of LiDAR BEV features. Finally, the LiDAR BEV features and image BEV features are fused to provide enhanced features for the detection head. Experimental results show that LGMMfusion achieves 71.1% NDS and 67.3% mAP on the nuScenes validation set, while also improving the detection of small objects and enhancing the detection accuracy of most objects.

Machine learning predictions of unplanned readmissions using electronic medical records: Predictor importance across medical and surgical patient populations

Hospital readmissions prolong patient suffering and increase healthcare expenditures. While several studies have attempted to develop prediction models to reduce readmissions, most have demonstrated modest predictive accuracy. To improve upon prior approaches, we conducted an overview of systematic reviews to identify the most relevant predictor variables, then subsequently developed machine learning models in a retrospective, multisite study across eight hospitals. The patient sample comprised 200,799 inpatient stays from eligible hospitalizations, based on the Centers for Medicare and Medicaid Services (CMS) definition of unplanned readmissions within 30 days of discharge. We constructed random forest models and evaluated out-of-sample performance using the area under the receiver operating characteristic curve (AUC) across different train–test splits. The hospital-wide sample was divided into medical and surgical cohorts to investigate predictor importance across different patient populations. The average AUC score was 0.78 ± 0.01 (mean ± standard deviation [SD]). Patients’ diagnoses were the most important predictor variables (contributing 18.4% ± 0.15 to the model’s decision, mean ± standard error [SE]), followed by nursing assessments (11.2% ± 0.04, mean ± SE) and procedural information (10.8% ± 0.09, mean ± SE). Comparing medical and surgical patients, we found that medications and prior healthcare use (e.g., prior emergency encounters) were more important in the medical compared with the surgical cohort, whereas procedural information and healthcare provider information (e.g., physician caseload) were more relevant in the surgical relative to the medical cohort. In conclusion, we have established the feasibility of using Swiss electronic medical record (EMR) data to accurately predict unplanned readmissions. The reported variable importances may guide future research and inform development of clinical decision support systems aimed at reducing readmissions.

A framework to enhance the signal-to-noise ratio for quantitative fluorescence microscopy

Single-cell fluorescence characterization has gained much attention for studying the dynamics of individual cells in human diseases such as cancer. Despite the abundance of literature on quantitative fluorescence microscopy and its advantages in measuring cell-to-cell variation and spatial variation over other high-throughput instruments, there lacks a concise model that one can follow to maximize the quality of images. Here, we used the signal-to-noise ratio (SNR) model to verify marketed camera parameters and optimize microscope settings to maximize SNR for quantitative single cell fluorescence microscopy (QSFM). We determined the microscope camera’s readout noise, dark current, photon shot noise, the clock-induced charge, and validated the additive noise model for each noise source. The dark current and the clock-induced charge were both higher than reported in literature, compromising camera sensitivity. We also reduced excess background noise and improved SNR by 3-fold, by adding secondary emission and excitation filters as well as by introducing wait time in the dark before fluorescence acquisition. Additionally, our work opens new avenues for enhancing superresolution microscopy techniques such as single-molecule localization microscopy (SMLM).

Association between Systemic Immune-Inflammation Index and female breast cancer based on NHANES data (2001–2018): A cross-sectional study

Worldwide cancer statistics have shown that breast cancer dominates female cancer incidence and remains a leading cause of death. The Systemic Immune-Inflammation Index (SII) is a new prognostic indicator of systemic inflammation used to assess systemic immune-inflammatory response levels in the human body. It is associated with the prognosis of various diseases, such as malignant tumors, cardiovascular diseases, and autoimmune diseases. Although SII offers valuable information for diagnosing and predicting the risk of female breast cancer (FBC), the association between SII and FBC has not yet been analyzed. Therefore, the relationship between SII and FBC was investigated in this study. Multivariate logistic regression, model fit assessment using Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), and smoothing curve fitting were applied to examine the correlation between SII and FBC using data from the National Health and Nutrition Examination Survey (NHANES) 2001-2018. Then the stability of their association was further examined using subgroup analysis and interaction tests among populations. Results showed a positive correlation between SII and FBC in 17,044 participants with age ≥ 20 years. In the fully adjusted model, every 100-unit increase in SII was accompanied by a 3% increased odds of FBC prevalence [OR = 1.03 (95% CI: 1.01, 1.05)]. Individuals in the highest quartile of SII exhibited 44% increased odds of FBC prevalence than those in the lowest quartile [OR = 1.44 (95% CI: 1.11, 1.88)]. Model fitness assessment using AIC and BIC criteria demonstrated that multivariable-adjusted models exhibited better fit compared to unadjusted models for both continuous and categorical SII specifications. Receiver Operating Characteristic (ROC) curve analysis demonstrated that SII exhibited excellent diagnostic capability for breast cancer, with the area under the ROC curve (AUC) of 0.816 (95% CI: 0.801–0.831), comparable to NLR (AUC = 0.816) and neutrophil counts (AUC = 0.815). In disease-specific performance comparison, SII’s predictive ability for breast cancer (AUC = 0.816) was slightly superior to that for hypertension (AUC = 0.799), with the difference being statistically significant (P = 0.0407). Our findings confirmed that SII was a promising biomarker associated with FBC prevalence, and it may provide valuable insights into early screening and personalized treatment strategies.

Analysis of the quadruple evolutionary game of ecosystem service payment empowered by farmers’ cooperatives

Payments for Ecosystem Services (PES) are essential for ecosystem restoration and promoting sustainable economic development. Farmer cooperatives serve as key intermediaries in implementing PES. This study constructs a game model involving four stakeholders—local government, enterprises, cooperatives, and cooperative members—while considering their bounded rationality. Numerical simulations using Matlab are conducted to test the stability and effectiveness of equilibrium strategies among these stakeholders. The results show that no matter which direction the system evolves, the strategy of farmers ‘ cooperatives is to produce ecological products, that is, to effectively promote the realization of the value of ecological products, farmers ‘ cooperatives need to actively participate in the production of ecological products. The payment model of ecosystem services can not only provide practical basis and motivation for the payment of ecosystem services, promote the optimization and improvement of relevant mechanisms, but also enhance the market competitiveness of farmers ‘cooperatives and enhance their brand value.

The influence of perceived threat on the motive attribution asymmetry bias for groups in conflict

Previous research shows higher perceived threat is related to more intergroup bias, usually via greater ingroup positivity. Newer research has identified the Motive Asymmetry Attribution Bias in which ingroup and outgroup members make very different explanations for the motives about why their groups are in conflict. We were interested in this Motive Asymmetry Bias and its relationship to perceived threat with groups in conflict, so we designed two studies to investigate it cross-sectionally (Study 1) and longitudinally (Study 2). We recruited samples of American Republicans and Democrats to complete an online survey measuring perceived threat and Motive Asymmetry Bias. Regression analyses indicated that perceived threat was not related to ratings of one’s own party; however, higher perceived threat was related to more negative ratings of the other party. This discovery is important to help inform different ways to intervene to improve intergroup relations, especially for groups in conflict.

RFK Jr demanded a vaccine study be retracted — the journal said no

Study on the alteration of gut microbiota in ovariectomized rats and the impact of estrogen intervention over a 3-month period

Postmenopausal osteoporosis (PMOP) is a common primary osteoporosis. With the aging of the population, it is becoming a major disease that endangers health and quality of life. The purpose of this study was to explore the effect of gut microbiota on PMOP by observing the changes in the levels of estradiol, bone density, and gut microbiota diversity in rats after 3 months of OVX surgery. 60 female SD rats were randomly divided into four groups: baseline group (6 rats), sham-operated group (18 rats), model group (18 rats), and estrogen-treated group (18 rats). The ovariectomy model of postmenopausal osteoporosis was established by performing bilateral ovariectomy. After surgery, 6 rats from each group were randomly selected for sacrifice every 30 days and subsequent assessment. At the end of 90 days, all rats were sacrificed for evaluation of body weight, bone mineral density (BMD), tissue mineral density (TMD), trabecular bone parameters, femoral bone morphology, hormone levels, and gut microbiota diversity. The analysis revealed that OVX led to a decrease in BMD, TMD, and serum estradiol levels in rats, and increased TNF-α levels. The bone micro-architecture and tissue morphology were also changed, with trabecular fractures, thinning, and decreased numbers. Meanwhile, there was also a shift in the diversity of gut microbiota. The administration of estrogen could potentially ameliorate these alterations. Overall, OVX leads to a persistent decline in estrogen levels in rats. This results in gradual bone loss, which is related to gut microbiota imbalance.

US Supreme Court allows NIH to cut $2 billion in research grants

Implementing dynamic distance-time conversion factors through real-time traffic data for enhancing urban mobility and service accessibility in South Korea

This study introduces an expanded methodology for smart regional planning tailored to improve public service accessibility. We develop a city-level distance-time conversion factor (DCF) that utilizes regional characteristics to offer more intuitive estimates of travel times and distances in public service planning. This approach integrates three key variables: road network distances, Euclidean straight-line distances, and minimum travel times derived from both speed limits and actual traffic speeds. The DCF, formulated from the circuity factor (CF) and the delay factor (DF), identifies areas with elevated DCF values, particularly in major metropolitan areas. These metrics serve as critical indicators for densely populated areas, marking a substantial improvement over traditional methods of uniform location planning. Our analysis addresses underdevelopment and population density challenges, underscoring the need for adaptable planning strategies. By incorporating real-time traffic data, the DCF provides insights crucial for strategically developing public infrastructure in high-demand regions. This research enhances the existing smart public service planning frameworks, emphasizing the significance of regional-specific strategies. Ultimately, our findings advocate for a tailored approach to infrastructure development, aiming to create more efficient and responsive public services.

Retraction: Investigating capital flight in South Asian countries: The dual influence of terrorism and corruption

Reaching a cell monolayer at the end of hiPSC differentiation enhances neural crest lineage commitment

Neural crest stem cells (NCSCs) compose a highly migratory, multipotent, stem cell population arising from the neural plate border of the embryonic ectoderm. Investigating the development of NCSCs is critical in understanding both embryonic development and abnormal events that underlie neurocristopathies. Suggested seeding densities in in vitro human induced pluripotent stem cells (hiPSCs) differentiation protocols, varying between 10,000 cells/cm2 and 200,000 cells/cm2, demonstrate a lack of consensus on the optimal conditions to obtain NCSCs. Aiming to maximize the differentiation efficiency of hiPSCs towards the NCSCs lineage, we investigated the effect of the initial seeding density on NCSCs lineage commitment, both in fibroblast- and human peripheral blood mononuclear cell (PBMC)-derived hiPSCs. Cultures were characterized with gene and protein expression analysis assessing stemness (OCT3/4 and NANOG), neural crest identity (SNAI2 and SOX10) and neuroectoderm identity (PAX6 and SOX1). We demonstrate that reaching a confluent monolayer of cells by the end of the differentiating protocol is crucial to obtaining NCSCs from hiPSCs. To achieve this, our results indicated 17,000 cells/cm2 is the optimal initial seeding density. Under this protocol, a confluent monolayer was reached after 8 days of differentiation and an average of 89% SOX10 positive cells were obtained. The fold change of SNAI2 and SOX10 expression was 11-fold and 17-fold higher, respectively, in cultures seeded with 17,000 cells/cm2, compared to the highest tested density of 200,000 cells/cm2. In contrast, seeding 200,000 cells/cm2 induced neuroectoderm-like cells, confirmed by an average of 45% of cells marking positive for PAX6. With this work, we demonstrate the importance of achieving cellular confluency during NCSCs differentiation.

Does dark energy spawn from black holes? Could be a bright idea

The “most beautiful place” where “it’s not possible to live”: A qualitative study of relational well-being in an area of climate vulnerability, Bangladesh

Purpose Climate change is the greatest global health threat of the 21st century, but little is known about well-being in climate vulnerable populations. We investigate how well-being is shaped by common and unique stressors in an area of climate vulnerability in Bangladesh. Methods We present findings from 60 semi-structured in-depth interviews. We inductively analyzed our data following a Reflexive Thematic Analysis approach and then applied a Relational Well-being (RWB) framework. Results We found that well-being was influenced negatively by factors such as financial worries, forced migration, social pressure, and natural disasters. Well-being was influenced positively by factors such as financial satisfaction, voluntary migration, social support, and place attachment. Conclusions Using relational well-being as a conceptual lens allowed us to explore the dynamism and complexity of factors shaping well-being that were partly specific to the local context and partly rooted in wider societal and global structures. Policies which aim to improve the well-being of climate vulnerable populations should consider relational well-being as a conceptual tool to leverage locally available informal resources, such as suppotive reciprocal relationships.

Validation in diabetic rats of a fully automated insulin delivery system based on impulsive offset-free MPC control

The development of an impulsive automated insulin delivery system (i-AiDS) for type 1 diabetes mellitus aims to provide real-time blood glucose regulation with minimal human intervention. This study presents the validation of an offset-free impulsive zone model predictive control strategy designed to cope with external disturbances such as meal intake and plant-model mismatch in a diabetic rat model. Fourteen male Wistar rats induced diabetes with streptozotocin were monitored using an continuous glucose monitoring and regulated by delivering insulin with a customized low-cost pump. After acquiring diabetes condition, the procedure for installing the devices in the rat is carried out. During the first day, manual insulin injections are made by the pump, the glucose response is recorded by the interface and an off-line parametric estimation is executed. Based on the parameters found, simulations are used for the first tuning of the controller and the estimator. During the second day, the parameters of the model and the control are tested and adjusted. Finally, on the third day, a 72-hour test of the impulsive begins in full autonomous mode. Results showed that the controller achieved an average of 83.4% of the time within the target range of 80-180 mg/dL, with no severe hypoglycemic or hyperglycemic events. The median absolute relative difference between model predictions and actual sensor data was 24.66%, indicating the presence of plant-model mismatch that was effectively handled by the controller. Peak hyperglycemic events reached 320 mg/dL, but were regulated within 50 minutes, while mild hypoglycemic events occurred in 3.62 ± 1.8 cases per subject. The study demonstrates the efficacy of the controller in managing unannounced carbohydrate intake and physiological disturbances in a real-world preclinical environment. These findings provide a foundation for future clinical trials, emphasizing the importance of in vivo validation of control strategies to refine the i-AiDS for human use. Improvements in model accuracy and dynamic parameter tuning could further enhance performance, particularly in longer experimental periods.

Event-driven architecture and intelligent decision tree facilitated sustainable trade activity monitoring model design

This paper introduces a groundbreaking monitoring model tailored for sustainable trade activity surveillance, which synergistically integrates event-driven architecture with an intelligent decision tree. Confronting the constraints of conventional trade monitoring approaches that falter in adapting to the intricate and ever-changing market landscape, our model innovatively establishes an efficient, adaptable, and sustainable monitoring framework. By embedding an intelligent decision tree, it enables dynamic resource allocation, thereby optimizing operational efficacy. Initially, we devise rules that align data injection and processing velocities, ensuring expedient data processing. Subsequently, we implement an optimal binary tree decision-making algorithm, grounded in dynamic programming, to achieve precise allocation of elastic resources within data streams, significantly bolstering resource utilization. Throughout the monitoring continuum, the model employs intelligent agents to assess resource status in real-time and dynamically adjusts resource allocation strategies triggered by events, prioritizing the seamless execution of pivotal trade activities. Empirical findings underscore the model’s superiority across critical metrics, including data accumulation efficiency, processing latency, resource utilization, and throughput. Specifically, it attains an average data accumulation value of 15.46, curtails latency by 14.67%, achieves an average resource utilization of 60.29%, and registers a throughput of 336.5 Mbps. Consequently, the model not only exhibits rapid responsiveness to market fluctuations and curtails resource energy consumption but also fosters a harmonious equilibrium between economic gains and environmental preservation, ensuring the uninterrupted operation of trade activities.

Impact of health promotion strategies on HPV vaccination uptake: A descriptive epidemiological study (2019–2024)

Introduction/objectives The effectiveness of vaccines depends not only on resource availability but also on broad public acceptance and the uptake of widely accessible vaccines. A vaccination campaign is a strategically coordinated initiative designed to enhance vaccine coverage within a specific population. This study evaluated the impact of health promotion strategies (HPS), including social marketing and education, on human papillomavirus (HPV) vaccine uptake among adolescents aged 9–19 years in City of Novi Sad, Serbia (population ~300,000). Since 2020, efforts have transitioned from individual to organized, publicly funded initiatives. Materials and methods A descriptive epidemiological study was conducted using anonymized data from the electronic immunization registry (2019–2024). Data on HPS implementation were obtained from the Institute of Public Health of Vojvodina. Statistical analyses included Mann–Whitney U test, ANOVA, correlation, and multiple regression with time lag. Results and discussion From 2019 to 2024, 6,395 adolescents received the HPV vaccine, with sharp increases in 2023–2024. Among the different strategies analyzed, health promotion through educational media content delivered by physicians (TV, radio, social networks) and the implementation of the “Open Door” initiative had the most consistent and positive association with increased vaccination coverage. Vaccination uptake was strongly associated with frequent and accessible promotional activities, especially the “Open Door” initiative (r =.668, p <.01; β =.687, p <.001). Media activities showed moderate effects (r =.270, p <.05). Educational activities show only a weak or non-significant correlation with vaccination rates, except for their modest association with website updates (r =.282, p <.05). Conclusions Accessible, action-oriented interventions, particularly “Open Door” days, were the most effective strategy for increasing adolescent HPV vaccination. Social marketing combining convenience and multi-channel communication significantly enhanced uptake. These findings support the implementation of targeted, barrier-reducing public health strategies to improve vaccine coverage.

Peer reviewers more likely to approve articles that cite their own work

The disease and economic burden of notified and underestimated Campylobacter enteritis cases and associated sequelae in Germany

Background According to surveillance data, Campylobacter enteritis (CE) has been the most frequently notified bacterial gastrointestinal disease in Germany and Europe for many years. Presumably, the total number of cases is underestimated because an unknown number of cases is not diagnosed and some diagnosed cases are not reported in the surveillance system. The aim of this study was to estimate the disease and economic burden of CE and its related sequelae in Germany. Methods The disability-adjusted life years (DALY) as well as the direct and indirect costs associated with the five-year (2018–2022) mean number of CE cases and related sequelae were estimated in a Monte Carlo simulation. The age- and gender-specific characteristics were integrated where possible. The underestimated CE cases were quantified by reconstructing the surveillance pyramid using age group-specific health care seeking parameters. Results The estimated incidence rate was 553 CE cases (95%-CI: 551–555 cases) per 100,000 inhabitants per year. This corresponds to 7.7 underestimated cases per notified case. Underestimation was lowest in the age group <5 years and highest in the age group 15–29 years. The notified plus underestimated CE cases and associated sequelae resulted in a loss of 6,764 DALY (95%-CI: 6,689-6,839 DALY), 88% of which were due to sequelae. The total economic burden amounted to 263.5 million Euros (95%-CI: 262.5–264.4 million Euros). Approximately 25% of the total DALY and costs were attributable to the notified cases. Conclusions The results suggest a substantial burden due to CE, both in terms of DALY and costs, in Germany. Especially the high number of underestimated cases and associated sequelae contribute to the health and economic burden – although some remaining uncertainties cannot be ruled out. By using age-specific multipliers to determine the underestimated cases, age-related differences in DALY and cost of illness can be accounted for, thereby preventing an overestimation of the total burden.

CAT: Class-aware adaptive-thresholding for robust semi-supervised domain generalization

Domain Generalization (DG) seeks to transfer knowledge from multiple source domains to unseen target domains, even in the presence of domain shifts. Achieving effective generalization typically requires a large and diverse set of labeled source data to learn robust representations that can generalize to new, unseen domains. However, obtaining such high-quality labeled data is often costly and labor-intensive, limiting the practical applicability of DG. To address this, we investigate a more practical and challenging problem: semi-supervised domain generalization (SSDG) under a label-efficient paradigm. In this paper, we propose a novel method, CAT, which leverages semi-supervised learning with limited labeled data to achieve competitive generalization performance under domain shifts. Our method addresses key limitations of previous approaches, such as reliance on fixed thresholds and sensitivity to noisy pseudo-labels. CAT combines adaptive thresholding with noisy label refinement techniques, creating a straightforward yet highly effective solution for SSDG tasks. Specifically, our approach uses flexible thresholding to generate high-quality pseudo-labels with higher class diversity while refining noisy pseudo-labels to improve their reliability. Extensive experiments on multiple benchmark datasets demonstrate the superior performance of our method, with improvements of 3.45% on PACS, 9.47% on OfficeHome, and 10.90% on miniDomainNet datasets, highlighting its effectiveness in achieving robust generalization under domain shifts.