Browse Articles
Discover research articles across all indexed journals
Feasibility study of texture-based machine learning approach for early detection of neonatal jaundice
Exploration of stagnation-point flow of Reiner–Rivlin fluid originating from the stretched cylinder for the transmission of the energy and matter
A new cut-off value of FRAX tools as an osteoporosis screening tool for Thai geriatric population
Abstract Identifying osteoporosis in geriatric populations is essential for fragility fracture prevention. While dual-energy X-ray absorptiometry (DXA) remains the gold standard for diagnosing osteoporosis, its availability and cost for mass screening are limited. This study aims to determine an effective fracture risk assessment tool (FRAX) cut-off value for screening osteoporosis in the Thai geriatric population. The demographic data, FRAX hip fracture (HF), major osteoporotic fracture (MOF), and Bone mineral density (BMD) of community-dwelling Thai adults aged ≥ 60 years, conducted between March 2021 to August 2022 were analyzed. Osteoporosis is defined as a BMD T-score ≤ − 2.5. The accuracy of FRAX in identifying osteoporosis was assessed using the area under the receiver operating characteristic curve (AUC). Among 2991 participants (average age 69.2 ± 6.5 years), the discriminative ability was acceptable for both FRAX hip fracture (HF) (AUC = 0.75) and major osteoporotic fracture (MOF) (AUC = 0.72). A cut-off value of 1.5 for FRAX HF and 4.5 for FRAX MOF demonstrated excellent sensitivity (90.4%) and a high negative predictive value (89.7%) in osteoporosis detection. This study identifies FRAX cut-off values that can effectively screen for high-risk osteoporosis in the Thai geriatric population and suggests that FRAX could be a valuable tool for initial osteoporosis screening in Thai seniors.
Improving Malaria diagnosis through interpretable customized CNNs architectures
Abstract Malaria, which is spread via female Anopheles mosquitoes and is brought on by the Plasmodium parasite, persists as a serious illness, especially in areas with a high mosquito density. Traditional detection techniques, like examining blood samples with a microscope, tend to be labor-intensive, unreliable and necessitate specialized individuals. To address these challenges, we employed several customized convolutional neural networks (CNNs), including Parallel convolutional neural network (PCNN), Soft Attention Parallel Convolutional Neural Networks (SPCNN), and Soft Attention after Functional Block Parallel Convolutional Neural Networks (SFPCNN), to improve the effectiveness of malaria diagnosis. Among these, the SPCNN emerged as the most successful model, outperforming all other models in evaluation metrics. The SPCNN achieved a precision of 99.38 $$\pm$$ 0.21%, recall of 99.37 $$\pm$$ 0.21%, F1 score of 99.37 $$\pm$$ 0.21%, accuracy of 99.37 ± 0.30%, and an area under the receiver operating characteristic curve (AUC) of 99.95 ± 0.01%, demonstrating its robustness in detecting malaria parasites. Furthermore, we employed various transfer learning (TL) algorithms, including VGG16, ResNet152, MobileNetV3Small, EfficientNetB6, EfficientNetB7, DenseNet201, Vision Transformer (ViT), Data-efficient Image Transformer (DeiT), ImageIntern, and Swin Transformer (versions v1 and v2). The proposed SPCNN model surpassed all these TL methods in every evaluation measure. The SPCNN model, with 2.207 million parameters and a size of 26 MB, is more complex than PCNN but simpler than SFPCNN. Despite this, SPCNN exhibited the fastest testing times (0.00252 s), making it more computationally efficient than both PCNN and SFPCNN. We assessed model interpretability using feature activation maps, Gradient-weighted Class Activation Mapping (Grad-CAM) and SHapley Additive exPlanations (SHAP) visualizations for all three architectures, illustrating why SPCNN outperformed the others. The findings from our experiments show a significant improvement in malaria parasite diagnosis. The proposed approach outperforms traditional manual microscopy in terms of both accuracy and speed. This study highlights the importance of utilizing cutting-edge technologies to develop robust and effective diagnostic tools for malaria prevention.
Multiple respiratory assessment and thresholds for noninvasive ventilation in adult patients with spinal muscular atrophy
Clinical and pathological risk factors for postencephalitic epilepsy after herpes simplex virus-1 encephalitis in children
Evaluating the impact of self myofascial release and traditional recovery strategies on volleyball athletes using thermal imaging and biochemical assessments
Impacts of land use change on carbon storage in the Guangxi Beibu Gulf Economic Zone based on the PLUS-InVEST model
Measurements of face mask’s capability to block ionizing radiation
Abstract Experts suggest wearing a face mask during a radiation emergency if it is impossible to get inside immediately and high-level protective respirators are unavailable. This study quantitatively investigated seven face mask materials’ ability to block radioactive alpha and beta radiation. Rayon fiber, pure cotton, paper fiber, polyester fiber, nonwoven fiber, advanced nonwoven fiber, and N95 were examined. The results suggest that the abovementioned mask materials can block more than 90% of alpha particles. Rayon fiber, polyester fiber, and N95 can block almost all radioactive alpha particles. On the other hand, the measurements suggest that all tested materials could not effectively block beta particles. Polyester fiber and N95 block more than 10% of beta particles, which outperform other mask materials. In addition, the results imply that the electret fiber might help block beta particles. This study suggests that wearing a relatively thick polyester or N95 mask may be a better choice than wearing a thin nonwoven mask to prevent inhaling alpha and beta particles during a radiation emergency.
Fast, smart, and adaptive: using machine learning to optimize mental health assessment and monitor change over time
Identifying mosquito plant hosts from ingested nectar secondary metabolites
Abstract Establishing how plants contribute food and refuge to insects can be challenging for small species that are difficult to observe in their natural habitat, such as disease vectoring mosquitoes. Currently indirect methods of plant-host identification rely on DNA sequencing of ingested plant material but are often unsuccessful for small insects that feed primarily on plant sugars or have little contact with plant cells. Here we developed an innovative approach to determine species-specific phytophagy by detecting taxon-specific plant secondary metabolites (PSMs) in nectar. Two mosquito species were exposed to three PSMs, each present in the nectar of a known plant host, firstly from dosed sucrose solutions and secondly from flowers. Both experiments yielded high rates of PSM detection in mosquitoes using liquid chromatography-mass spectrometry (LC-MS). PSMs were consistently detected in mosquitoes up to 8 h post-ingestion. In experiments consisting of two or three plant species, multiple PSMs from different host plants could be detected. These positive results demonstrate that PSMs could be useful indicators of insect plant-hosts selection in the wild. With expanded knowledge of nectar-based PSMs across a landscape, improved knowledge of plant-host relationships could be achieved where direct observations in their natural habitat are lacking. Increasing understanding of vector insect ecology will have an important role in tackling vector-borne disease.
Design and motion analysis of a coal mine robot with variable wheel diameter
Diagnosis-related groups study of uterine leiomyoma patients based on E-CHAID
Tendency to overeat predicts an elevated body mass index trajectory across school-age years
Abstract Overeating is a complex appetite trait, cross-sectionally linked to an elevated weight in children. However, little is known about longitudinal associations. Therefore, we studied how a tendency towards overeating predicts weight development between 8 and 16 years of age. In this study among 4517 children from the Finnish Health in Teens cohort, parents reported their child’s tendency to overeat when children were on average 11.2 (SD 0.8) years old. Children were then categorised as overeating, possibly overeating, or not overeating. Height and weight measurements from two data collection periods were combined with growth data from a national health register, and age- and sex-standardised body mass index z-scores (BMIz) were calculated using the International Obesity Task Force reference. Children also reported their lifestyle factors, including food consumption, physical activity, screen time, and sleep patterns. We examined the association between overeating and BMIz using a linear mixed model, adjusting for age, sex, specific food consumption frequencies, physical activity, screen time, and sleep duration. We further analysed whether associations differed by age, food consumption frequencies, or physical activity. The average BMIz in the overeating group was 1.18 (95% CI 1.10─1.26) units higher compared with those without overeating, but remained stable as age increased. Among those without overeating, BMIz increased 0.043 units per year of age (p < 0.001). Physical activity, but not food consumption, modified the association between overeating and BMIz (p for interaction = 0.038). In the lowest third of physical activity (≤ 5 h/week), BMIz was 1.28 units higher (95% CI 1.15─1.41) in the overeating compared with the no overeating group, while in the highest third (≥ 9 h/week) the effect size was 1.08 (95% CI 0.93─1.24). In conclusion, children with a parent-reported tendency to overeat exhibited an elevated, but stable mean BMIz across adolescence. Public health programmes tackling the obesity epidemic should consider the differences in appetite self-regulation among children.
SCCA-YOLO: A Spatial and Channel Collaborative Attention Enhanced YOLO Network for Highway Autonomous Driving Perception System
Polyextremotolerant, opportunistic, and melanin-driven resilient black yeast Exophiala dermatitidis in environmental and clinical contexts
Enterovirus D68 infection in cotton rats results in systemic inflammation with detectable viremia associated with extracellular vesicle and neurologic disease
Association between maternal anemia during pregnancy with low birth weight their infants
Semi-supervised tissue segmentation from histopathological images with consistency regularization and uncertainty estimation
Abstract Pathologists have depended on their visual experience to assess tissue structures in smear images, which was time-consuming, error-prone, and inconsistent. Deep learning, particularly Convolutional Neural Networks (CNNs), offers the ability to automate this procedure by recognizing patterns in tissue images. However, training these models necessitates huge amounts of labeled data, which can be difficult to come by due to the skill required for annotation and the unavailability of data, particularly for rare diseases. This work introduces a new semi-supervised method for tissue structure semantic segmentation in histopathological images. The study presents a CNN based teacher model that generates pseudo-labels to train a student model, aiming to overcome the drawbacks of conventional supervised learning approaches. Self-supervised training is used to improve the teacher model’s performance on smaller datasets. Consistency regularization is integrated to efficiently train the student model on labeled data. Further, the study uses Monte Carlo dropout to estimate the uncertainty of proposed model. The proposed model demonstrated promising results by achieving an mIoU score of 0.64 on a public dataset, highlighting its potential to improve segmentation accuracy in histopathological image analysis.