Reconstructed Graph Neural Network With Knowledge Distillation for Lightweight Anomaly Detection.

The proliferation of Internet-of-Things (IoT) technologies in modern smart society enables massive data exchange for offering intelligent services. It becomes essential to ensure secure communications while exchanging highly sensitive IoT data efficiently, which leads to high demands for lightweight models or algorithms with limited computation capability provided by individual IoT devices. In this study, a graph representation learning model, which seamlessly incorporates graph neural network (GNN) and knowledge distillation (KD) techniques, named reconstructed graph with global-local distillation (RG-GLD), is designed to realize the lightweight anomaly detection across IoT communication networks. In particular, a new graph network reconstruction strategy, which treats data communications as nodes in a directed graph while edges are then connected according to two specifically defined rules, is devised and applied to facilitate the graph representation learning in secure and efficient IoT communications. Both the structural and traffic features are then extracted from the graph data and flow data respectively, based on the graph attention network (GAT) and multilayer perceptron (MLP) techniques. These can benefit the GNN-based KD process in accordance with the more effective feature fusion and representation, considering both structural and data levels across the dynamic IoT networks. Furthermore, a lightweight local subgraph preservation mechanism improved by the graph attention mechanism and downsampling scheme to better utilize the topological information, and a so-called global information alignment defined based on the self-attention mechanism to effectively preserve the global information, are developed and incorporated in a refined graph attention based KD scheme. Compared with four different baseline methods, experiments and evaluations conducted based on two public datasets demonstrate the usefulness and effectiveness of our proposed model in improving the efficiency of knowledge transfer with higher classification accuracy but lower computational load, which can be deployed for lightweight anomaly detection in sustainable IoT computing environments.

Personalized Federated Graph Learning on Non-IID Electronic Health Records.

Understanding the latent disease patterns embedded in electronic health records (EHRs) is crucial for making precise and proactive healthcare decisions. Federated graph learning-based methods are commonly employed to extract complex disease patterns from the distributed EHRs without sharing the client-side raw data. However, the intrinsic characteristics of the distributed EHRs are typically non-independent and identically distributed (Non-IID), significantly bringing challenges related to data imbalance and leading to a notable decrease in the effectiveness of making healthcare decisions derived from the global model. To address these challenges, we introduce a novel personalized federated learning framework named PEARL, which is designed for disease prediction on Non-IID EHRs. Specifically, PEARL incorporates disease diagnostic code attention and admission record attention to extract patient embeddings from all EHRs. Then, PEARL integrates self-supervised learning into a federated learning framework to train a global model for hierarchical disease prediction. To improve the performance of the client model, we further introduce a fine-tuning scheme to personalize the global model using local EHRs. During the global model updating process, a differential privacy (DP) scheme is implemented, providing a high-level privacy guarantee. Extensive experiments conducted on the real-world MIMIC-III dataset validate the effectiveness of PEARL, demonstrating competitive results when compared with baselines.

Comparison and Evaluation of the Accuracy for Thoracic and Lumbar Pedicle Screw Fixation in Early-Onset Congenital Scoliosis Children.

BACKGROUND: Compared to adult scoliosis, correcting scoliosis in children often presents greater challenges. This is attributed to two key factors. Firstly, it involves accounting for the growth potential of children. Secondly, the thinner pedicles in children can complicate screw insertion, particularly when dealing with existing deformities. The utilization of intraoperative navigation technology offers a modest improvement in the precision of screw placement but does come with the drawback of increased radiation exposure. The aim of this study is to investigate and assess the accuracy of manually inserting pedicle screws in the thoracic and lumbar spine to rectify deformities in children with early-onset congenital scoliosis. METHODS: In this retrospective study, 26 hospitalized patients diagnosed with early-onset congenital scoliosis between December 2014 and December 2019 were selected. The cohort comprised 16 boys and 10 girls, aged between 2 and 10 years, with an average age of 4.68 ± 2.42 years. Pedicle screw fixation was applied in the segment spanning from T1 to L5. Pedicle screws were inserted manually, guided by the positioning of the C-arm and anatomical markers. The assessment of pedicle screw placement was based on the distance of penetration into the medial, lateral, or anterior bone cortex of the vertebral body, including the pedicle, categorized into three grades: Grade 1 (placement <2 mm), Grade 2 (placement between 2-4 mm), and Grade 3 (placement >4 mm). Grade 1 indicates accurate pedicle screw placement, while Grades 2 and 3 signify abnormal pedicle screw placement. Complications related to pedicle screw insertion were also recorded, both during and after the surgical procedure. RESULTS: A total of 173 pedicle screws were inserted in this study, with an average of 6.65 screws per patient. Accurate screw placement was achieved in 143 cases (82.7%), while 30 pedicle screws were found to be abnormal. Among the abnormal screws, 24 were categorized as Grade 2 (13.9%), and 6 as Grade 3 (3.5%). Grade 2 abnormalities were distributed across 20 thoracic vertebrae and 4 lumbar vertebrae, while Grade 3 abnormalities affected 5 thoracic vertebrae and 1 lumbar vertebra. When comparing the lumbar and thoracic vertebral regions, a significant difference in the rate of abnormal screw placement was observed (χ2 = 5.801, p < 0.05). The rate of abnormal screw placement was higher in the thoracic vertebral region with abnormal vertebral bodies than in the lumbar vertebral regions. Furthermore, a statistically significant difference in the rate of abnormal screw placement was found between the concave and convex sides (χ2 = 23.047, p < 0.05). The concave side of the abnormal vertebral body had a higher rate of abnormal screw placement (55.6%, 15/27) compared to the convex side (20.1%, 7/34), and this difference was statistically significant (p < 0.05). Throughout the intraoperative and postoperative follow-up period, spanning from 12 to 56 months, only one patient experienced issues with wound healing, and no complications related to pedicle screw placement occurred, such as hemopneumothorax, pedicle fracture, nerve root injury, aortic injury, screw loosening, pullout or breakage, or spinal cord injury. CONCLUSIONS: In children under 10 years of age with early-onset congenital scoliosis, the freehand placement of thoracic and lumbar pedicle screws demonstrates a high level of accuracy. Moreover, complications associated with pedicle screw insertion are infrequent following surgery. It is advisable to exercise caution when placing pedicle screws in thoracic vertebral bodies and morphologically abnormal vertebral bodies, with particular attention to the concave side when screw placement is required in these regions.

Modified Bilhaut-Cloquet procedure for Wassel type III-IV polydactyly.

BACKGROUND: To investigate the functional and aesthetic results of a new modified Bilhaut-Cloquet procedure for the treatment of Wassel type III-IV thumb polydactyly. METHODS: Thirteen patients with Wassel type III-IV thumb polydactyly who visited the Department of Orthopedics of Hebei Provincial Children's Hospital from 2019 to 2022 were selected. The surgical procedure involved a modified Bilhaut-Cloquet surgery, where two-thirds of the distal part of the dominant finger was retained as the p body of the reconstructed thumb. The triangular bone block of the ablated distal thumb that did not contain the epiphysis and articular cartilage was sutured and fixed, and the neurovascular flap of the ablated distal thumb was used as an augmenting segment of the reconstructed thumb, with the nail bed and nail matrix exquisitely sutured. The evaluation performed according to the Japanese Society for Surgery of the Hand (JSSH) system. RESULTS: All 13 children showed bone healing, no wound infection, nonunion, or deformity healing. None of the children showed a significant reduction in the active and passive mobility of the thumb postoperatively compared with preoperatively. Postoperative evaluation was performed based on the JSSH score, with a mean of 17.15 points (14-19 points), with 11 children rated as excellent and two as good. No severe nail ridges, nail gaps, or nail split deformities of the thumb were observed postoperatively. Postoperative metacarpophalangeal and interphalangeal joint movements were not reduced compared with preoperative movements. All parents were satisfied with the appearance and function of the reconstructed thumb. CONCLUSION: The modified Bilhaut-Cloquet procedure designed in this study was satisfactory for Wassel type III-IV thumb polydactyly without affecting the stability of the interphalangeal joints and preserving joint mobility. The postoperative thumb has a comparable circumference and nail width and was cosmetically and functionally satisfactory, especially for the asymmetric two thumbs, which achieved favorable results.

Logistic regression analysis of multiple rapid admission hematological indicators to determine the risk of osteomyelitis in infected children.

To analyze the differences of multiple rapid admission hematological indicators between children with acute osteomyelitis (AO) and children with other orthopedic infectious diseases and clarify the characteristics of admission inspection hematological indicators of children with AO. Retrospective analysis of this pilot study was proceeded on 144 children with limbs infectious diseases, who were treated in our hospital. According to their final diagnosis, they were divided into osteomyelitis group (n = 57) and non-osteomyelitis group (n = 87). Case data were collected, including sex, age, body temperature, white blood cell (WBC), C-reactive protein (CRP), etc. The differences in these indexes between the two groups of patients were compared, and then, the index with significant differences was selected for univariate and multivariate logistic regression analysis. There were significant differences between the two groups in age, body temperature, CRP, ESR, fibrinogen, total bilirubin, alanine aminotransferase, aspartate aminotransferase (AST), glutamyl transpeptidase, creatinine, PCT, albumin (ALB), and ALB globulin ratio (A/G) (P < 0.05). The results of univariate and multivariate logistic regression analysis showed that the age of ≥5 years (4.592, 1.711-12.324), WBC (>1.5 × 109/L) (0.271, 0.102-0.718), ESR (>50 mm/h) (6.410, 2.291-17.936), PCT (>0.06 µg/L) (3.139, 1.066-9.243), and AST (>40 U/L) (11.174, 1.718-72.666) was an independent risk factor of AO in children with orthopedic infectious diseases (P < 0.05). For newly admitted children with orthopedic infectious diseases, if the age ≥ 5 years, WBC ≤ 1.5 × 109/L, ESR > 50 mm/h, PCT > 0.06 µg/L, and AST > 40 U/L, the occurrence of AO should be alerted.

Information Theoretic Learning-Enhanced Dual-Generative Adversarial Networks With Causal Representation for Robust OOD Generalization.

Recently, machine/deep learning techniques are achieving remarkable success in a variety of intelligent control and management systems, promising to change the future of artificial intelligence (AI) scenarios. However, they still suffer from some intractable difficulty or limitations for model training, such as the out-of-distribution (OOD) issue, in modern smart manufacturing or intelligent transportation systems (ITSs). In this study, we newly design and introduce a deep generative model framework, which seamlessly incorporates the information theoretic learning (ITL) and causal representation learning (CRL) in a dual-generative adversarial network (Dual-GAN) architecture, aiming to enhance the robust OOD generalization in modern machine learning (ML) paradigms. In particular, an ITL-and CRL-enhanced Dual-GAN (ITCRL-DGAN) model is presented, which includes an autoencoder with CRL (AE-CRL) structure to aid the dual-adversarial training with causality-inspired feature representations and a Dual-GAN structure to improve the data augmentation in both feature and data levels. Following a newly designed feature separation strategy, a causal graph is built and improved based on the information theory, which can enhance the causally related factors among the separated core features and further enrich the feature representation with the counterfactual features via interventions based on the refined causal relationships. The ITL is incorporated to improve the extraction of low-dimensional feature representations and learn the optimized causal representations based on the idea of "information flow." A dual-adversarial training mechanism is then developed, which not only enables the generator to expand the boundary of feature distribution in accordance with the optimized feature representation from AE-CRL, but also allows the discriminator to further verify and improve the quality of the augmented data for OOD generalization. Experiment and evaluation results based on an open-source dataset demonstrate the outstanding learning efficiency and classification performance of our proposed model for robust OOD generalization in modern smart applications compared with three baseline methods.

Effects of sugarcane variety and nitrogen application level on the quality and aerobic stability of sugarcane tops silage.

To better understand the effects of sugarcane variety and nitrogen application level on silage, we analyzed the fermentation quality, microbial community dynamics, and aerobic exposure of sugarcane tops silage from three sugarcane varieties (B9, C22, and T11) treated with three levels of nitrogen (0, 150, and 300 kg/ha urea). After 132 days of silage, the sugarcane tops silage produced from variety B9, with strong nitrogen fixation ability, treated with nitrogen had the highest crude protein (CP) contents, pH, and yeast counts (P < 0.05), as well as the lowest Clostridium counts (P < 0.05), and the CP increased with increasing N application level (P < 0.05). In contrast, the sugarcane tops silage produced from variety C22, with poor nitrogen fixation ability, treated with 150 kg/ha nitrogen had the highest lactic acid bacteria (LAB) counts, dry matter (DM), organic matter (OM) and lactic acid (LA) contents (P < 0.05), as well as the lowest acid detergent fiber (ADF) and neutral detergent fiber (NDF) contents (P < 0.05). However, these results were not present in the sugarcane tops silage produced from variety T11, with no nitrogen fixation ability, whether it was treated with nitrogen or not; although the silage was treated with 300 kg/ha nitrogen, the ammonia-N (AN) content was the lowest (P < 0.05). After 14 days of aerobic exposure, Bacillus abundance increased in the sugarcane tops silage produced from variety C22 treated with 150 kg/ha nitrogen and from varieties C22 and B9 treated with 300 kg/ha nitrogen, while Monascus abundance increased in the sugarcane tops silage produced from varieties B9 and C22 treated with 300 kg/ha nitrogen and from variety B9 treated with 150 kg/ha nitrogen. However, correlation analysis showed that Monascus was positively correlated with Bacillus irrespective of nitrogen level and sugarcane variety. Our results indicated that sugarcane variety C22, with poor nitrogen fixation ability, treated with 150 kg/ha nitrogen produced the highest sugarcane tops silage quality and inhibited the proliferation of harmful microorganisms during spoilage.

The Role of OH Cards in Psychological Interventions for Children With Fractures.

Context: Fractures are traumatic events, with psychological effects that can have a negative impact on children hospitalized with fractures. They can seriously affect children's physical rehabilitation and quality of life and even produce psychological disorders The OH card is a metaphorical card that allows access to an individual's inner world and can have a positive effect in psychotherapy. Objective: The study intended to investigate the use of OH Cards during psychological interventions with children with fractures and to provide a methodological reference for the use of OH Cards in therapy. Design: The research team performed a randomized controlled study. Setting: The study took place in the Department of Trauma Surgery at Children's Hospital of Hebei Province in Shijiazhuang, China. Participants: Participants were 74 children with fractures who had been admitted to the hospital between September 2020 and November 2021. Intervention: The research team randomly divided participants into two groups using a random number table: (1) 37 in the intervention group, who received a conventional nursing intervention and also an OH-card intervention, and (2) 37 in the control group, who received conventional nursing interventions only. Outcome Measures: At baseline and postintervention, the research team: (1) measured the participants' posttraumatic growth scores, using the children's version of the Post-Traumatic Growth Inventory (PTGI); (2) assessed their coping styles, using the Medical Coping Modes Questionnaire (MCMQ); (3) determined the existence of any stress disorders, using the Child Stress Disorder Checklist (CSDC); (4) evaluated their mental statuses using the Depression Self-Rating Scale (DSRSC) and the Screen for Child Anxiety-related Emotional Disorders (SCARED); and (5) measured participants' Fracture Knowledge Questionnaire scores. Results: At baseline, no significant differences existed between the groups for any outcome measure at baseline. Postintervention, the intervention group's scores: (1) on the PTGI, were significantly higher for mental change, appreciate life, individual force, new possibilities and personal relation than those of the control group; (2) on the MCMQ, were significantly higher for facing and significantly lower for avoidance and yield than those of the control group; (3) on the CSDC, were significantly lower for trauma incidents and acute response than the control group did; (4) on the DSRSC were significantly lower and on SCARED were significantly higher than those of the control group; and (5) on the Fracture Knowledge Questionnaire were significantly higher than those of the control group. Conclusions: OH Cards can increase the posttraumatic growth scores of children with fractures, improve their coping styles, reduce stress disorders, decrease depression and improve their psychological state, increase their knowledge about fractures, and promote their recovery.

MiR-1275 Targeting SPARC Promotes Gambogic Acid-Induced Inhibition of Gastric Cancer.

Gambogic acid (GA) has been observed to effectively impede the progression of numerous types of cancers. In this study, we investigated the effects of miR-1275 and Secreted Protein Acidic and Cysteine Rich (SPARC) on GA in gastric cancer (GC). miR-1275 and SPARC expression were determined in GC cell lines and tissues using reverse transcription-quantitative polymerase chain reaction (RT-qPCR). The correlation between miR-1275 and SPARC expression was ascertained using Pearson's correlation coefficient. Cell proliferation was assessed using the cell counting kit-8 (CCK-8) assay. The Transwell assay was conducted to examine cell migration. A dual-luciferase reporter assay was used to verify the regulatory relationship between miR-1275 and SPARC. The levels of SPARC, Bcl-2, and Bax proteins were estimated using western blotting. To verify the effects of GA on the growth of GC cells in vivo, a tumorigenesis experiment was performed in nude mice. GA suppressed GC cell viability and migration, facilitated apoptosis, and inhibited tumor growth in vivo and in vitro. Low levels of miR-1275 been observed in GC cell lines and tissues. GA-treated GC cells manifested high miR-1275 levels. In functional experiments, miR-1275 enhanced the influence of GA on cell apoptosis, migration, and proliferation. Furthermore, GA treatment suppressed SPARC upregulation in GC cell lines and tissues. Pearson's correlation coefficient revealed that miR-1275 expression negatively correlated with SPARC expression. Mechanistically, miR-1275 promoted growth inhibition in GA-treated GC cells by targeting SPARC. Our study indicates that miR-1275 enhances the suppressive effect of GA on GC progression by inhibiting SPARC expression. Through this study, we contribute to the knowledge of a new mechanism by which GA suppresses GC progression.

Federal learning edge network based sentiment analysis combating global COVID-19.

As one of the important research topics in the field of natural language processing, sentiment analysis aims to analyze web data related to COVID-19, e.g., supporting China government agencies combating COVID-19. There are popular sentiment analysis models based on deep learning techniques, but their performance is limited by the size and distribution of the dataset. In this study, we propose a model based on a federal learning framework with Bert and multi-scale convolutional neural network (Fed_BERT_MSCNN), which contains a Bidirectional Encoder Representations from Transformer modules and a multi-scale convolution layer. The federal learning framework contains a central server and local deep learning machines that train local datasets. Parameter communications were processed through edge networks. The weighted average of each participant's model parameters was communicated in the edge network for final utilization. The proposed federal network not only solves the problem of insufficient data, but also ensures the data privacy of the social platform during the training process and improve the communication efficiency. In the experiment, we used datasets of six social platforms, and used accuracy and F1-score as evaluation criteria to conduct comparative studies. The performance of the proposed Fed_BERT_MSCNN model was generally superior than the existing models in the literature.

Deep Transfer Learning for Cross-Species Plant Disease Diagnosis Adapting Mixed Subdomains.

A deep transfer learning framework adapting mixed subdomains is proposed for cross-species plant disease diagnosis. Most existing deep transfer learning studies focus on knowledge transfer between highly correlated domains. These methods may fail to deal with domains that are poorly correlated. In this study, mixed domain images were generated from source and target image groups for improving the correlation between the mixed domain (training dataset) and the target domain (testing dataset). A subdomain alignment mechanism is employed to transfer knowledge from the mixed domain to the target domain. The proposed framework captures the fine-grained information more effectively. Extensive experiments were conducted and prove that the proposed method produces a more effective result compared with existing deep transfer learning technologies for poorly related subdomains.

Toward real-time and efficient cardiovascular monitoring for COVID-19 patients by 5G-enabled wearable medical devices: a deep learning approach.

Patients with deaths from COVID-19 often have co-morbid cardiovascular disease. Real-time cardiovascular disease monitoring based on wearable medical devices may effectively reduce COVID-19 mortality rates. However, due to technical limitations, there are three main issues. First, the traditional wireless communication technology for wearable medical devices is difficult to satisfy the real-time requirements fully. Second, current monitoring platforms lack efficient streaming data processing mechanisms to cope with the large amount of cardiovascular data generated in real time. Third, the diagnosis of the monitoring platform is usually manual, which is challenging to ensure that enough doctors online to provide a timely, efficient, and accurate diagnosis. To address these issues, this paper proposes a 5G-enabled real-time cardiovascular monitoring system for COVID-19 patients using deep learning. Firstly, we employ 5G to send and receive data from wearable medical devices. Secondly, Flink streaming data processing framework is applied to access electrocardiogram data. Finally, we use convolutional neural networks and long short-term memory networks model to obtain automatically predict the COVID-19 patient's cardiovascular health. Theoretical analysis and experimental results show that our proposal can well solve the above issues and improve the prediction accuracy of cardiovascular disease to 99.29%.

Medical Image Super-Resolution Based on Semantic Perception Transfer Learning.

Medical images are an important basis for doctors to diagnose diseases, but some medical images have low resolution due to hardware technology and cost constraints. Super-resolution technology can reconstruct low-resolution medical images into high-resolution images and enhance the quality of low-resolution images, thus assisting doctors in diagnosing diseases. However, traditional super-resolution methods mainly learn the mapping relationships among modal pixels from low resolution to high resolution, lacking the learning of high-level semantic features, resulting in a lack of understanding and utilization of semantic information, such as reconstructed objects, object attributes, and spatial relationships between two objects. In this paper, we propose a medical image super-resolution method based on semantic perception transfer learning. First, we propose a novel semantic perception super-resolution method that empowers super-resolution models to perceive high-level semantics by transferring features of the image description generation network in natural language processing. Second, we construct a semantic feature extraction network and an image description generation network and comprehensively utilized image and text modal data to learn transferable, high-level semantic features. Third, we train an end-to-end, semantic perception super-resolution model by fusing dynamic perceptual convolution, a semantic extraction network, and distillation polarization self-attention. Experiments show that semantic perception transfer learning can effectively improve the quality of super-resolution reconstruction.

Multi-omics analyses reveal MdMYB10 hypermethylation being responsible for a bud sport of apple fruit color.

Apple bud sports offer a rich resource for clonal selection of numerous elite cultivars. The accumulation of somatic mutations as plants develop may potentially impact the emergence of bud sports. Previous studies focused on somatic mutation in the essential genes associated with bud sports. However, the rate and function of genome-wide somatic mutations that accumulate when a bud sport arises remain unclear. In this study, we identified a branch from a 10-year-old tree of the apple cultivar 'Oregon Spur II' as a bud sport. The mutant branch showed reduced red coloration on fruit skin. Using this plant material, we assembled a high-quality haplotype reference genome consisting of 649.61 Mb sequences with a contig N50 value of 2.04 Mb. We then estimated the somatic mutation rate of the apple tree to be 4.56 × 10 -8 per base per year, and further identified 253 somatic single-nucleotide polymorphisms (SNPs), including five non-synonymous SNPs, between the original type and mutant samples. Transcriptome analyses showed that 69 differentially expressed genes between the original type and mutant fruit skin were highly correlated with anthocyanin content. DNA methylation in the promoter of five anthocyanin-associated genes was increased in the mutant compared with the original type as determined using DNA methylation profiling. Among the genetic and epigenetic factors that directly and indirectly influence anthocyanin content in the mutant apple fruit skin, the hypermethylated promoter of MdMYB10 is important. This study indicated that numerous somatic mutations accumulated at the emergence of a bud sport from a genome-wide perspective, some of which contribute to the low coloration of the bud sport.

CSPP-IQA: a multi-scale spatial pyramid pooling-based approach for blind image quality assessment.

The traditional image quality assessment (IQA) methods are usually based on convolutional neural networks (CNNs). For these IQA methods using CNNs, limited by the feature size of the fully connected layer, the input image needs be tailored to a pre-defined size, which usually results in destroying the original structure and content of the input image and thus reduces the accuracy of the quality assessment. In this paper, a blind image quality assessment method (named CSPP-IQA), which is based on multi-scale spatial pyramid pooling, is proposed. CSPP-IQA allows inputting the original image when assessing the image quality without any image adjustment. Moreover, by facilitating the convolutional block attention module and image understanding module, CSPP-IQA achieved better accuracy, generalization and efficiency than traditional IQA methods. The result of experiments running on real-scene IQA datasets in this study verified the effectiveness and efficiency of CSPP-IQA.

Machine learning techniques for CT imaging diagnosis of novel coronavirus pneumonia: a review.

Since 2020, novel coronavirus pneumonia has been spreading rapidly around the world, bringing tremendous pressure on medical diagnosis and treatment for hospitals. Medical imaging methods, such as computed tomography (CT), play a crucial role in diagnosing and treating COVID-19. A large number of CT images (with large volume) are produced during the CT-based medical diagnosis. In such a situation, the diagnostic judgement by human eyes on the thousands of CT images is inefficient and time-consuming. Recently, in order to improve diagnostic efficiency, the machine learning technology is being widely used in computer-aided diagnosis and treatment systems (i.e., CT Imaging) to help doctors perform accurate analysis and provide them with effective diagnostic decision support. In this paper, we comprehensively review these frequently used machine learning methods applied in the CT Imaging Diagnosis for the COVID-19, discuss the machine learning-based applications from the various kinds of aspects including the image acquisition and pre-processing, image segmentation, quantitative analysis and diagnosis, and disease follow-up and prognosis. Moreover, we also discuss the limitations of the up-to-date machine learning technology in the context of CT imaging computer-aided diagnosis.

An Improved Neural Network Based on SENet for Sleep Stage Classification.

Sleep staging is an important step in analyzing sleep quality. Traditional manual analysis by psychologists is time-consuming. In this paper, we propose an automatic sleep staging model with an improved attention module and hidden Markov model (HMM). The model is driven by single-channel electroencephalogram (EEG) data. It automatically extracts features through two convolution kernels with different scales. Subsequently, an improved attention module based on Squeeze-and-Excitation Networks (SENet) will perform feature fusion. The neural network will give a preliminary sleep stage based on the learned features. Finally, an HMM will apply sleep transition rules to refine the classification. The proposed method is tested on the sleep-EDFx dataset and achieves excellent performance. The accuracy on the Fpz-Cz channel is 84.6%, and the kappa coefficient is 0.79. For the Pz-Oz channel, the accuracy is 82.3% and kappa is 0.76. The experimental results show that the attention mechanism plays a positive role in feature fusion. And our improved attention module improves the classification performance. In addition, applying sleep transition rules through HMM helps to improve performance, especially N1, which is difficult to identify.

Microbiome and fermentation parameters in the rumen of dairy buffalo in response to ingestion associated with a diet supplemented with cysteamine and hemp seed oil.

In this study, high-throughput gene amplicon sequencing was used to investigate the effects of 6 treatments [2 levels of hemp seed oil (HSO) × 3 levels of cysteamine (CS)] on bacterial and fungal communities in the rumen of 30 crossbred dairy buffalo. Our results indicate that the total numbers of bacterial and fungal taxa were unaffected regardless of diet (p > 0.05), while the total number of archaea was affected (p < 0.05) by the interaction of HSO and CS. Compared with control treatment, microbial composition of archaea was strongly influenced by CS (p < 0.05), while the addition of HSO, CS or both had a weak effect on fungus and bacteria. In addition, there was a significant increase in the lactic acid content with the addition of HSO, and the addition of CS to the feed caused a significant decrease in the ratio of acetic acid to propionic acid, compared with control treatment (p < 0.05). Correlation analysis showed that Acetobacter was significantly positively correlated with the genera Pichia, Klebsiella and Acinetobacter. pH was found to have a significant effect on the methanogens, and total volatile fatty acids (VFA) had a strong correlation with Butyrivibrio. The strong influence of CS on some methanogens shows that it may have potential in the development of methane reduction interventions.

Smartphone Interventions Effect in Pediatric Subjects on the Day of Surgery: A Meta-Analysis.

Background: A meta-analysis was performed to evaluate the effect of smartphone interventions on the anxiety of the pediatric subjects at induction on the day of surgery compared to oral midazolam or standard care as control. Methods: A systematic literature search up to June 2021 was performed and nine studies selected 785 pediatric subjects on the day of surgery at the start of the study; 390 of them were using smartphone interventions, 192 were control, and 203 were using oral midazolam. They were reporting relationships between the effects of smartphone interventions on the anxiety of the pediatric subjects at induction on the day of surgery compared to oral midazolam or control. The mean difference (MD) with its 95% CIs was calculated to assess the effect of smartphone interventions on the anxiety of the pediatric subjects at induction on the day of surgery compared to oral midazolam or control using the continuous method with a fixed or a random-effects model. Results: Smartphone interventions in pediatric subjects were significantly related to lower anxiety at induction on the day of surgery (MD, -19.74; 95% CI, -29.87 to -9.61, p < 0.001) compared to control and significantly related to lower anxiety at induction on the day of surgery (MD, -7.81; 95% CI, -14.49 to -1.14, p = 0.02) compared to oral midazolam. Conclusion: Smartphone interventions in pediatric subjects on the day of surgery may have lower anxiety at induction compared to control and oral midazolam. Further studies are needed to confirm these findings.

CNN-RNN Based Intelligent Recommendation for Online Medical Pre-Diagnosis Support.

The rapidly developed Health 2.0 technology has provided people with more opportunities to conduct online medical consultation than ever before. Understanding contexts within different online medical communications and activities becomes a significant issue to facilitate patients' medical decision making process. As a subcategory of machine learning, neural networks have drawn increasing attentions in natural language processing applications. In this article, we focus on modeling and analyzing the patient-physician-generated data based on an integrated CNN-RNN framework, in order to deal with the situation that patients' online inquiries are usually not very long. A so-called DP-CRNN algorithm is developed with a newly designed neural network structure, to extract and highlight the combination of semantic and sequential features in terms of patient's inquiries. An intelligent recommendation method is then proposed to provide patients with automatic clinic guidance and pre-diagnosis suggestions, in which a clustering mechanism is utilized to refine the learning process with more precise diagnosis scope and more representative features. Experiments based on the collected real world data demonstrate the effectiveness of our proposed model and method for intelligent pre-diagnosis service in online medical environments.