This PDF file contains the front matter associated with SPIE Proceedings Volume 13208, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

OBJECTIVE: This study aims to investigate the intervention effectiveness and feasibility of the computer programs of rapid eye movement (REM) and relaxation on acute stress in adults. METHODS: A randomized controlled trial was conducted in which 90 healthy adult male subjects were randomly assigned to use the control, REM or relaxation computer program after acute stress induction. The biological indicator of Salivary Cortisol, Positive and Negative Affect Schedule and Stress Self-Assessment Scales were collected from the subjects before and after stress elicitation and intervention to assess their stress state. RESULTS: Repeated-measures ANOVA showed that salivary cortisol [t (75) = 12.61, P < 0.001], stress self-assessment [t (86) = 4.45, P < 0.001], and negative affect [t (85) = 19.76, P < 0.001] were significantly elevated in all groups after stress induction, which was successful. Reductions in negative affect were more significant in both the REM group (P = 0.030) and relaxation group (P = 0.036) than in the control group at post-intervention, with no significant differences between the two groups. CONCLUSION: The present study demonstrated that both computer programs of REM and relaxation effectively attenuated negative affect and reduced the acute stress in adults. The experiment results provide valuable insights into the construction and application of computer intervention system for acute stress management.

In order to study the molecular properties and pharmacological activity of artemisinin, computer high performance computing (HPC) and density functional theory (DFT) were used to simulate and calculate the structure of this drug. The microscopic structure characteristics of artemisinin were obtained, the electrostatic potential (ESP) distribution and its infrared spectrum (IR), ultraviolet-visible absorption spectrum (UV) and fluorescence spectra in artemisinin were simulated and predicted. It was concluded that artemisinin could be used for characteristic absorption detection of infrared vibration, and artemisinin could be quantitatively detected by ultraviolet absorption spectrum and molecular fluorescence spectrum. More importantly, it was confirmed that artemisinin molecules required low energy to be activated, and had active molecules of nucleophilic and electrophilic reactions. It was expected to provide reference and guidance for the structural properties of artemisinin and the modification of artemisinin-like drugs molecules.

The exploration of defense peptides holds significant research potential for fungal inhibition. This study optimized the antifungal gene (PtAFP) from Pyrenophora tritici-repentis using synonymous codon bias. The resulting antifungal protein PtAFP exhibited a characteristic antiparallel β-folded barrel structure, featuring three disulfide bonds, with the initial 23 amino acids forming a signal secretion sequence. Successful cloning of the optimized PtAFP gene into Escherichia coli yielded an antifungal protein of approximately 10 kDa. Fluorescence analysis assessing the thermal stability of PtAFP revealed a remarkable retention of 87% fluorescence intensity after 6 h at 50°C, underscoring its robust protein stability. Inhibition rate analysis against Penicillium funiculosum CICC40279 indicated a 15% inhibition rate after 48 h of cultivation, persisting at 18% even after 72 h. This study highlights the effective application of the optimized gene sequence in E. coli, laying the groundwork for future advancements in the development of more potent antifungal proteins.

Single-cell RNA sequencing allows to discovery of new cell subtypes based on transcriptomic information. Clustering analysis is an effective approach for exploring single-cell heterogeneity. Nevertheless, current single-cell clustering methodologies are challenged by the high-dimensional, sparse, and dropouts on the one hand, and on the other hand, they neglect the thorough exploration of potential relationships between cells and noise removal. To tackle these challenges, this study introduces a novel self-optimized single-cell clustering algorithm named scZVEA, which combines ZeroInflated Negative Binomial (ZINB) model and variational graph attention autoencoder. The scZVEA framework comprises three key modules. Firstly, a Deep Count Autoencoder (DCA) is designed to model data distribution and eliminate noise. Then, a variational graph attention autoencoder to extract latent features from the data. Lastly, the self-optimized clustering module enables the two previously independent clustering and feature modules to mutually benefit each other by iteratively updating cluster centers to further enhance clustering performance. Experimental findings based on six authentic scRNAseq datasets illustrate that the proposed clustering algorithm significantly enhances clustering accuracy when compared to alternative methods.

BRCA1 gene testing is a trump card for liquid biopsy and early diagnosis in breast cancer, which has been limited, however, by the fact that the current mainstream methods such as gene sequencing are expensive, time-consuming and complex. In this study, a semiconductor carbon nanotube (CNT)-based field-effect transistor (FET) DNA sensor was constructed to detect BRCA1. The CNT-FET biosensors employed a floating gate (FG) structure and Y2O3 / HfO2 high κ bilayer gate dielectric. A key advance here is that the detection performance of three nucleic acid probes, DNA, phosphorodiamidate morpholino oligos (PMO) and peptide nucleic acid (PNA), was systematically investigated for the first time to define the optimal BRCA1 sensor configuration. The PNA-functionalized FG CNT-FET sensor stood out with an excellent sensitivity (LOD: 1.38 aM) and good specificity. What is more, the developed sensor also exhibited good anti-interference capability in serum BRCA1 testing, owing to the passivation of the bilayer dielectric. Importantly, the breast cancer patients and healthy individuals were successfully distinguished by our method, which is fully consistent with the gene sequencing results. This work represents a desirable analytical platform for the simple, sensitive, and accurate liquid biopsy of breast cancer. And we envision that the developed method will provide ideas for the construction of future DNA sensors and the selection of their recognition elements.

While Yinma Jiedu Granules are frequently employed for Acute bronchitis treatment, research on their network pharmacology and molecular docking remains scarce. Utilizing network pharmacology and molecular docking techniques, this study forecasts potential targets and signaling routes for Yinma Jiedu Granules in managing Acute bronchitis. The active components and related targets of Yinma Jiedu Granules were obtained by TCMSP, GeneCards and UniProt. GeneCards, DrugBank and OMIM were used to obtain targets related to Acute bronchitis. By merging the STRING database, the network of protein interactions was created, and the PPI network linking Yinma Jiedu Granule and Acute bronchitis was established using the Cytoscape v3.8.2 software. GO bioinformatics analysis and KEGG pathway enrichment analysis on intersecting genes were conducted using the David database, and the bubble map and histogram were made using the micro-bioinformatics platform and RStudio software. The key targets were verified by molecular docking with the main active components by AutoDockTools 1.5.7, Pymol, BatchDocking and LigPlus software. Findings indicated that employing Oral Bioavailability and Drug Likeness for screening led to the acquisition of 117 active elements, 154 genes intersecting drug-disease, and 111 primary targets. The KEGG enrichment study forecasted that Yinma Jiedu Granules' therapy for Acute bronchitis primarily targeted advanced glycation end products, including receptor and TNF signaling pathways, IL-17 signaling pathways, among others. Molecular docking outcomes verified that Yinma Jiedu Granules' primary active elements, quercetin and luteolin, maintained a consistent binding affinity with key targets STAT3 and HSP90AA1.

This research delves into the intricate mechanisms of fish sex determination, with a particular emphasis on lampreys, exploring phenomena such as sex reversal. We investigate the influence of lamprey sex conversion on their adaptability to environmental changes, resilience to adverse conditions, and ecological interactions. By applying the Verhulst-Volterra model, this study evaluates the effects of gender ratio variability on lamprey reproductive success and the overall stability of ecosystems across different resource scenarios. Our results indicate that variability in gender ratios significantly improves lamprey survival rates and provides valuable perspectives on the implications of gender diversity for the stability of the global biosphere.

In order to explore the influence of methylation on molecular characteristics, the molecular properties of cytosine and cytosine deoxynucleotides before and after methylation were compared through a computer-simulated prediction under the density functional theory method, including the natural charge distribution, infrared vibration spectrum, ultraviolet absorption spectrum, and the change of the energy gap of the front molecular orbital. The results showed that with the methylation of cytosine and cytosine deoxynucleotides, the amino nucleophilic ability and molecular activity of cytosine and cytosine deoxynucleotides were improved. In addition, it had obvious ultraviolet characteristic absorption spectrum after methylation, which could be used for qualitative analysis of cytosine and cytosine deoxynucleotide methylation. And the nucleophilic ability and molecular activity of cytosine and cytosine deoxynucleotides were improved after methylation, which could have an important effect on DNA inheritance.

In gene regulatory networks, due to the influence of paradigm, the number of parent nodes of a gene node is always smaller than the sum of the number of other gene nodes in the network, so there must exist some nodes that are not the parents of that node. In the Hidden Markov's Non-Chiral Dynamic Bayesian Network (HMM-DBN) model, the sampling strategy of the parent node set is to sample all the nodes randomly, and does not take into account the correlation between the nodes in the network, and the no-joints are sampled many times, which leads to poor accuracy and slows down the rate of convergence in the process of constructing the network. To address this problem, this paper proposes a Hidden Markov Non-homogeneous Dynamic Bayesian Network (MICHMM-DBN) model based on node screening. The model identifies other nodes that have high correlation with a particular node through the maximum information coefficient and treats these nodes as candidate parent nodes for that particular node. Following this, Markov chain Monte Carlo (MCMC) sampling is performed on these candidate parent nodes to determine the final set of parent nodes. This method effectively reduces the search space and eliminates redundant connections, thus improving the accuracy of network construction. In a series of experiments conducted on synthetic datasets as well as real biological network data, the new model demonstrates higher network construction accuracy compared to the original model.

With the progress of molecular diagnostic technologies, manual nucleic acid extraction methods no longer suffice to meet current demands. In order to enhance the positioning precision and liquid handling accuracy of a 25-well high-throughput automated pipetting workstation, we investigated an S-shaped acceleration-deceleration algorithm based on the Sigmoid function for controlling stepper motors along the X, Y, and Z axes. Subsequently, we constructed a corresponding mechanical device. Through this apparatus, we conducted experiments to assess positioning precision, liquid handling accuracy, and biological performance. The experimental results revealed that the absolute error in positioning precision was less than 0.1 mm, and the coefficient of variation (CV) for liquid handling accuracy was below 1%. Furthermore, post-purification nucleic acid A260/A280 ratios were measured at approximately 1.911, falling within the range of 1.8 to 2.0. Our devised system further improved both positioning and liquid handling precision, thereby advancing high-throughput automated nucleic acid extraction technology.

HIV spreads in the body through two patterns: virus-to-cell (VTC) infection and cell-to-cell (CTC) transmission, where CTC transmission refers to the actively and latently infected cells transmitting the virus to target cells through direct cellular contacts via the virological synapse. Latently infected cells hide within dormant memory CD4+ T cells to escape immunity, and they will become actively infected cells after a time lag that can produce new infectious viruses. This paper establishes an HIV latent model with both infection patterns. For the model, the basic regeneration number and the equilibria under different infection states are given to show the disease transmission nature and prevalence, and the local asymptotic stability of the disease-free equilibrium is proved by applying the Routh-Hurwitz criterion. Innovatively, for high-dimensional virus dynamics system, we confirm the global asymptotic stability of the endemic equilibrium by the Lyapunov method. Finally, the theoretical conclusions are verified by numerical simulations.

This study delves into the complex relationship between exercise, muscle physiology, and vascular dynamics, highlighting the role of Near-Infrared Spectroscopy (NIRS) in elucidating these intricacies. The study design, involving the assessment of muscle oxygen levels and hemodynamic parameters during exercise, provides valuable insights into the physiological responses underlying physical activity. The consideration of multiple influencing factors underscores the multifaceted nature of blood flow regulation in skeletal muscle. Experimental findings, such as the initial decline followed by an increase in total hemoglobin concentration ([tHb]) during exercise, reveal the dynamic interplay between mechanical factors and vascular regulation. This highlights the importance of considering the temporal evolution of physiological responses. Future research directions include the utilization of advanced imaging techniques and pharmacological interventions to unravel the complexities of muscle-blood flow interactions.

Two-photon microscopy plays a significant role in biological applications, as it enables non-invasive, high-resolution imaging of thick tissue samples. This paper begins by elucidating fundamental concepts of two-photon microscopy to facilitate a deeper understanding of its luminescent mechanism. Subsequently, we provide a comprehensive overview of the imaging system, encompassing the light source, scanning unit, fluorescence collection, and detection components. As two-photon microscopy continues to evolve, We undertake assessments of cutting-edge optical instrumentation, emphasizing novel approaches for imaging across wider fields of view, at higher speeds, and with increased volumetric imaging capabilities. Furthermore, we summarize and compare the technical specifications of advanced two-photon imaging. Following this, the manuscript showcases the transformative applications of these innovative technologies in biological imaging, providing further evidence of the indispensable role of two-photon microscopy in biomedical imaging research.

The three-dimensional cell spheroids hold great potential for accurately simulating the microenvironment of cells or tissues within living organisms, making them indispensable in various fields including drug screening and tissue engineering. However, traditional methods employed to generate three-dimensional cell spheroids suffer from limitations such as slow production speed, high operational complexity, and inadequate uniformity. In light of these challenges, we have developed an intelligent system designed to address these issues and facilitate the efficient fabrication of three-dimensional cell spheroids. This innovative system facilitates precise control over the movement of the printing head through the utilization of a development board and software. Moreover, the printing chip, manufactured using CMOS-MEMS technology, enables the placement of a 3×3 grid of cell spheroids within each well of a 96-well plate lid. Remarkably, the production of 864 cell spheroids across the entire lid of the 96-well plate requires only 54 minutes, thus effectively achieving high-throughput production of three-dimensional cell spheroids. Furthermore, the coefficient of variation (CV) pertaining to the diameter of the printed cell spheroids can be remarkably reduced to 4.97%. To enhance the uniformity and size precision of the cell spheroids, we propose two optimization techniques: extending the settling time for the cells and incorporating the use of an ultrasonic cleaner. Application of these methods individually resulted in reductions of 4.7% and 1.77%, respectively, in the diameter of the cell spheroids. This integrated system not only exhibits promising implications for advancing the preparation of proteins and organ-like structures but also opens new avenues for further exploration in these areas.

Arteriosclerosis obliterans (ASO) of the lower limbs is a common peripheral arterial disease encountered in clinical practice, which has caused a large numbers of amputation in middle-aged and elderly individuals. The treatments of ASO are frequently associated with restenosis and adverse events after surgical recanalization. In this study, bioinformatics were used to analyze 75 ASO patients from two GEO databases. Multi-algorithm joint filtering of key mutation sites, biological function annotation and intervened disease model verification were performed to screen out the core genes that may be related to pathological vascular remodeling. Genes with higher expression in occlusive lower extremity artery than the normal, which may undergo pathogenic mutations, were conjectured to result a more profound adverse influence on vascular structure. Our study reveales that matrix metalloproteinase 9 (MMP9) has recurring harmful up-regulated in occlusive artery and inhibition of MMP9-related proteins decreases the degree of foam in smooth muscle cells in vitro and in vivo. Based on this, we conclude that MMP9 may play a key role in the onset and progression of ASO in lower extremity artery by controlling smooth muscle cell phenotypic shift.

Examining the state of research worldwide and current areas of interest in the field of cell-laden scaffold was the goal of this study. Using the pertinent keywords "cell-laden scaffolds," we searched the Web of Science Core Collection database and obtained the published literature from 2012 to 2024. Information was extracted about journal title, author details, institutional affiliation, nation, and keywords. The obtained data was then subjected to bibliometric analysis using Bibliometrix. China and the United States emerged as the most prolific countries in the field, contributing significantly to the research output. The journals that were most frequently utilized for publishing in this domain were Biofabrication, Acta Biomaterial, Advanced Healthcare Materials. Additionally, the current research landscape primarily focuses on tissue engineering, 3d/4d bioprinting, and gelama, indicating their significance as emerging research hotspots. Cell-laden scaffolds hold significant potential in biomedical engineering as the seeded cells were embedded into while the scaffolds were fabricated and offer promising avenues for further research.

Over the last decade, photothermal agents are developing rapidly in disease diagnosis and treatment. In particular, near-infrared (NIR) photothermal agents, which have potential for clinical application, have attracted more and more attention. Herein, we report a supramolecular complex with high biocompatibility obtained by supramolecular PEGylation between host and guest. In addition, this supramolecular complex exhibits excellent NIR photothermal conversion ability and photothermal stability. This supramolecular complex is expected to be applied to in vivo photothermal therapy.

Primary liver cancer is a prevalent malignancy in the digestive system and represents the third leading cause of cancerrelated deaths globally. The tumor microenvironment, consisting of various immune cells and the extracellular matrix, interacts with cancer cells, influencing liver cancer progression. Our comprehensive investigation into the immune microenvironment utilized RNA-seq data from the TCGA database, analyzed via CIBERSORTx software. We identified key immune cell populations, including CD4 memory resting T cells, CD8 T cells, and Macrophages M0 and M2, pervasive throughout all liver cancer stages. Notably, resting NK cell levels were significantly higher in stage 1 and stage 3 patients, while Macrophages M0 were lower in stage 1 compared to stage 2. CD4 memory resting T cells were elevated in stages I and IV. Furthermore, longer survival correlated with lower Tregs and Macrophages M0, but higher CD8 T cells. Histological grading showed a decrease in resting NK cells, Monocytes, Macrophages M2, and resting Mast cells from primary to secondary patients. This study highlights the heterogeneity of immune cell infiltration patterns in liver cancer across different stages, survival durations, and histological grades, advocating for personalized immunotherapeutic strategies to accommodate the diverse disease backgrounds of patients.

Observation of actin in cells is crucial for the investigation of cell motility, protein transcription, pathogen infection, and cell cycle research. In this study, we have identified Brilliant Green (BG) as a water-soluble AIEgen (luminogen with aggregation-induced emission properties) that can be utilized for actin staining. Our experimental findings highlight the effectiveness of BG for long-term live cell imaging, as well as its remarkable stability and resistance to photobleaching.

This study aims to combine Raman spectroscopy technology with deep learning algorithms to achieve precise classification and identification of cancer cells. Firstly, a large-scale spectral database is established using datasets. Subsequently, deep learning techniques are employed to design and train various deep learning network models for cancer cell classification. The models can extract high-order features from complex spectral data, enabling accurate identification of different types of cancer cells. Through validation and optimization of the models, their accuracy and reliability in cancer diagnosis will be verified, providing clinicians with an effective auxiliary diagnostic tool that is expected to play a significant role in improving the accuracy of cancer diagnosis and treatment efficacy.

This study aims to explore the molecular mechanisms and protective effects of Monoisobutyrate on lipopolysaccharide (LPS)-induced inflammatory damage in microglial cells by integrating GEO data analysis, molecular docking techniques, and in vitro experiments. Transcription data of BV2 microglia post-LPS induction were analyzed using the Limma package, with protein interactions of differentially expressed genes obtained from the STRING database. The resulting PPI network was clustered and analyzed using the Mcode plug-in, and GO and KEGG pathway enrichment analyses were conducted using the clusterProfiler package. Molecular docking technology facilitated the screening and validation of target drugs via in vitro cellular experiments on LPS-induced BV2 microglial cells. The study identified 249 up-regulated genes and 116 down-regulated genes through GEO database screening. Key pro-inflammatory proteins such as TNF, IL1b, and IL6 were highlighted in the PPI results, indicating their significance in the biological process of BV2 microglia regulated by the NF-kB pathway.The small molecule morroniside, identified through molecular docking, demonstrated the ability to reduce nuclear translocation of NF-κB transcription factor, block TNF-α generation, decrease intracellular ROS levels, and mitigate apoptotic cell death in BV2 microglia. Ultimately, morroniside was found to protect against LPS-induced neuroinflammation by targeting the NF-kB pathway and interacting with downstream pathways, showcasing its potential as a therapeutic agent.

Thermophilic proteins are proteins that can maintain good activity at high temperatures, they have important application value in the fields of enzyme engineering and biopharmaceuticals. Therefore, accurate identification of thermophilic proteins provides important information for their engineering applications. In this study, we proposed an improved predictor of thermophilic protein using ProtBert protein language model, called PB-TMP. Firstly, we used the ProtBert protein language model to extract the hidden features in protein sequences, and 1024-dimensional features are obtained. Secondly, Recursive Feature Elimination (RFE) are employed to remove the noise or redundant features, 31-dimensional features are selected as the final features set. Finally, Multi-Layer Perceptron (MLP) is trained on the final features set, then an identification model is obtained. The model yielded an accuracy of 98.53% on 10-fold cross-validation, and performed best among the existing methods. The model will accelerate the discovery of thermophilic protein and important application in protein engineering and biotechnology.

Surface electromyography(sEMG)-based hand pose estimation is a promising technique for muscle-computer interface (MCI) control, allowing users to predict finger movements using electromyography signals. However, the high sampling rate required for accurate prediction may increase device costs and hinder widespread adoption. This study aims to investigate the relationship between sEMG-based hand pose estimation and sampling rate, and provide recommendations for more economical sampling rates in future hardware devices. The study focuses on two key factors: features and prediction models, which significantly impact hand pose estimation results. Four sEMG features, including both classic and novel ones, are compared, along with three prediction models, at different sampling rates. Experimental results demonstrate that a sampling rate of 400Hz can achieve comparable recognition performance to a higher rate of 2000Hz. This conclusion holds true for both intact and amputated subjects. By exploring the relationship between sampling rate and hand pose estimation, this work provides valuable insights for optimizing hardware device production. Lowering the sampling rate to 400Hz can help reduce costs without compromising the accuracy of finger movement prediction.

Myocardial hypoperfusion triggers ischemic cascade reactions and ultimately leads to myocardial infarction. Unfortunately, medical imaging alone is not the optimal method for quantitative evaluation, and the accuracy of clinical diagnosis and subsequent treatment is limited by restricted image resolution and vascular scale identification. Numerical simulation of myocardial perfusion is expected to fill the gaps in imaging and quantitative techniques. This study introduces MultipleNetwork Poroelastic Theory (MPET) into the myocardial numerical simulation. A 3D subject-specific model with boundary conditions obtained from multimodal magnetic resonance imaging is constructed, and fluid flow and fluidstructure interaction are solved within the established MPET framework. This model incorporates three fluid components, including arterial blood, arteriole/capillary blood and venous blood. Myocardial blood flow (MBF) is the core output of the model. In this study, a preliminary analysis of the values and distribution of MBF was performed, demonstrating that the numerical results are physiological and could capture the distributional characteristics of myocardial perfusion. The simulation provides some potential insights into understanding the perfusion process from a biomechanical point of view, aiding clinical diagnosis and disease risk assessment.

In order to assist doctors in making better medical decisions, the model recommends accurate and safe drug combinations for patients by analyzing diagnosis, surgery, historical medication records and drug interaction knowledge base of a large number of patients. This paper uses a double attention mechanism to capture the influence of multiple internal members of diagnosis and surgery in the stage of patient feature representation, and considers the similarity of patients' personal history medication information in drug recommendation. In order to reduce the adverse reactions of the recommended drugs, the algorithm uses a graph neural network to model the fusion of the historical drug combination and drug interaction knowledge base and calculates the correlation with patients on this basis. The method has been extensively tested on public electronic medical record data. Experimental results of the proposed model outperformed all baselines in predicting accurate and safe drug recommendations, with a 1.02 percent improvement in accuracy and a 27 percent reduction in the proportion of drug interactions compared to several advanced methods. The drug recommendation method does not take into account the dynamic change of drug use over time of disease development. The drug recommendation model combining attention mechanism and graph neural network is effective and safe, and can perceive the influence of patient characteristics on drug use, so as to provide more accurate reference for doctors to recommend clinical drug use.

Steady-State Visually Evoked Potentials (SSVEPs) based Brain-Computer Interfaces (BCIs) offer broad application prospects and significant research value due to their capacity for multiple output commands and high performance. Effectively leveraging the inherent information of SSVEPs across various dimensions, including time, frequency, and spatial domains, can significantly enhance their recognition performance. Existing research has revealed that different independent harmonic components of SSVEPs contain distinct information and represent unique physiological functions. Consequently, this study proposes a method to enhance the detection performance of SSVEP-based BCIs by utilizing the information from each independent harmonic of SSVEP. This method hypothesizes that for specific users and stimulus conditions, the independent harmonics of SSVEP responses should each possess stable and repeatable waveform and spatial propagation characteristics. This method utilizes Task-Related Component Analysis (TRCA) for the extraction and utilization of these characteristics. By integrating the comprehensive information within SSVEPs and the localized information within each independent harmonic, the proposed method has demonstrated superior performance on public datasets compared to traditional filter bank-based FBCCA and TRCA methods. This research confirms the application potential of analyzing and utilizing each independent harmonic of SSVEP, offering new strategies and perspectives for enhancing the performance of SSVEP-based BCIs.

Multimodal imaging genetics can integrate imaging data and genetic data to help understand the pathogenesis of Alzheimer's disease (AD) and provide insights for early diagnosis. Non-negative matrix factorization (NMF) has been widely use d in the field of multimodal analysis as an effective joint feature extraction and dimensionality reduction method that can help to reveal underlying structures and patterns among data. However, these methods ignore the importance of the coefficient matrix. Based on this, this paper proposes a sparse network nonnegative matrix factorization (SnetNMF) algorithm based on the sparse constraints of netNMF. By establishing a linear relationship between structural magnetic resonance imaging (sMRI) and corresponding gene expression data under sparse constraint, the SnetNMF algorithm can effectively integrate biological information, thereby better exploring the correlations among data. By optimizing the similarity between modules, SnetNMF also can accurately identify and analyze the modules and structures within the network, thus provi ding a deeper understanding for AD research. Experiments show that the SnetNMF algorithm can accurately identify risk regions (ROIs) and key genes simultaneously for AD, genes such as MMP9, GATA1, ITGA2B, SNCA, HIF1A, etc. RO Is such as l4thVen/r4thVen, rLatVen, etc. Meanwhile, comparison experiments show that the algorithm has better reconstruction performance as well as faster running speed.

To address the issues of extensive artifacts and noise in sparse-angle computed tomography (CT) reconstruction, we propose a neural radial field-based sparse-angle CT reconstruction algorithm. Specifically, we utilize both newly synthesized projection views and original projection views as inputs to increase the projection views, thereby enhancing the quality of the reconstruction results. However, the newly synthesized projection images exhibit a certain degree of blur. To address this issue, we incorporate a deblurring module into the network to optimize sampling during the process, thereby achieving deblurring effects. This approach overcomes the challenges of sparse-angle CT reconstruction. We demonstrate CT reconstruction with fewer projection data (20 projections) and validate the effectiveness of our algorithm on a publicly available clinical dataset. Experimental results show that the proposed algorithm can improve the quality of CT image reconstruction, providing a feasible and effective method.

Named Entity Recognition (NER) plays a crucial role in Chinese Natural Language Processing, particularly in medical texts containing numerous nested entities. Existing NER techniques often rely on Bidirectional Long Short-Term Memory networks combined with Conditional Random Fields (BiLSTM-CRF) to identify entities. However, their effectiveness in handling nested entities is limited. Therefore, this paper proposes a model based on multilevel convolutional networks and Biaffine networks. While single layer convolutional networks capture local features, they lack the ability to capture long-range dependencies in text. To address this issue, multilevel convolutional networks are utilized to integrate information from multiple scales. Subsequently, nested entities are identified using Biaffine networks. The proposed method demonstrates promising performance on various medical text entity recognition tasks, outperforming existing methods in nested entity recognition, as evidenced by experimental results.

Ballistocardiogram (BCG) is a technique for monitoring cardiac function, and the characteristic peaks of the BCG signal are important. However, due to individual differences, especially when the subject has cardiac dysfunction, waveform distortion occurs in the BCG signal, making traditional peak localization algorithms insufficiently adaptive. This study, based on the principle of waveform consistency in the BCG signals of the same subject, first locates the J peak, then segments the cardiac cycle, builds a template, and finally achieves peak localization by dynamic time warping matching between the template and the remaining cardiac cycles. A fiber optic sensing monitoring system is used to collect BCG signals from 10 healthy subjects. Three technicians annotate the feature peak groups according to the rules and ensure reliability using the Kappa coefficient. The experimental result shows that the overall accuracy of the algorithm was 91.7%. This method can method can help monitor cardiac function by detecting characteristic peaks in BCG signals

Channel switching is a key technology for the development of molecular communication. The realization of the channel switching function is of great significance for the construction of a mature molecular communication network. In this paper, we propose a channel switching model that is considered the communication between cells at the receiving end. The channel switching model proposed in this work is more consistent with the intercellular communication process. The proposed model is validated by numerical simulation experiments in Matlab. Compared with the previous model, the stimulus interval which can detect periodic Ca²⁺ oscillations in abnormal cell B is changed from the original (0.52, 0.9) to the present (0.39, 0.9). The stimulus interval in normal cell C is changed from (0.007, 0.77) to (0.007, 0.84). This paper also finds that attenuation in Ca²⁺ transport is more likely to affect communication between abnormal gap junctions.

Registration methods based on deep learning have been gotten popular for its performance and efficiency by automatically extracting deep features of images. Currently, the registration performance is further improved by using cascade networks. However, there is few studies pay attention to the registration of local complex areas of the brain. This paper combines the advantages of cascade networks and attention mechanism to solve the problem of registration in complex small areas of the brain by introducing a localized attention block. Specifically, we introduce an attention mechanism module (called Res-non-local-block), which consists of the attention mechanism (non-local-block), allowing the block to more flexibly adapt to complex anatomical structures and image changes. We also extract deformation fields of different scales by down-sampling the predicted deformation fields to further improve the local registration performance. Experiments on the public datasets LPBA40 and HBN demonstrate that our proposed method exhibits higher registration accuracy on local areas of the brain than several popular deep learning based on registration models.

This study aims to investigate the neural modulation effects of transcutaneous auricular vagus nerve stimulation (taVNS) on brain activity through the application of functional magnetic resonance imaging (fMRI). Thirty-two healthy individuals were allocated randomly to undergo either taVNS or a sham treatment involving earlobe stimulation. Regional homogeneity (ReHo) analysis revealed significant changes in neural activity post-taVNS, with increased activity observed in the left lingual gyrus, left calcarine fissure, and cerebellum, and decreased activity in the frontal gyrus and precuneus. No statistical changes were observed in the control group. Classification analysis based on the regions of interest (ROIs) derived from ReHo analysis demonstrated a high accuracy rate of 96.33% in distinguishing between pre- and post-taVNS conditions, highlighting the potential of machine learning techniques in recognizing brain activity patterns associated with taVNS and further emphasizing the monitorability of taVNS as a neuroregulatory technique. The contribution scores of the ROIs indicated that regions including the left calcarine fissure, right parahippocampal gyrus, and dorsolateral superior frontal gyrus play a crucial role in the neural modulation effects of taVNS. These results offer evidence for the effectiveness of taVNS as a neuroregulatory treatment method and contribute to our understanding of its neural mechanisms and potential therapeutic applications. These findings have potential for translating into clinical practice or developing personalized treatment approaches based on individual brain activity patterns.

The characteristics of electrocardiogram (ECG) signal play a crucial role in assessing human health. Accurate identification of frequency components within ECG signal can greatly assist physicians in evaluating patients' health conditions and devising appropriate treatment strategies. The process of acquiring ECG signal involves the following several stages: signal acquisition, preprocessing, feature extraction and classification. However, existing feature extraction methods have limitations, which can compromise their effectiveness for clinical applications. The ESPRIT algorithm, based on eigen decomposition of correlation matrices, offers a solution for signal feature estimation. It enables the computation of signal frequency, phase, power, and other parameters, making it particularly useful in array signal processing for precise parameter estimation even with limited data. To address these limitations, this paper proposes a novel feature analysis method based on the ESPRIT algorithm. Our goal is to overcome the shortcomings of conventional feature extraction techniques, which are often computationally intensive, require extensive calculations, and may not accurately extract electrocardiogram signal features. The proposed method was evaluated using data from the MIT-BIH database, demonstrating the ESPRIT algorithm's ability to accurately estimate frequencies. These results provide critical insights and guidance that can enhance the accuracy of ECG signal acquisition.

With the continuous development of medical technology, the current two-dimensional tomographic images can no longer well meet the needs of clinical diagnosis. Three-dimensional reconstruction technology refers to the recovery of the three-dimensional structure of the object itself from the two-dimensional image, and rendering, and ultimately the expression of the objective world of virtual reality in the computer. Three-dimensional reconstruction of medical images can reconstruct two-dimensional images of human organs into three-dimensional forms, so as to reduce the two-dimensional image of the three-dimensional image does not visualise the reasons for the clinical analysis of the error, in order to help doctors diagnose all kinds of diseases more accurately. This article reviews the application of three-dimensional reconstruction in lung clinical diseases, and at the same time looks forward to the future application of three-dimensional reconstruction technology in the direction of clinical diagnosis and treatment.

The segmentation of brain tumors plays a crucial role in improving the accuracy of diagnosis and enhancing the overall diagnostic efficacy, given the severe impact of malignant brain tumors on human health. However, traditional methods face challenges such as information loss, weak cross-scale feature fusion, and low segmentation accuracy when dealing with multi-scale targets. To address these issues, we propose an intelligent brain tumor segmentation method based on improved U-Net, to enhance the accuracy and robustness of the model in brain tumor segmentation tasks. Firstly, we improve the ResNet50 backbone network, which provides feature maps extracted from multiple levels. These feature maps are summarized and mapped to the final number of categories through the full connection layer, which can achieve the advantages of multi-layer feature fusion and improve segmentation performance. Secondly, we enhance the feature fusion module by incorporating a channel cross-attention mechanism and a dual-layer convolutional neural network, enabling a more comprehensive extraction of information from the feature maps. Trials indicate that our approach yields notable advancements in the accuracy of segmentation and superior integration of features across various scales. Such enhancements contribute substantially to the analytical processes involved in understanding brain tumor pathologies.

This study aims to propose and validate an innovative method for detecting and classifying muscle fatigue, based on surface electromyography (sEMG) signals, with a focus on accurately identifying and classifying static muscle fatigue. By recruiting seven healthy young volunteers (comprising two males and five females), this research conducted a series of fatigue tests, during which the participants' surface electromyographic signals and their subjective fatigue scores were meticulously recorded. The study extracted 18 features from the surface electromyographic signals, including time domain, frequency domain, and time-frequency domain characteristics. These features were then labeled as "normal" or "fatigued" based on the participants' Borg CR-10 subjective fatigue scale. In the data processing phase, 80% of the statistical features were selected as the training dataset for machine learning models (CNN-SVM, KNN, SVM, RF), with the remaining 20% serving as the test dataset to evaluate the performance of the four algorithms in terms of accuracy in muscle fatigue classification. The experimental results indicate that in the evaluation mode utilizing all features, the CNN-SVM model achieved the highest accuracy and F1-score, at 86.13% and 84.58% respectively. In the mode that only utilized the four principal sEMG signal features, the KNN model exhibited the best performance with an accuracy rate of 75.87%.The muscle fatigue recognition model proposed in this study is of paramount significance for monitoring muscle fatigue in static sitting and standing postures.

Conductive intracardiac communication (CIC) can be utilized for the synchronization of multi-chamber leadless pacemakers (LLPMs) to overcome complications arising from traditional pacemaker lead connections. Current CIC methods are mainly based on pulse modulation (PM) and On-Off Keying (OOK), which suffer from high power consumption and poor interference resilience, with transmitter output bit energy (Ebit) exceeding 5 pJ. Addressing these concerns, we proposed a method of conductive intracardiac communication based on Gaussian Minimum Shift Keying (GMSK). In the transmitter, a power-optimized GMSK modulation method is employed for signal transmission. In the receiver, a variable gain amplifier is utilized for CIC signal reception, and signal recovery is achieved through low-power demodulation and bit synchronization methods. A prototype transceiver was designed for measuring the bit error ratio (BER) and transmitter output E_bit of GMSK and OOK methods in in-vitro experiments using porcine hearts as channels. The transceiver was connected to the right atrium and right ventricle via stainless steel needle electrodes. At data rates of 75-500 kbps and a BER of 1e-4, the average Ebit for OOK ranged from 0.06-0.61 pJ, whereas for GMSK, it was lower, ranging from 0.02-0.35 pJ. This study demonstrates that LLPM achieving reliable CIC using GMSK at transmission powers below 35 pW is feasible in practical channel conditions. The proposed GMSK-based CIC method is more preferable to PM and OOK due to its higher reliability and lower transmission power requirements.

Ling Six-Sound is a typical sound material widely used in hearing assessment for speech perception, which consists of /m/, /u/, /a/, /i/, /sh/ and /s/. At present, the Ling Six-Sound clinical hearing assessment mainly through subjective audiometry or question-and-answer, there are inaccurate and incomplete drawbacks. With the development of brain-computerinterface, objective perception based on EEG has made great progress. In order to construct an objective evaluation method for Ling Six-Sound perception, this study used oddball paradigm to collect passive EEG induced by Ling Six-Sound, analyzed the time and spatial domain features of the EEG signals, and used four kinds of deep learning models to classify the collected EEG signals. The experimental results show that the performance of the temporal convolutional network with the fusion layer (TCNet-Fusion) is better than other models, and the accuracy for six-classification is 79.78%, which is much higher than the random level. Based on the above results, it can be concluded that the objective Ling Six-Sound perception assessment is feasible, and provide reference for introducing this objective assessment method into clinical hearing assessment in the future.

Human beings serve as a crucial resource and labor force in daily life, industrial, and agricultural production. To complete specific tasks, certain muscle groups must maintain a certain level of contraction during manual labor. Sustaining the same posture for a long time can cause muscle fatigue and chronic strain diseases. Quantifying fatigue effectively and predicting its intensity can help address these issues. Therefore, this paper investigates the development of a quantitative evaluation and prediction method for muscle fatigue intensity during isometric contractions. It focuses on seven standard upper limb work postures, leveraging surface electromyography (sEMG) collected from these postures. The study extracts and selects time-domain and frequency-domain features to identify characteristic parameters that effectively represent the fatigue state. Based on these parameters, a back propagation (BP) neural network-based muscle fatigue intensity classification model is established, enabling the categorization and quantitative assessment of muscle fatigue levels. Considering the temporal nature of fatigue progression, a nonlinear autoregressive dynamic neural network model is constructed for predicting fatigue characteristics. Bayesian regularization is employed as the training algorithm following comparative selection. By integrating the fatigue classification model with the feature prediction model, the study enables the forecast of muscle fatigue severity at future time points, achieving an accuracy rate of up to 95%.

Pathological myopia (PM) is a serious public health issue which potentially leads to severe visual impairment. Traditionally, fundus color photography (FCP) is employed to detect PM for its precision and image quality. However, manual analysis of FCP is time-consuming and susceptible to errors. Current automated detection methods lack the necessary granularity in classifying different stages of PM lesions. To address this problem, we developed an intelligent method for automatic PM detection, utilizing Vision Transformer (ViT) to perform detailed classification of PM. This system achieved an accuracy of 98.66% and an area under the curve (AUC) of 99.98% in detecting PM. Compared to traditional ophthalmological diagnostic methods, this model demonstrates higher efficiency and accuracy

Thyroid nodules exhibit variations in size, shape, and location, while ultrasound images feature intricate textures and spatial structures, hindering traditional convolutional neural networks from accurately extracting edge information, thus impeding complete nodule segmentation and affecting segmentation accuracy. This paper introduces a novel segmentation method named Fusion of Multi-scale Feature Net (FMF-Net). Built upon residual networks, this approach can extract multi-scale image feature information and leverage self-attention mechanisms to introduce global in-formation into deep features, enhancing segmentation performance. By introducing the Fusion Feature Block (FF-Block), this method integrates global and local feature information at different scales, achieving more precise segmentation results. Experimental results on the publicly available dataset TG3K demonstrate that the proposed method achieves a Dice coefficient of 0.9222 and a Hausdorff distance (HD) index of 0.935. Comparative experimental results demonstrate that FMF-Net exhibits more reliable segmentation accuracy in addressing complex thyroid nodule ultrasound image segmentation challenges compared to existing methods.

This paper introduces an integrated deployment and operation system for a pulmonary CT image analysis model aimed at patients with Acute Respiratory Distress Syndrome (ARDS). The system is based on Inspur Cloud servers and Cambricon MLU220 edge computing boxes, designed to assist in the clinical examination and targeted treatment of ARDS. The research team uses Xshell and Xmanager for remote connection to cloud servers without a graphical interface to deploy the model. They have developed an interactive interface for deployment and operation on mobile servers or edge computing boxes with graphical interfaces using a pseudo-encapsulation method. The paper details the preparation of the system and hardware, the installation of the 3D Slicer software and plugins upon which the model relies, as well as the deployment process based on Inspur Cloud servers and Cambricon MLU220 edge computing boxes. Experimental results show that edge computing boxes with acceleration cards have a significant advantage in terms of running time, with error control within 0.5%.The integrated deployment and operation system is simple to operate, with a user-friendly interface, capable of real-time stable transmission of remote data, while ensuring the security of user data. However, the study also points out the shortcomings of the system, such as a limited number of experimental cases and the complexity of cloud serverside deployment, and suggests directions for future improvements.

Deep learning has emerged as a powerful tool for accurate and efficient registration of medical images. However, traditional evaluation methods based on Euclidean distance (MSE) or normalized cross-correlation (NCC) suffer from low contrast or large deformation of the image. To address these issues, inspired by the concept of Large Language Model (LLM), this work trains a feed-forward neural network on multi-site medical image datasets to construct the Perceptual Loss which is applied to the registration process replacing the traditional loss functions. The method of ours is validated on the LPBA40 and improves the Dice value by 2% compared to the second ranked method. At the same time, faster convergence and smoother deformation fields are achieved by this method, which adapts to handling complex deformations effectively. Our code can be found at https://github.com/nicetry0724/smilecode.git.

Invasive ductal carcinoma (IDC) represents the most prevalent form of breast cancer. Automatic and precise IDC grading is essential for the clinical assessment of tumor status. However, current methodologies struggle with accurate grading due to the intricate spatial structure within IDC tumor regions, coupled with significant similarities between different classes of images and high variability within the same class, posing a substantial challenge for IDC grading. To tackle these challenges, we propose a novel multi-scale attention model named MS-HiFuse. This model integrates the multi-scale convolution and the multi-head attention mechanism into the local feature block of the HiFuse network, creating an enhanced multi-scale local feature block that more effectively captures the nuanced features of the tumor area and facilitates the learning of fine-grained feature representations. Furthermore, to overcome the limited number of original data and to ensure data quality, we augment the dataset through operations such as maximum rectangle cropping, subdivision of plots, and image flipping for each sample. Comparison experiments demonstrate that our proposed MSHiFuse network attains an AUC of 80.84%, outperforming both the current competing networks and the original model. Hence, the application of the MS-HiFuse network proposed herein holds significant promise for the grading of IDC pathology images.

Dynamic Visual Acuity (DVA) is a component of visual function that pertains to the ability to track moving targets. Traditional DVA assessments, which utilize moving dots or letters as stimuli, are subjective in nature. However, with advancements in brain-computer interface techniques, the steady-state visual evoked potential (SSVEP) has been employed for objective visual acuity assessment. This has prompted researchers to explore the use of moving stimuli paradigms in DVA evaluation. While previous studies have investigated factors such as moving speed and direction, stimuli pattern, and starting phase, they have not provided specific stimulus design suggestions for DVA assessment purposes. In this study, we approached the issue from a color and frequency perspective and designed stimuli that flicker at selected frequencies—7.5Hz, 9Hz, 11Hz, and 13Hz—while also moving at five different speeds: 4 deg/s, 8deg/s 12deg/s ,20deg/s ,and30deg/s . A total of forty stimuli were created by combining different colors,speeds,and frequencies,and each subject underwent five runs with these stimuli.Ten subjects participated in this trial. Furthermore,Cannonical correlation analysis(CCA)and signal-to-noise ratio(SNR)were used to evaluate EEG performance when presented with black-white and red-green stimuli under identical movement speeds and flashing frequencies.The results revealed that black-white at 11Hz was found to be the most effective combination of color and frequency regardless of speed; whereas red-green was more effective under a frequency of7.5 Hz.One interesting finding was that CCA identified peak frequency shifts from the target while the stimulus was in motion.The likelihood of frequency shift increased as stimulus speed increased.

The increasing popularity of artificial heart pump therapy due to insufficient donor numbers and the growth of patients with advanced heart failure. The size of artificial heart pump leaf is one of the important parameters of artificial heart pump. The purpose of this study is to simulate and analyze the flow field of artificial heart blood pump by using dynamic grid technology in CFD and on the basis of existing medical data, and improve the flow state of blood in the blood pump by improving the impeller inlet structure, so as to enhance its anti-hemolysis performance. For this purpose, key parameters (such as tip clearance) of artificial heart pump blades were optimized based on CFD flow field simulation. The result showed that decreasing the tip clearance from 100 μm to 60 μm, would cause a stability in the pump efficiency. The study shows that CFD in the blade size of artificial heart pump is beneficial to the optimal design of blood pump chamber

Breast cancer, as the most common malignant tumor in women, poses great harm to women's health. Among breast cancer patients, there is a relatively high proportion of HER2-positive cases, making the prediction of HER2 status extremely important. Currently, many deep learning methods based on Whole Slide Images (WSIs) have been proposed, but these methods still face the problem of imbalanced datasets, with significant differences in the number of samples in each class, causing the model to overfit categories with an advantage in quantity while ignoring minority categories. Therefore, to address the above issues, this paper proposes a weakly supervised Dual-tier Cluster-based Multiple Instance Learning (DCMIL) framework for the classification of breast cancer images. This framework uses ResNet-101 as the backbone and optimizes the pseudo bag partitioning process by introducing clustering mechanisms in the DTFD algorithm. A series of experiments were conducted on the HER2 dataset. The results showed that our model outperforms the latest algorithms in terms of performance.

The monitoring of respiratory holds significant importance in clinical applications, encompassing parameters such as respiratory rate, tidal volume, and maximal airflow velocity. Respiratory monitoring is categorized into contact and noncontact methods. Non-contact respiratory monitoring, owing to its minimal interference with the user's breathing and comfort, has garnered considerable attention in research. However, motion artifacts significantly affect respiratory monitoring, leading to a substantial reduction in the accuracy of calculation of tidal volume, respiratory rate and so on. Hence, the suppression of motion artifacts is a crucial aspect of non-contact respiratory monitoring. This paper reports a method for suppressing motion artifacts in respiratory signals obtained from fiber optic sensing mattress. The proposed method uses empirical mode decomposition (EMD) to suppress motion artifacts by decomposing and reconstructing the signals. The method demonstrates effective suppression of motion artifacts, resulting in an improvement of 0.015 in the Pearson correlation coefficient between the respiratory signal after suppressing motion artifacts and the standard respiratory signal obtained from airflow sensors.

Incomplete multi-view clustering (IMVC) for liver tumor CT images provide common and complementary information, that essential to support liver tumor diagnosis of without labeling. These IMVC techniques still have two problems, though: 1) they only heed the following to the resemblance between the same samples of different views, ignoring the resemblance between different samples of the same view, which can provide more information for liver tumor diagnose; 2) most of the existing diagnostic processes are staged, greatly increasing the cost of time. We suggest a solution to these problems by a dual contrastive learning framework to learning mutual information from cross-view contrast learning and intra-view contrast learning at the same time. On this basis, the cosine similarity between views is constructed using positive and negative samples from these dual contrastive learning, and the whole process is carried out synchronously. In this paper, two multi-view liver tumor data sets are compared with 4 methods in recent years, indicating that this method is the most advanced.

Coronary artery disease (CAD) is a cardiovascular disease characterized by coronary stenosis or occlusion due to atherosclerosis, which may result in a number of symptoms, including myocardial ischemia, angina and heart failure. Coronary computed tomography angiography (CCTA) is a diagnostic assessment for CAD. Radiology encompasses a vast amount of quantitative, high-dimensional features and transform medical images into a rich dataset that can be explored for insights. This study introduces an approach leveraging radiology features for the automated detection of coronary artery stenosis. We extract curved planar reconstruction (CPR) images along with the segmentation of the coronary arteries from three-dimensional CCTA images and extract radiomic features from the segmented regions of interest. Considering the high-dimensional nature of radiology features, we utilize techniques like LASSO regression to reduce the dimensionality of these features. We construct a graph convolutional network (GCN) block to fuse radiomic features and deep features embed this block within an encoder-decoder network. In the visualization analysis of coronary radiology features, there is a qualitative distinction between lipid and calcification regions, demonstrating the diagnostic value of radiology in coronary stenosis detection.

In view of the shortcomings of existing medical information intelligent question-and-answer systems in terms of data sources, user needs understanding and answer accuracy, this research aims to build a medical information questionanswering system based on deep learning and knowledge graph. Through the utilization of the BERT-GCN model, as well as knowledge graph, intelligent question-and-answer, and network crawler technologies, we have gathered, extracted, and integrated medical data from the Internet extensively. The system's key features include a knowledge graph, keyword matching, and automatic question-and-answer capabilities, facilitating the retrieval and visualization of medical information knowledge. By efficiently answering questions, users are able to access the necessary medical knowledge and knowledge graph in a user-friendly manner. This system effectively assists patients and healthcare professionals in accessing, comprehending, and utilizing medical information, thereby advancing the progress of the healthcare industry

In the realm of advanced cognitive processes, animals employ both egocentric and allocentric reference frames for navigation. While prior studies have largely concentrated on differentiating and transitioning between these frames, the present research explores the concurrent use of both strategies in navigation tasks. By simulating the entorhinal-hippocampal neural circuit and applying reinforcement learning (RL), we developed three agents characterized by egocentric, allocentric, and a novel integrated strategy. Testing across diverse navigation scenarios revealed that the integrated model substantially improved navigation adaptability and efficiency. Despite the challenges associated with the "black box" nature of RL, the artificial combination of individual reference frames notably enhanced performance. These results shed light on novel mechanisms of biological spatial information processing and the flexible organization of navigation strategies

Depression poses significant threats to individuals’ health and even life safety. However, once accurately diagnosed, patients can receive effective treatment. Currently, screening for depression primarily involves questionnaire assessments, blood biochemical tests, and imaging studies, but the definitive diag- nosis is determined by questionnaire assessments, with blood biochemical tests and imaging studies serving as supplementary diagnostic methods. Recent advancements in medical research have revealed that various emotions can trigger specific physio- logical, biochemical, and behavioral responses. These responses, through underlying muscle activity and microvascular circulation changes, alter body temperature distribution, thus positioning body temperature distribution as a potential indicator for de- tecting somatization disorders related to depression, offering a new perspective for imaging studies. This article introduces a diagnostic model for somatization depression and anxiety disorders based on infrared images by training with human body images captured by an infrared thermal imager and utilizing an improved YOLOv5 algorithm for human body part detection, followed by extracting lesion features for secondary classification. Testing has shown that this algorithm has a high accuracy rate of 82% in screening for depression. As a means of assisting in the diagnosis of depression, such a combination of imaging studies and deep learning not only achieves real-time detection but also significantly reduces the workload of doctors

This research presents a machine vision system for automated test tube identification and sorting in medical laboratories, improving upon manual methods that are error-prone and inefficient. The system employs digital image processing to detect specimen presence via tube caps, enhancing accuracy and throughput. The setup includes a transparent test tube rack with CCD cameras capturing images against a black background. Images are processed from RGB to HSI format for better color differentiation and smoothed to reduce noise. Edge and circle detection algorithms pinpoint the locations of test tubes for robotic arm sorting. Experiments across various filling scenarios validate the system's effectiveness. The system promises to reduce human error, optimize lab workflows, and increase automation in healthcare diagnostics.

Liver cancer is one of the most common cancers in the world, and the most important current clinical treatment for liver cancer is surgery. Posture matching with augmented reality navigation for laparoscopic hepatic resection facilitates the surgeon to achieve precise liver resection, and the liver ridgeline and falciform ligament of intraoperative laparoscopic images are very important features. Since the traditional methods of extracting liver ridgeline and ligament are extracted from liver CT images or surface grids, there are drawbacks such as many manual interactions. In order to solve the above problems, this paper proposes PointNet-based feature extraction of falciform ligaments and ridgeline in preoperative 3D point clouds of the liver, which is able to extract falciform ligaments and ridgeline directly on the 3D model of the liver, and the extraction results are concise, clear and efficient.

For reasons that the cadavers infected with Ebola are contagious and infectivity varies with age, we establish an Ebola virus epidemic model with stochastic perturbation and age structure. First, by constructing a appropriate Lyapunov functions and then applying Ito's formula, the existence and uniqueness of global positive solution of this stochastic model are derived. Moreover, studied the existence of a sole stationary distribution that leads to the threshold R^s₀ , it reflects whether the disease is spreading. If s 0 R  1 , the disease disappears, and when R^s₀ 1, the disease will be prevalent for a long time. In addition, we give some sufficient conditions for disease extinction. At last, numerical simulation verifies the feasibility for the model.

Human Papillomavirus (HPV) is a prevalent sexually transmitted virus and skin-to-mucous membrane contact transmitted virus that infects a significant proportion of the global population. In this paper, we propose a mathematical model based on the pathology development of HPV to explain the difference in HPV viral load among patients in different stages of infection results in variations in infectivity. Compute the disease-free equilibrium point and basic reproduction number utilizing the next generation approach. We performed the normalized forward sensitivity index and LHS-PRCCs to analyze the basic reproduction number, determine the key factors influencing HPV transmission. The results indicate that immune decline and the rate of recovery related to host immune response as well as vaccination plays a significant role in suppressing HPV transmission. Based on this, the effective control of transmission strategies and suggestions for enhancing protective measures are proposed to suppress the spread of HPV.

The existing named entity recognition model cannot accurately identify the entity about Lactic Acid Bacteria(LAB), and there is a lack of an annotated corpus in the task of lactic acid bacteria named entity recognition. Thus we have constructed, for the first time, a LAB-related biomedical entity annotated corpus based on the Medical Subject Headings(MeSH) of the National Center for Biotechnology Information (NCBI). This corpus includes five types of entity categories, totaling 13,871 entities. Subsequently, this paper proposed a LAB-related Named Entity Recognition method (LABNER) based on BERT-BiGRU-CRF. The method first uses BERT to encode the input text, vector representations for each word were generated. Then, BiGRU is utilized to capture the dependencies of the context. Finally, CRF is used for sequence labeling to obtain the tag with the highest probability. Contrast experiments were conducted on the LAB-related biomedical entity annotated corpus. Ultimately, LABNER achieved 90.33%, 90.75%, and 90.54% in precision, recall, and F1 score, respectively

Anxiety significantly impacts individuals' lives. This study employs a machine learning framework to detect anxiety among college students in their daily environments, focusing particularly on pre-sleep anxiety. Data encompassing electrocardiogram (ECG) and triple-axis acceleration (T-ACC) were gathered from 45 college students across 254 days, and calculated the continuous RR intervals in the ECG data. Subsequently, 30 linear and non-linear RR parameters, along with T-ACC parameters, were extracted to form the physiological model features indicative of anxiety. Feature selection was conducted using Sequential Forward Selection and Leave-One-Out Cross Validation methods. A binary classification model was then developed using these selected features through a classical machine learning algorithm, achieving an optimal accuracy of 68.42%. Statistical analysis of the features indicated that anxiety elevates parasympathetic nervous activity and significantly influences heart rate variability on certain features.

With advancements in bioinformatics, an accumulating body of evidence highlights the substantial implication of piRNA in the progression of various diseases. piRNAs are acknowledged as potential therapeutic targets and biomarkers for disease diagnosis and therapy. Exploring the intricate relationship between piRNA and diseases through computational methods can streamline costs, improve efficiency, and mitigate experimental risks. However, existing computational methods face challenges related to the sparsity of heterogeneous adjacency matrix and inherent noise in similarity networks. Therefore, in this study, we propose a matrix completion-based multi-kernel learning approach for predicting the association between piRNA and diseases, termed MCMKL-PDA. The primary goal of MCMKL-PDA is to populate the sparse adjacency matrix and mitigate the impact of noise from similarity networks on model performance. Specifically, firstly, compute piRNA and disease similarity networks using data from multiple sources. Subsequently, employ bounded matrix completion (BMC) to prefill piRNA–disease association matrices. Next, employ the low-rank fast kernel learning (LRFKL) algorithm fusion similarity networks, and the Network Enhancement (NE) algorithm refines the fused similarity networks. Finally, the Random Forests (RF) methodology is applied to forecast potential associations. The experimental results unequivocally establish the superior performance of MCMKL-PDA over other state-of-the-art methods. Moreover, independent test sets further substantiate the practical effectiveness of MCMKL-PDA.

Synthetic lethal interactions are critical genetic interactions that have been discovered for identifying new drug targets and potential cancer drug combination strategies. As a targeted approach to selectively kill cancer cells, it has attracted great attention in the field of cancer treatment. However, with the rapid growth of high-throughput data, effective identification of synthetic lethal interactions remains challenging. Although many graph topology methods focus on predicting SL interactions, heterogeneous graphs that exist in the real world tend to be scale-free, where the number of high-order neighbor nodes of high-degree nodes grows exponentially. The graph embedding method based on Euclidean space may not be able to effectively capture the internal hierarchical structure of the scale-free network, and there is not enough space to accommodate the nodes, making the node embedding highly distorted, which greatly limits the modeling ability. In this paper, we propose a model named HLASL based on the hyperbolic Lorentz attention mechanism, it ensures that the learned node features follow a Lorentz manifold. Specifically, in hyperbolic space, we use Lorentzian manifolds to encode heterogeneous information from KGs and information about synthetic lethality. In addition, we introduce the Lorenz knowledge graph attention mechanism to distinguish the weights of different information from the hyperbolic space to optimize gene embedding. Experiments on two datasets show that HLASL outperforms several state-of-the-art baseline methods in synthetic lethality prediction.

Accurate clustering of cell type classification by integrating multi-omics data is necessary for subsequent data analysis. However, related algorithms face difficulties in dealing with noise and missing values present in multi-omics data. The network-based integrative clustering algorithm (NIC) can overcome the problem of data heterogeneity but does not fully utilize the spatial structure information in the data. To solve this problem, this study proposes a new algorithm-graph regularization-based NIC (GR-NIC) to improve the integration accuracy, remove redundant features, and preserve the geometric structure of the data by fusing single-cell multi-omics data, single-cell sequencing data (scRNA-seq or scATACseq) and histone modification data. To verify the effectiveness of the GR-NIC algorithm, this paper compares three state-of-the-art multi-omics clustering algorithms through five datasets. The results show that GR-NIC can select features that are more representative of the structure of the data and improve the accuracy of clustering. This new approach is effectively validated on mouse brain and peripheral blood sample datasets and shows significant improvements.

In our research, we obtained resting-state functional magnetic resonance imaging (rs-fMRI) data from 23 newly diagnosed, drug-free Major Depressive Disorder (MDD) patients and 20 age- and sex-matched healthy controls (HC). We examined Amplitude of Low Frequency Fluctuation (Alff) and fractional Alff (fAlff) across various frequency ranges to assess frequency-specific neural modifications associated with MDD sufferers. The investigation was conducted across the entire brain, avoiding the limitations of seed-based approaches, which are constrained to predefined brain areas. We found that certain regions, particularly those with varying spatial patterns across frequencies like the medial prefrontal cortex, showed more pronounced abnormalities in specific frequency ranges. These findings suggest improved methods for detecting functional brain changes in MDD.

This research focuses on establishing a mathematical model to delve into the variation in gender ratios of sea lampreys (Petromyzon marinus). The aim is to comprehend their dynamic adjustments under different environmental conditions and provide decision support for fisheries management and ecological conservation. Firstly, using Analytic Hierarchy Process (AHP), we assigned weights to the influence of different resources on larval development speed, revealing the relationship between gender transformation rates, developmental speed, and environmental resources through a mathematical model. We established a mathematical model of the sex ratio adjustment process of Lampreys and an Adaptive sex variation (ASV) model, in which the ASV model was combined with Lotka-Volterra and Logistic model. By using Matlab to solve the model, two important conclusions are drawn. On one hand, sea lampreys contribute to ecosystem stability by controlling the population of other fish species, presenting a scenario conducive to overall ecological balance. On the other hand, the research illustrates a scenario where parasitic behavior leads to a drastic decline in fish populations and the collapse of the food chain, causing harm to ecosystem stability. In conclusion, the model provides a valuable strategy for fisheries management in understanding dynamic changes of sea lamprey populations under different environmental conditions.

In response to the demand of consumers for improving the comfort of electric vehicle driving, this article considers typical road scenarios and vehicle driving parameters, and designs an experiment to evaluate the degree of passenger motion sickness. Based on actual vehicle test data, the correlation between subjective motion sickness degree of passengers and objective physiological indicators is analyzed, and objective physiological indicators strongly related to subjective motion degree are identified. Subsequently, sensitivity analysis was conducted on the factors that affect correlation, and the trend changes of various physiological indicators with the degree of passenger motion sickness were explored. Research has found that there is a significant correlation between objective physiological indicators such as pupil diameter of eye movement signals and EDA of electrodermal signals and subjective degree of motion sickness, and this correlation is more sensitive to age. As subjective degree of motion sickness increases, pupil diameter and EDA of electrodermal signals show a clear trend. At the end of this article, a graded evaluation model for passenger motion sickness degree was constructed by weighting the combination of subjective and objective evaluation indicators, achieving scientific quantification and testing evaluation of passenger motion sickness degree, providing a reference for improving user driving comfort.

A proposed transcutaneous monitoring system utilizing a CO₂ sensor with ion-selective electrodes for monitoring transcutaneous carbon dioxide partial pressure (TcpCO₂) is detailed in this study. The system incorporates an STM32F103 processor as the control chip, designed with a modular solution. The hardware circuit design scheme include power supply circuit, PID temperature control circuit, signal conditioning circuit, digital-analog conversion. Additionally, an Microsoft Foundation Classes(MFC)-based human-computer interface is designed for data communication, display and storage. The above system was validated through experimentation, demonstrating that the designed TcpCO₂ monitoring system can achieve a temperature measurement accuracy of ±0.1o C and a temperature control accuracy of ±0.2o C. Furthermore, the system was able to detect changes in partial pressure of CO₂ (pCO₂) within the range of 0-100 mmHg, with a detection limit(LOD) of 1.38 mmHg and a relative standard deviation(RSD) of 7.07% for pCO₂. Furthermore, the electrochemical workstation system, initially small and simple, has been enhanced with the addition of a weak current detection channel and a constant potentiostat circuit, thereby enabling the system to broaden its functionality. Ultimately, this system shows promising potential for application in clinical medicine and health monitoring

The fusion of single-cell multi-omics data not only provides us with a more comprehensive and in-depth understanding of cellular states and biological processes, but also greatly facilitates the study of the full range of intracellular regulatory mechanisms. However, single-cell data usually contain much noise, making data analysis difficult. The Joint SemiOrthogonal Nonnegative Matrix Factorization (JSNMF) algorithm can effectively reduce the noise in different histological data from the same sample by applying the coherence graph fusion method to achieve the integration of two histological data. However, JSNMF takes a step of dimensionality reduction followed by clustering when dealing with single-cell data, which limits its ability to uncover unique patterns between different datasets. In this study, we propose an improved method, Joint Learning Dimension Reduction Semi-Orthogonal Nonnegative Matrix Factorization (JLDRSNMF), which mines unique patterns among different datasets by incorporating Joint Learning Dimension Reduction into the JSNMF algorithm to improve the accuracy of the model. Meanwhile, we applied JLDRSNMF to five multi-omics datasets and compared it with two other JSNMF algorithms. The results show that JLDRSNMF can effectively integrate the information of the same cell from different histological data and significantly improve the accuracy of cell labeling prediction, thus greatly facilitating the subsequent analysis.

Eye tracking technology is a critical component in human-computer interaction (HCI). For HCI systems that incorporate eye tracking, the precise classification of eye movement patterns is crucial. Yet, current deep learning methods face several difficulties such as data scarcity issue and class imbalance problem. To overcome these challenges, we introduce the ForestEye Classifier (FEC), which features a dual-component strategy: (1) Feature Extraction: The utilization of a multiscale time window approach enables the extraction of features spanning a variety of temporal scales. This method is crucial for the precise classification of different patterns, effectively addressing the challenge of data scarcity. Specifically, the classification of smooth pursuit relies on features from extended time scales. (2) Classification Model: FEC employs the deep forest ensemble learning method and utilizes a layered-ensemble architecture to integrate multiple forests, thereby overcoming the class imbalance problem and achieving improved classification performance. This marks a pioneering application of ensemble learning in the field of eye movement classification. In the end, the experiment indicates that, compared to the existing deep learning-based models, FEC demonstrates superior F1 scores (classification accuracy) across all eye movement patterns and within two distinct datasets. Moreover, FEC demonstrates consistent and robust performance across a range of participants.

This study aimed to elucidate the temporal trends in overall survival (OS) and disease-specific survival (DSS) among cancer patients, establish associated Cox proportional hazards (Cox PH) for risk stratification, and develop predictive models using deep learning artificial neural networks (ANN) and machine learning random forest (RF) methods. Data from the Surveillance, Epidemiology, and End Results (SEER) database from 2004 to 2020 were analyzed, comprising 682,917 patients diagnosed between 2004 and 2014. Prognostic factors influencing OS were identified using various statistical methods, including Cox regression analyses and machine learning algorithms. Five-year survival prediction models were developed and evaluated for both OS and DSS, with the optimal models achieving high predictive performance (AUC of 0.899 for OS and 0.923 for DSS). The study revealed an increasing incidence of cancer in the United States from 2004 to 2014, particularly among middle-aged and older individuals, highlighting cancer as a growing global health concern. Older age at diagnosis was associated with lower survival rates, emphasizing the importance of age in prognostic assessments. Furthermore, factors such as gender, tumor characteristics, and treatment regimen significantly influenced patient outcomes. The developed prognostic models offer valuable insights for clinical decision-making, enabling the calculation of expected survival rates under different treatment regimens, thus aiding in personalized treatment selection.

Congenital Heart Disease(CHD) complicated with Pulmonary Artery Hypertension(CHD-PAH) is one of the more harmful CHD-related complications. The objective of this paper was to develop a algorithm to study CHD-PAH from the perspective of heart sounds. This could enable adolescents to be detected and treated early without the need for expensive equipment or specialist skills. For this purpose an algorithm was designed. According to the pathological signs of CHDPAH in terms of heart sounds, i.e., it is mainly reflected in the hyperactive P2 of pulmonary valve component and the presence of A2 and P2 splitting. An image generator was used to feed the Choi-Williams distribution features of the Second Heart Sound(S2) into a convolutional neural network for training. A accuracy of 94.1% and a sensitivity of 94.9% were achieved using novel signal processing techniques and an ensemble of neural networks as classifier.

This study proposes a Bayesian agent based on a hierarchical Bayesian model and Bayesian decision theory as a mentalizing model of players involved in the prisoner’s dilemma, aiming to simulate the decision-making process of humans in the game. By mimicking the cognitive reasoning process of humans, this model seeks to enhance the understanding of cooperative and betrayal behaviors of biological agents in the Prisoner’s Dilemma, thus providing a robust computational model framework for social cognition research. Moreover, we demonstrate through simulations that the model effectively tracks its own state and that of its partner, as well as makes effective gaming decisions.

With the exacerbation of global population aging, the issue of falls among the elderly is increasingly drawing attention from various sectors of society. Consequently, the development of an effective human fall detection system is crucial for timely identification of fall incidents and provision of emergency assistance. This paper proposes a fall detection method based on non-contact sensors, aiming to achieve real-time monitoring and accurate identification of fall events through advanced sensor technology and data processing algorithms. Initially, this study employs low-frequency ultra-wideband radar (UWB) as the primary tool for data acquisition, followed by the utilization of deep learning techniques to extract key dynamic features from radar data, thus providing a reliable foundation for fall event identification. Furthermore, this paper designs and implements a fall recognition model based on Convolutional Neural Networks (CNN), which, through training on a large amount of radar image data, learns to distinguish patterns between normal activities and falling behavior. The results demonstrate that the proposed method can effectively detect falls under various common daily activities, with a recognition rate as high as 99.338%.

With the increasing prevalence of adolescent idiopathic scoliosis, it is particularly important to understand the mechanical mechanism of paravertebral muscles in patients with scoliosis. The purpose of this study is to establish a personalized scoliosis model through OpenSim software and to simulate the mechanical behavior of paravertebral muscles in static flexion and extension postures of scoliosis patients, to deepen the understanding of the musculoskeletal system of scoliosis patients. Numerical simulation analysis provides theoretical support for symptomatic treatment and provides an accurate basis for clinical evaluation and intervention treatment.

Addressing the limitations of computational and storage resources in the current Internet of Medical Things (IoMT), this study establishes a model of an IoMT heart disease classification system leveraging UAVs equipped with edge computing. A novel algorithm based on deep reinforcement learning is proposed to optimize system energy consumption and reduce delay as the primary objectives within a multi-constrained optimization framework. This optimization problem is formulated as a Markov Decision Process (MDP). Subsequently, a rational reward function is designed in accordance with the optimization objectives to learn the optimal task offloading strategies by maximizing long-term rewards. The model uses a dataset for classifying heart disease to train the classification model. Simulation experiments demonstrate that the proposed algorithm outperforms DDPG and SAC algorithms in terms of performance while effectively reducing system energy consumption and delay.

Optical Coherence Tomography Angiography (OCTA) provides an innovative method for obtaining information on the depth of retinal vascular blood flow. However, the accuracy and efficiency of OCTA vessel segmentation limits its clinical application. To address this problem, we explore a novel model to generate vascular saliency maps based on random walks, thereby improving the accuracy and stability of segmentation. On this basis, three salient features are innovatively used to generate seed points, which are combined with random walks to generate a saliency map and incorporated into a level set image segmentation model to propose a saliency-guided retinal vessel segmentation algorithm to solve the application problems effectively. The results of experiments conducted on OCTA-DB image datasets demonstrate that the proposed algorithm is capable of markedly enhancing the precision of vessel segmentation.

Building knowledge graphs based on relational databases has increasingly become a research focus in the field of knowledge engineering. This paper explores the mapping methods for converting relational data into graph data, using the construction of clinical knowledge graphs from electronic medical record (EMR) relational data as a case study. By analyzing the differences in data expression structures between EMR relational data and knowledge graph models, the paper proposes a semantic model for mapping relational data to graph data that meets application semantic requirements, defines rules for converting relational data with varying granularity and semantic constraints into graph structures, and designs a mapping algorithm where the order of mapping is determined by node degree. Experiments demonstrate that the proposed method can effectively implemente knowledge graph construction based on relational data.

Prodigiosin is a kind of tripyrrole cyclic pigment produced by diverse microorganisms with a wide spectrum of biological activities and extensive applications. Here, bibliometric analysis was used to evaluate the biological activities, mechanisms, and applications of prodigiosin. A total of 1,227 related publications between 1990 and 2023 collected from the Web of Science Core Collection database were extracted for in-deep research. The results showed that the number of publications of prodigiosin was increasing rapidly. Bibliometric analysis showed that change for the growth conditions of the strain influenced prodigiosin yields, and its unique molecular structures with pyrrolyl pyrromethane provided a wide variety of biological activities. Applications and mechanisms of prodigiosin for anticancer, antibacterial, antimalarial, and immunosuppressive are important research topics and hotspots in medicine, new energy, and cosmetic field. However, many challenges are still around the field of prodigiosin. High production cost has limited the application and some regulation mechanisms remains unclear. The bibliometric analysis in this study would provide a reference for researchers, and support the development of prodigiosin in the future.

Hepatitis B virus (HBV) infection is a major risk factor for hepatocellular carcinoma. Currently, there are still many limitations in the therapeutic approaches for HBV infection. During the antiviral process, inhibiting viral transmission, blocking the viral replication cycle, and slowing its progression to hepatocellular carcinoma are the keys to treatment. With this goal in mind, this paper proposes a new therapeutic strategy in addition to antiviral therapy for patients who are infected with HBV but have not yet become cancerous by mining transcriptomics data from relevant samples. Our contribution is mainly to find specific receptors and ligands in the cell cycle of HBV-infected patients that can be used as targets for drug administration, which is useful for understanding and intervening in the progression of hepatitis B to hepatocellular carcinoma, and also brings a fresh perspective to antiviral therapy.

Variant novel peptides in cancer samples can provide guidelines and evidence for cancer research. In order to obtain more valuable novel peptides, protein databases need to contain more non-canonical proteins. However, the dramatic increase of database brings high false positive rate of peptides, the identification of novel peptides faces great challenges. We filter peptide results with high false positive rates by combining peptide retention time and peptide fragment ion informations using the prediction tool DeepLC and the validation algorithm SpectrumAI. Both methods are currently integrated into the pypgatk Python package and provide multiple parameters to meet individual needs.

In recent years, Computational Fluid Dynamics (CFD) simulation software has become increasingly prevalent in the study of hemodynamics. The choice between artery bypass and stent surgeries has long been a subject of debate within this field. This paper use ANSYS FLUENT for numerical simulations to investigate the flow dynamics post-surgery, aiming to compare the advantages and disadvantages of these surgical methods and thus inform decision-making for medical professionals. The simulations reveal that post-artery bypass surgery, blood flow near the vessel wall is accelerated while it slows in the central region, which could effectively mitigate atherosclerosis. However, due to external pressure on the vascular adventitia, the internal diameter in the stent region is reduced compared to a normal vessel, leading to a highspeed flow area within the stent that enhances blood circulation. Ultimately, the choice of surgical intervention should consider the patient's specific conditions, including age, disease duration, and severity, to determine the most suitable approach. It's crucial to recognize that surgery is merely one component of treatment; adopting healthy lifestyle habits is essential to prevent the recurrence of vascular blockages.

Nowadays, lymphoma has become a tumor type characterized by high morbidity and a low survival rate. Lymph spreads throughout the body, and the tumor's volume and size are irregular, making it cumbersome for doctors to diagnose. At the same time, artificial intelligence is developing rapidly, and its application in the medical field is becoming more and more extensive, so it is possible to detect lymphoma and assist doctors' work by using artificial intelligence. The purpose of this paper is to automatically complete the detection of lymphoma segmentation and the calculation of prognostic indicators with the help of artificial intelligence, and to assist doctors in determining a good treatment plan, as well as assessing the metabolic information of the tumor, predicting the progression of the disease, and reducing the workload of doctors. The experiment used the NIFTI data of 45 patients from Inner Mongolia Autonomous Region People's Hospital, and the data were labeled under the guidance of doctors. The data were divided into training and test sets at a 7:1 ratio, and 8-fold cross-validation was performed. The segmentation of lymphoma was performed by a 3D UNET network, and three prognostic indicators, tumor metabolic volume TMTV, maximum diffusion distance Dmax, and total lesion glycolysis ratio TLG, were measured based on segmentation. In the experiment, the training DSC was 90.50%, and the validation DSC was 86.80% after 350 rounds of Epoch training. The researchers measured the three prognostic indicators simultaneously. The experiment proved the feasibility of using the method of artificial intelligence for the detection and evaluation of lymphoma.

This paper aims to conduct a visual analysis of the research status and emerging trends of hospital pharmacy performance appraisal. The paper retrieved 366 literatures related to hospital pharmacy performance evaluation from the CNKI database for the years 2002 to 2022.Through the visualization analysis of knowledge graph, the development trends of publishing, author cooperation, institutional cooperation and keywords are analyzed. The results indicate that the number of documents related to hospital pharmacy performance appraisal is continuously increasing. However, the cooperation between authors and institutions remains relatively loose, with no core team formed yet. The concepts, objects, methods, and indices of hospital pharmacy performance appraisal are hot topics in this field. The researchers predict that the number of research on hospital pharmacy performance appraisal will continue to increase, with the selection of pharmaceutical performance assessment objects and indicators expanding. The paper forecasts that the construction of a pharmaceutical performance appraisal index system should be a focus for future research.

The proper use of medical data plays a vital role in the development of medical research. However, there are challenges related to privacy leakage in medical data sharing and the controllability of data by patients during its use. To address these issues, this study proposes a blockchain-based scheme for controlling access to medical data. Firstly, a medical data sharing scheme based on blockchain is constructed, incorporating a medical data sharing scheme based on proxy re-encryption. This solution achieves secure sharing of medical data while protecting the privacy and security of patient medical data through proxy re-encryption technology, with consensus nodes acting as proxy nodes responsible for proxy re-encryption. Secondly, a mechanism for dynamically adjusting data access permissions is designed. Access control policies are stored in smart contracts, allowing patients to interact with smart contracts to update medical data access permissions by modifying attribute lists, thereby achieving controllable sharing of medical data by patients. Finally, security analysis indicates that the proposed scheme can protect medical data privacy, enable dynamic data sharing, and resist collusion attacks. Performance analysis shows that the scheme has advantages in terms of communication overhead and computational costs, making it suitable for controlled data sharing between patients or hospitals and research institutions.

For a long time, researchers have been facing the challenges of population stratification in genetic association studies. Due to the potential of population stratification to cause false positives and false negatives, if not properly corrected, it may mask true association signals. In the analysis of rare variant associations, this issue can become even more challenging because rare variants are difficult to detect. To address the aforementioned issue, this paper introduces a method, the optimal weighted aggregate (C-TOWA), for assessing the effects of variants in admixed populations and detecting associations between genetic loci and trait values. The method takes into account the weights of variants that are strongly associated with the phenotype, that is, optimally deriving weights based on existing phenotype and genotype data. We assessed the execution of the C-TOWA method through wide-ranging simulation experiments. The simulation results show that C-TOWA can effectively control population stratification effects and is the most powerful in nearly all scenarios.

Atherosclerosis is a main cause of cardiovascular and cerebrovascular diseases. Blood flow velocity profile (BFVP) is usually used as an important hemodynamic index for the course of atherosclerosis. Ultrasonic speckle tracking (ST) has been widely used in the estimation of blood flow velocity profile. To solve the problem of long time-consuming and low accuracy of the conventional ST with a global search (ST-GS), an improved ST method based on the Moth-Flame optimization (ST-MFO) is proposed. Based on the Field II (namely, an ultrasound simulator), the performance of the STMFO algorithm has been evaluated, and experimental results have been compared with those of the ST-GS. The ST-MFO method reduces the NRMSE of the BFVP measurements by 17.44% on average and the running time by 18.79% on average. It indicates that MFO method can improve the accuracy and efficiency of BFVP measurement, which is of great significance to the prevention and course monitoring of cardiovascular and cerebrovascular diseases.

Alzheimer's disease (AD) is a neurodegenerative disease with an insidious onset and a brain disorder often found in the middle-aged and elderly population, posing a serious health risk to the middle-aged and elderly. For the large database of AD patients that can be used for patient diagnosis, there is a general concern that its reliance on current medical classifications leads to ambiguous results and hides important scientific information. Therefore, it is important to know how to use mathematical models to assist experts in constructing AD recognition models and accurately diagnosing the disease. This paper propose a search and diagnosis classification strategy that combines supervised and unsupervised learning to obtain data information about patients from ANDI databases using external medical knowledge, which includes structural brain features and cognitive-behavioral features. To reduce the data volume of clinical measures, after preprocessing the data, supervised random forest classification is applied to filter features with feature importance greater than 0.025, and then unsupervised learning including PCR and K-means clustering is used to create new clinical performance categories based on the features selected by the clusters. The results of the defined tri- and quintuple classification are more precise, and this paper's strategy aims to establish a connection between potential biomarkers and clinical functional performance. The screened data is also meaningful, providing a basis for interpretation and comparison for the classification and diagnosis of AD by AD clinical medicine experts and opening up more opportunities for the early detection of AD patients.

Gastrointestinal diseases have emerged as a significant factor in disability-adjusted life years and mortality in recent years. Currently, the standard method for diagnosing gastrointestinal disorders is endoscopy. Traditional invasive endoscopy not only has a long pre-procedure preparation time but also causes discomfort to the patients. The emergence of capsule robots can largely avoid the above problems. The capsule robot not only enables all-round examination of the digestive tract, targeted drug administration, and biopsy sampling but also greatly improves examination efficiency and accuracy. When robots are targeting specific areas of the intestine for examination, biopsy sampling, and drug delivery, most capsule robots are anchored by anchoring forces generated by an anchoring mechanism to enable the positioning of the capsule that robot, which in turn provides the possibility of targeted drug delivery, biopsies, and other manipulations. The current research has primarily concentrated on the design, motion control, and attitude stabilization of these robots, but there has been limited exploration of their anchoring mechanisms. The study analyzes the effects influencing the anchoring force of capsule robots, including the effective working area of the anchoring structure, the length of the anchoring mechanism, surface roughness, and intestinal peristaltic forces. The paper uses a hyperelastic model of the intestine to assess the impact of intestinal contraction forces on the required anchoring force of the capsule robot. The analysis can optimize the anchoring mechanism of the capsule robot and provide a theoretical framework for its effective design.

Artificial intelligence (AI) and the brain have close connection. The brain contains a large number of brain cells, which form a complex network of brain cells and nerves. By studying the structure of brain networks, we can further understand human cognition and provide better support for artificial intelligence and brain computer interface technology. Firstly, we analyze common structures and their advantages and disadvantages. Then, we analyze the multipath problem in brain network structures, including the number of paths and the probability of path selection. Finally, we study the impact of the number of neighboring nodes. The research shows that the brain network structure is crucial for maintaining the flexibility and reliability of the brain network, and the number of neighboring nodes has significant impact.

Background: Hematologic disease diagnosis, particularly for malignancies, heavily relies on labor-intensive and variable bone marrow smear analysis. Deep learning holds promise for accurate blood disease diagnosis. However, existing methods have challenges such as data dependencies and limited generalization. Moreover, research integration of bone marrow morphology with prognosis remains insufficient. Method: To overcome these limitations, this study proposes a novel approach that integrates convolutional neural networks (CNNs) and graph neural networks (GNNs). This combined framework facilitates comprehensive analysis of global features and spatial relationships among cells in bone marrow smears. Notably, the introduction of a learnable Transformer fusion layer enables multitask analysis for both diagnosis and survival prediction in malignant hematological diseases. Results: Validation results on public datasets demonstrate the efficacy of our approach, achieving a diagnostic AUC of 0.905±0.029 and a prognostic AUC of 0.780±0.047. These findings indicate successful mitigation of challenges in bone marrow smear analysis through CNNs and GNNs integration, potentially enhancing diagnostic accuracy and patient outcomes. Implications: The proposed method represents a significant advancement in hematologic disease diagnosis and prognosis analysis. This innovation holds promise for improving diagnostic accuracy, facilitating timely interventions, and ultimately enhancing patient outcomes. These findings contribute significantly to the evolving field of artificial intelligence in healthcare and pave the way for further enhancements in hematologic disease diagnosis and prognosis analysis.

The aim of this study is to investigate the effects of mothers' physical and mental health on infants' behavior. The model of the relationship between infants' behavioral characteristics and mothers' physical and mental indicators was established by machine learning classification methods. Eventually, 20 groups of infant behavioral characteristic types were identified. The results of the study show that the physical and mental health of mothers has an important impact on the behavioral characteristics of infants, and the machine learning method can effectively predict the behavioral characteristics of infants and provide a basis for the development of a personalized treatment plan, which is of great significance for optimizing infant development and improving the health of mothers and infants.

In recent years, with the continuous expansion of hospital scale and the development of technology, the demand for intelligent logistics in hospitals has become increasingly urgent, especially in operating rooms, efficient consumable logistics can make the surgical process smoother, more efficient and safer. This paper selects Shenzhen People's Hospital as the experimental site, and selects the operating room consumable logistics data from November 2021 to March 2022, comparing the average logistics distribution time of surgical consumables under three modes: manual delivery, single robot delivery, and multi-robot collaborative delivery. The experiment shows that the average time of using multi-robot collaborative distribution is 2.7 minutes/time, compared with manual distribution, the efficiency has increased by almost 100%.

Given the paramount importance of treatment plan side effects to patients, they should be at the core of clinical decision-making. Traditional clinical decision support systems (CDSS) often rely on clinical guidelines for their knowledge bases, causing them to overlook the significance of side effects. Current research has shown limited examples of effectively integrating treatment plan side effects with clinical guideline knowledge. To address this deficiency, we can extract relevant knowledge by mining information about side effects from literature resources. This process involves analyzing literature abstracts and identifying matching literature knowledge based on the patient's specific symptoms, which is then integrated with the knowledge in clinical guidelines. By doing so, we establish a chain of associations linking the patient, treatment plan, and side effects. Clinicians can evaluate this chain of associations and reevaluate guideline-based decisions to optimize the clinical decision-making process. In order to validate the effectiveness of this approach, we have designed and implemented a prototype system for breast cancer patients. This system showcases the importance of combining side effect information with clinical guideline knowledge and demonstrates its potential in optimizing clinical decision making through weight reassessment. Through practical application, this system can provide more comprehensive and personalized treatment recommendations, thereby enhancing overall treatment program quality.

Drug discovery affects human health and has also been one of the key research topics in the medical field. However, due to the long and complex process of drug development, it is difficult to develop a new drug. Fortunately, with the rise and development of artificial intelligence technology, many intelligent tools and algorithms have been developed to accelerate drug discovery. In this paper, we propose a method of drug discovery using knowledge graph. Firstly, we obtained data from a variety of public medical databases and constructed a knowledge graph of medical research and development containing a variety of omics data including drugs, genes, diseases, and the relationships between them. Then, in order to make better use of knowledge graphs, we use various knowledge graph embedding methods to convert the graph into a low-dimensional continuous embedding vector to capture the semantic association between entities. Finally, a case study exploring drug repurposing was conducted to demonstrate the effectiveness of knowledge graph-based link prediction for drug discovery. By integrating diverse biomedical data sources and combining different embeddings, we ensure that the semantic relationships between different entities in knowledge graphs can be fully understood by the model. Our method effectively takes advantage of the knowledge graphs, which will ultimately accelerate drug repurposing and lead to reduced costs.

Allergic asthma is one of the most common phenotypes of asthma, affecting over 330 million people worldwide. The aim of this study is to explore the pathogenesis of allergic asthma by establishing an in vitro model of allergic asthma. Firstly, a microfluidic chip model for studying human epithelial-dendritic cell co-culture has been constructed by combining microfluidics and electrospinning technologies. Then, by utilizing electrospinning technology, a PLGA basement membrane was constructed, and non-contact co-culture of BEAS-2B epithelial cells and dendritic cells on the microfluidic chip was achieved based on this basement membrane. IL-25 and IL-33, as two major pro-inflammatory factors of asthma, were simultaneously detected by combining microfluidic technology and digital ELISA, through the design of optical path. The experimental results showed that the concentrations of IL-25 and IL-33 increased after the addition of dust mite antigen and tended to normalize after the addition of therapeutic drugs. Therefore, the in vitro co-culture microfluidic chip model designed in this study has revealed the pathogenesis of allergic asthma. Moreover, the drug experiments provide important references for future clinical diagnosis and targeted drug therapy.

This research work focuses on addressing the persistent challenges faced by citizens of Qingdao, China, in accessing efficient healthcare services, including repetitive medical examinations and long waiting times during medical consultations. To resolve these issues, the study proposed the implementation of an intelligent healthcare scenario termed "One Hospital for the Whole City." This work makes use of advanced digital technologies such as big data, cloud computing, the Internet of Things (IoT), artificial intelligence (AI), and blockchain to construct a regional information platform that integrates core operation systems of different hospitals. The platform facilitates cross-institutional recognition and sharing of inspection results and enables access to electronic health records across hospitals. To evaluate the effectiveness of the proposed system, detailed performance parameters were utilized, including average waiting time, redundant medical examinations, patient satisfaction scores, resource utilization, patient outcomes, and financial performance. Statistical analysis, including hypothesis testing using t-test method, was conducted to compare outcomes between a control group and an experimental group. Experimental results demonstrated significant improvements in healthcare efficiency and patient satisfaction in the experimental group compared to the control group. Specifically, the average waiting time reduced from 45 to 20 minutes, redundant medical examinations decreased from 30% to 10%, and patient satisfaction scores increased from 3.5 to 4.5.

The active components and target points of traditional Chinese medicine are highly complex and difficult to ascertain. In recent years, computational methods have become an effective approach for predicting compound-target interactions. However, effective utilization of the topological structural features of the network remains an urgent challenge. Traditional prediction models tend to focus excessively on the relationships between nodes while overlooking the features of the edges. To address these issues, we propose a TCMCPI model to predict the interactions between herbal compounds and target points. First, we transform the CPI network into a line graph by establishing a topological structure of the edges. We then used GCN and CNN to extract features from the compounds and targets. These features are fused together to represent the nodes of the line graph, and GAT is utilized to learn the node features. Finally, we map the node features of the line graph back to the edge features in the CPI network by merging the adjacent edge features into node features. The experimental results demonstrate that TCMCPI outperforms other baseline methods in predicting interactions between herbal compounds and target proteins, and high AUROC and AUPR evaluation metrics attest to its superior predictive performance.

Keeping the balance of flexibility and stability is difficult in orthopedic surgery when the doctor and the robot operate together. This paper gives a fuzzy theory-based variable admittance control method to solve this question. For the fuzzy controller, the robot's traction force and current velocity are inputs, the damping parameters are the output, calculation rules are the fuzzy rules through expert experience, the fuzzy controller can get different damping parameters according to the doctor’s operating intention, different damping parameters represent different desired speed of the robot. The stability of the robot is characterized by the change of acceleration at the end of the arm during the dragging process, and the flexibility of the robot is characterized by the amount of traction force exerted by the doctor on the end of the arm. A 7- DOF robotic arm experimental platform was built, and the dragging experiment was carried out. The experimental results show that the stability of the fuzzy-based variable-admittance controller is improved by 54.5% over the fixed-admittance controller for the same flexibility. Under the optimal variation range of damping parameters, the fuzzy-based variable admittance-controlled robot has a maximum speed of 0.047 m/s, a maximum traction force of 9.43 N, and a maximum trajectory deviation of 1.74 mm during human-robot collaboration. These improvements effectively enhance the speed and precision of human-robot collaboration in orthopedic surgery.

To investigate the application effects and value of the refined management system for hospital consumables based on collaborative design of software and hardware, we selected the consumable management data from Shenzhen People's Hospital and divided the management modes into traditional and informatized ones. Through a comparative analysis of order confirmation time, inventory counting time, warehouse expenses, and other aspects, we aim to demonstrate that the refined management system for medical consumables significantly improves the quality and efficiency of hospital consumable management, enhancing the management benefits, social benefits, and economic benefits of medical institutions

Artificial intelligence (AI) and brain networks are two very broad and complex research fields. They each have different priorities, but there are also intersections. Some AI systems, especially neural networks and deep learning models, are inspired by the structure of the human brain. These models attempt to process information by mimicking neurons and their connections. On the other hand, simulating and understanding human cognitive processes through AI models can help us better understand the working principles of the brain. Both fields are rapidly developing and providing us with profound insights into the essence of intelligence and how the brain works.

There are disrupted brain functional connectivity (FC) patterns in subjects with autism. The traditional method for constructing brain FC networks is to compute the Pearson correlation coefficient between the average BOLD signal of regions of interest. But it ignored the complex higher-order interaction information. In this study, we propose a novel framework for the auxiliary early diagnosis of children with autism, leveraging high-order brain functional connectivity (HOFC) networks and graph attention networks (GAT). We focus on children under the age of 12, constructing traditional FC networks, two types of HOFC networks from their functional magnetic resonance imaging data. Subsequently, we construct three types of graphs based on these three FC networks, and finally, we apply GAT to learn discriminative features for autism diagnosis. Our research findings demonstrate that this method based on associated HOFC outperforms traditional FC in accurately distinguishing between autism and typically developing children, offering a promising approach to improve the accuracy and efficiency of autism auxiliary diagnosis.

Assessing the binding affinity of pharmaceutical agents to their respective protein targets is a crucial phase in the formulation of new drugs. Current research on prediction algorithms based on deep learning is extensive but primarily focuses on improvements in the representation of proteins and drug molecules, neglecting the representation of interactions between proteins and small molecule drugs, which leads to a lack of biological interpretability. Moreover, the rapid development of protein large language models has introduced new methods for protein representation. Consequently, this study introduces a novel approach for forecasting drug-target affinity, employing bilinear networks coupled with protein large language models, designated as BplmDTA. Initially, BplmDTA employs a protein language model to obtain feature vectors for the target; it also uses extended connectivity fingerprints (ECFP) to represent drug molecules, thereby obtaining drug feature vectors from the constructed embedding layer. Subsequently, the model calculates the interaction between the two feature vectors using a bilinear network. Finally, the affinity between the drug and the target is predicted through a multi-layer feedforward neural network. The findings from the conducted experiments demonstrate that the introduced model surpasses current techniques in forecasting drug-target interactions, offering enhanced efficacy and interpretability.

In this study, we introduce a novel multimodal graph representation learning framework (MGRL-DR). By integrating bilinear graph convolution, graph attention networks, and autoencoders, the representational capability of MGRL-DR is significantly enhanced. In data preprocessing, the integration of heterogeneous networks and the introduction of a molecular feature enhancement module strengthen the feature representation of drugs and diseases. Ten-fold cross-validation on four datasets (Fdataset, Cdataset, LRSSL, and Ldataset) demonstrates that MGRL-DR outperforms comparative methods across AUROC, AUPR, and other performance metrics. In the case study, MGRL-DR discovered new associations not recorded in the dataset, validated by clinical trials and authoritative data sources. This method provides an efficient new avenue for drug discovery

Purpose: To evaluate the effectiveness of digital interventions for preventing adolescent anxiety. Methods: Two researchers conducted a systematic search of randomized controlled studies of internet-based and computerized interventions for preventing adolescent anxiety in Pubmed, APA PsycInfo, Science Direct, Web of Science, Scopus, The Cochrane Library, CNKI, WanFang Data and CQVIP, and screened the articles, extracted the data and reviewed the quality of the articles based on Cochrane 5.1.0. Meta-analyses were conducted using software Review Manager Version 5.4. Results: A total of 12 articles were included in the meta-analyses. The meta-analyses showed that digital interventions were effective in preventing anxiety (g=0.07, 95% CI 0.02 - 0.12, p = 0.005) and depression (g=0.06, 95% CI 0.01 - 0.11, p = 0.02) in adolescents. Conclusion: Internet-based and computerized interventions are effective in preventing anxiety and depression in adolescents.

In order to understand the research hotspots and frontier topics in the field of telemedicine, this article obtained 395 related articles from the China National Knowledge Infrastructure (CNKI) database and used COOC14.1 software to visualize the results. The results showed that scholars have a high level of interest in the field of telemedicine. The research hotspots in this field include the application of telemedicine in the treatment of chronic diseases and its role in the construction of medical consortia. In recent years, internet healthcare and telemedicine have become frontier topics in research. The development of telemedicine not only requires policy support but also the support of technological capabilities.

The assessment of gastrointestinal health through tongue image analysis is a significant aspect of traditional Chinese medicine. Utilizing computer vision technology for the analysis of tongue image features and disease diagnosis has emerged as a focal point in medical image processing research. However, the current integration of deep learning with traditional Chinese medicine remains relatively limited, particularly in the comprehensive exploration of tongue image features based on traditional Chinese medical diagnostic theories. In this study, a variety of deep learning models were employed to perform classification tasks on the presence of common tongue features such as thick coating, cracks, tooth marks, and the existence of gastric diseases. The deep learning models utilized include CNN, ResNet, AlexNet, and DenseNet. Subsequently, DenseNet was used as the reference model to evaluate the performance of pre-training with the three tongue image features for gastric disease classification. The training and validation were conducted on tongue image datasets collected and annotated at the Department of Traditional Chinese Medicine of Peking University Third Hospital. Experimental results demonstrate that DenseNet achieved an AUROC value of 0.9207 for certain tongue image features. Different networks exhibited favorable performance in metrics such as Accuracy, Precision, and Recall. Moreover, the injection of the three tongue image features as prior information significantly enhanced the model's accuracy in identifying gastric diseases. Our research validates the feasibility of deep learning in intelligent tongue image diagnosis, laying a foundation for the digitization and intelligence of traditional Chinese medicine tongue diagnosis.

Previous studies have indicated sex differences in Alzheimer’s disease (AD). There has not yet been an experiment examining the neural mechanisms and influencing factors of brain connectivity. We gathered resting-state functional magnetic resonance imaging (rs-fMRI) data from a total of 31 AD patients, including 17 females, and 50 healthy controls, comprising 28 females. We investigated brain connectivity from multiple topological variables. The small-world coefficient is significantly higher in the female subgroup than in the male subgroup, in both AD and CN groups. The probability distribution of brain connectivity degrees in males shows more pronounced changes compared to females. There are significant differences in topological variables among groups and sex. Our results demonstrate that brain connectivity exists in the sex differences between AD and CN, which provides new evidence for the existence of physiological protective mechanisms in males against AD.

This study introduces an Adaptive Attention mechanism designed for Multiple Instance Learning (MIL), focusing on the diagnosis of cervical lymph node metastatic cancers. By utilizing data from 505 cases and 1,036 Whole Slide Images (WSIs) provided by the First Affiliated Hospital of Sun Yat-sen University, this mechanism dynamically adjusts attention weights to enhance its performance in medical imaging. By prioritizing diagnostically significant patches within WSIs, the Adaptive Attention mechanism improves the model’s accuracy and interpretability. A comparative analysis within the ABMIL framework shows that Adaptive Attention surpasses conventional attention mechanisms in accuracy (ACC), F1 score, and Area under receiver operating characteristic curve (AUC) metrics. This substantiates its superior efficacy in complex MIL tasks, notably in trinary (non-tumor, lymphoma, metastatic cancer) and binary (lymphoma versus metastatic cancer) classification tasks. The findings highlight the potential of Adaptive Attention in advancing medical image analysis and artificial intelligence, marking a significant step toward improving diagnostic precision in healthcare.

Objective:The Network pharmacology and Drug Component-Target Molecular Docking Technology were used to study the molecular mechanism of knee osteoarthritis (KOA) in the treatment of knee osteoarthritis (KOA), so as to provide theoretical support for clinical application. Methods: Firstly, the main active ingredients and targets in the decoction were retrieved through the TCMSP database, and the known targets of the unobtained active compounds were predicted according to the published literature reports and Swiss Target Prediction. The relevant targets of KOA were retrieved through GeneCards, DisGeNET, OMIM and TTD databases, and the intersection targets of knee three visceral decoction and KOA were obtained by using the Wayne diagram. Subsequently, the drug-disease-target network diagram and PPI network were constructed by the STRING11.0 database combined with Cytoscape3.9.1 software, and the core components and targets of the treatment of KOA were screened. Results: A total of 79 drug-disease intersection targets were screened, with luteolin, β-dermatone and baicalein as the core compounds of drug action, and PTGS2, ESR1, ESR2 and CDK2 as the core targets. Molecular docking verification showed that PIK3CA had the best binding activity with β-desmatone. The mechanism of the treatment of KOA mainly involves the mitogen-activated protein kinase signaling pathway and the nuclear factor-activated B cell kappa-light chain enhancer signaling pathway. Conclusion Knee three organs decoction can treat osteoarthritis through multi-component, multiple target categories as well as multiple pathways, which provides ideas and theoretical basis for further research and experimental work.

With the rapid development of the Internet, Internetmedicine has become popular, and more patients choose online consultation and doctor consultation services. This articleisbased on the combination of TCM consultation and InternetInnovate new methods of TCM diagnosis and treatment. By developing WeChat applet, patients can use the platform to communicate with doctors at any time without time constraints and geographical location constraints, improve the efficiency of doctors' consultation and patients' consultation, provide a broader platform for the combination of the Internet and TCM culture, and also make online consultation more widely used.

At present, intelligent medical serves have become an essential component in serving system. With in the evolving landscape of healthcare logistics, where the integration of sophisticated technology is paramount to enhancing operational efficiency and patient care. This design is based on STM32 microcontroller as the core, using OpenMV camera to identify the delivery route and digital ward number, using HX711 pressure sensor to detect whether drugs are placed on the car, using the deceleration motor to drive the car, will carry drugs on the car to the designated pharmacy intelligent drug delivery car. When two vehicles work together, 2.4g communication module is needed to communicate with each other and transmit key location information. In the design, the OpenMV camera is used as the identification module to complete the function of line patrol and number recognition through the method of color block search and template matching. HX711 pressure sensor is used to measure the weight of placed drugs, so as to judge whether the car is loaded with drugs. Reduction motor is used to drive the trolley, and the attitude of the trolley is controlled by speed closed-loop, and the drug delivery is completed within the specified time. When the two vehicles carry out collaborative drug delivery, the 2.4g communication module is used for the interaction between the two vehicles to complete the functions of the two vehicles avoiding collision and the designated pharmacy delivery.

Aortic dissection (AD), commonly known as acute aortic syndrome, which also encompasses conditions such as hematoma, aortic ulcer, and thoracic aortic aneurysm rupture, is the most frequent and devastating diseases. In this study, we evaluate ChatGPT in diagnosis AD and compare the ChatGPT-4.0 and the ChatGPT-3.5. We conducted a prospective study via public case report literature to investigate whether ChatGPT-4.0 and ChatGPT-3.5 can provide correct diagnosis based on clinical information and image feature information. The results were analyzed using commonly used statistical methods regarding different category patterns (Gender, Disease, Age, Admission_date_time Or Published_time). For primary diagnose, the accuracy rates of ChatGPT-4.0 were very higher than the ChatGPT-3.5 (0.856 vs. 0.782, p<0.05). For secondary diagnoses, the accuracy rates of ChatGPT-4.0 were also significantly higher than ChatGPT-3.5 (0.856 vs. 0.782, p<0.05). ChatGPT-3.5 has limitations in its ability for patient history, symptom presentation, laboratory tests, and imaging data. ChatGPT-4.0 has limitations in identifying symptoms and laboratory test data by classifying and analyzing misdiagnosis causes. Both for primary and secondary diagnose, there was significant difference in Gender, Disease, Age, and Admission_date_time Or Published_time group in ChatGPT-4.0 and ChatGPT-3.5 (p>0.05). Both ChatChatGPT3.5 and ChatChatGPT4.0 satisfactory for addressing fundamental inquiries related to Aortic dissection disease. ChatGPT has potential in medical diagnosis, and accuracy of ChatGPT-4.0 is better than that of ChatGPT-3.5 in the disease diagnose, but ChatGPT-4.0 still has limitations regarding patient symptoms and laboratory data recognition. Further studies which performed in the dynamic clinical practice environment is needed.

Objective: This research aims to analyze the mechanisms of Fuyuan Capsule in treating Type 2 Diabetes (T2DM) using network pharmacology. Methods: We initially screened the active ingredients of Fuyuan Capsule, which contains 13 traditional Chinese medicines, from the TCMSP database and retrieved the relevant targets of these active ingredients. Subsequently, we identified T2DM-related targets through Genecards and Drug Banks databases. Using Venn diagrams, we obtained the intersection of active ingredients from traditional Chinese medicines and T2DM-related targets. Furthermore, we designed a PPI network model using Strings and Cytoscape 3.9.2. Finally, we conducted GO enrichment analysis and KEGG pathway analysis using the bioconductor program in R language. Use Autodock Tools to perform molecular docking between the active ingredient and the target. Results: In our study, we identified a total of 496 targets from Fuyuan Capsule, which comprises 17 ginseng components, 15 astragalus components, 1 prepared rhubarb component, 18 Epimedium components, 8 Cornus officinalis components, 5 Poria cocos components, 4 Hedyotis diffusa components, 2 Cistanche salsa components, 56 Salvia miltiorrhiza components, 1 Curcuma longa component, 11 Cassia obtusifolia components, 5 Schisandra chinensis components, and 14 Hawthorn components. After database screening and deduplication, we identified 1,541 T2DM-related targets, with 209 being the intersection targets of active ingredients from Fuyuan Capsule and T2DM. We also identified 5 core targets, including PTGS2, PTGS1, SCN5A, ADRB2, and RXRA. GO enrichment analysis yielded 476 Biological Process (BP) terms, 85 Cellular Component (CC) terms, and 185 Molecular Function (MF) terms. KEGG pathway analysis revealed 5 core pathways, including the PI3K-AKT signaling pathway, lipid and atherosclerosis, fluid shear stress, atherosclerosis, AGE-RAGE pathway. Conclusion: Our study demonstrates that various traditional Chinese medicines within Fuyuan Capsule primarily act on multiple pathways, including PI3K-AKT, AGE-RAGE, and lipid and atherosclerosis, through effective components such as quercetin, kaempferol, naringenin, salvianolic acid A, and resveratrol, among others. These mechanisms intervene and treat T2DM through multiple approaches, including improving insulin sensitivity, regulating insulin resistance, reducing inflammatory responses, and enhancing endothelial function.

The antibody is a type of protein in the human immune system that plays an important role in identifying and neutralizing antigens. However, the known types of antibodies are very limited. The emergence of computational protein design has to some extent addressed the drawbacks of traditional design methods, and deep diffusion probability models have been proven to be capable of antibody sequence and structure design. Existing models lack a prototype system, making it more difficult to implement them in practical applications, and the operational complexity of the models poses challenges. To overcome these challenges, we have designed and developed an antibody generation system based on a diffusion model, aiming to assist users in collaborative antibody sequence and structure design. This system integrates design, relaxation optimization, and evaluation functions, and is user-friendly, enabling users to easily carry out the entire process of antibody design. Through this system, users can rapidly and accurately generate antibody sequences and structures with specific functions and characteristics, thereby improving design efficiency and success rate.

This paper delves into the current state and future prospects of federated learning (FL) in the medical field, with a focus on its key advantages in privacy protection, data efficiency, and model performance improvement. The main innovations of the paper lie in the introduction of three strategies: federated pre-training, fine-tuning, and prompt engineering, aimed at balancing data privacy and model performance optimization. Additionally, the paper discusses the major technical challenges faced by federated learning, including security threats, privacy breaches, and issues with non-independent and identically distributed (Non-IID) data, proposing the use of differential privacy and secure multi-party computation techniques to enhance data privacy and model security. Finally, the paper outlines future research directions, including designing more comprehensive security and privacy protection mechanisms for federated learning, optimizing communication efficiency, and developing more effective algorithms for Non-IID data, all of which are crucial factors in promoting the widespread application of FL in the medical field.

Objectives: Use TgCDPK1, a kind of important protein in T.gondii cells, as the target protein, by main processes of “Proteins Sequences Alignment (from Uniprot database to RSCB.pdb database)”, “Drug Molecules Docking” and “Drug Toxicology” , to select molecules as the drugs against T.gondii which are able to combine with it. Methods: Use “TgCDPK1” as the key word, searching the relevant proteins of T.gondii in Uniprot database, and mark Q9BJF5 as “marker” when acquire proteins sequences, then perform the Protein Sequence Alignment for all proteins in Clustal Omega platform; Use Q9BJF5 as “marker”, searching the relevant proteins and their sequences in RSCB.pdb database and applying the same way of Protein Sequence Alignment. When acquire a protein whose sequence is match with Q9BJF5, it’s the target protein(3HX4), and it need to download the file of 3HX4 for the next Molecule Docking; Choose the tcmsp drug database and use the file of 3HX4 to perform Molecule Docking and calculate the “combination energy(Ec)”, and finally save the molecules whose combination energy is less than -8 kcal/mol as selected drugs; Eventually according to “ Lipinski Rule of Five” confirm the appropriate molecules as potential drugs. Results: Proteins in Uniprot are highly homologous in the key site(Gly128) with Q9BJF5, and the protein 3HX4 in RSCB.pdb chosen for Molecule Docking is also highly homologous with Q9BJF5; There are 35 kinds of molecules eligible for “Ec < -8 kcal/mol”; and finally choose the “Verticinone” as a drug molecule. Conclusions: Verticione can be used as a drug against T.gondii in the aspect of Bioinformatics(Ec= -10.23 kcal/mol, combines to Gly128), and it has little toxic side effect in the aspect of Drug Toxicology(there are 4 “No” in 5 kinds of “GYP inhibitor”).

With the booming development of artificial intelligence, deep learning, and machine learning, an important branch in the field of artificial intelligence, especially the intelligent algorithms represented by deep learning, has been gradually applied in the field of medicine. Liver cancer has been the focus of research in the field of medicine, and the development of deep learning technology is crucial for making precise diagnoses, lesion classification, risk factor prediction, and prognosis assessment of liver tumors. In this paper, we review the status of research related to deep learning in the field of liver cancer diagnosis.

Automatic colony counter is an important equipment used in the field of microbiology, and its mechanical structure design directly affects its performance and reliability. In order to improve the accuracy, stability and efficiency of the automatic colony counter, the key elements and principles of its mechanical structure design are discussed. Firstly, the working principle, design tasks and objectives of the automatic colony counter are analyzed, and then the key components of the mechanical structure design, such as sample tray, optical system and moving parts, are discussed in detail. Finally, through simulation analysis and experimental verification, the feasibility and superiority of the design scheme are verified, which provides a reliable theoretical basis and practical guidance for the mechanical structure design of the automatic colony counter.

Addressing the efficiency of large-scale vaccination campaigns presents a significant challenge. This paper introduces a Smart Internet of Things (IoT)-based vaccine supply chain platform designed to enhance vaccine distribution and administration. It provides an analysis of the current situation at grassroots vaccination sites and the traditional transportation models used in China. Through keyword analysis, the research identifies hotspots in the vaccine supply chain field. The paper presents a rapid-response logistics service model focused on grassroots sites, utilizing a smart IoT platform to maximize site satisfaction. This platform supports real-time tracking, appointment scheduling, and intake control. It also details the role of third-party cold chain logistics companies in integrating IoT, big data, and advanced technologies to improve timeliness. Furthermore, the paper offers recommendations for developing standardized cold chain systems.