Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

This study aims to investigate the mechanism by which resveratrol(RES) alleviates cerebral vascular endothelial damage in sepsis-associated encephalopathy(SAE) through network pharmacology and animal experiments. By using network pharmacology, the study identified common targets and genes associated with RES and SAE and constructed a protein-protein interaction( PPI) network. Gene Ontology(GO) analysis and Kyoto Encyclopedia of Genes and Genomes(KEGG) pathway enrichment analysis were performed to pinpoint key signaling pathways, followed by molecular docking validation. In the animal experiments, a cecum ligation and puncture(CLP) method was employed to induce SAE in mice. The mice were randomly assigned to the sham group, CLP group, and medium-dose and high-dose groups of RES. The sham group underwent open surgery without CLP, and the CLP group received an intraperitoneal injection of 0. 9% sodium chloride solution after surgery. The medium-dose and high-dose groups of RES were injected intraperitoneally with 40 mg·kg-1 and 60 mg·kg~(-1) of RES after modeling, respectively, and samples were collected 12 hours later. Neurological function scores were assessed, and the wet-dry weight ratio of brain tissue was detected. Serum superoxide dismutase(SOD), catalase( CAT) activity, and malondialdehyde( MDA) content were measured by oxidative stress kit. Histopathological changes in brain tissue were examined using hematoxylin-eosin(HE) staining. Transmission electron microscopy was employed to evaluate tight cell junctions and mitochondrial ultrastructure changes in cerebral vascular endothelium. Western blot analysis was performed to detect the expression of zonula occludens1( ZO-1), occludin, claudins-5, optic atrophy 1( OPA1), mitofusin 2(Mfn2), dynamin-related protein 1(Drp1), fission 1(Fis1), and hypoxia-inducible factor-1α(HIF-1α). Network pharmacology identified 76 intersecting targets for RES and SAE, with the top five core targets being EGFR, PTGS2, ESR1, HIF-1α, and APP. GO enrichment analysis showed that RES participated in the SAE mechanism through oxidative stress reaction. KEGG enrichment analysis indicated that RES participated in SAE therapy through HIF-1α, Rap1, and other signaling pathways. Molecular docking results showed favorable docking activity between RES and key targets such as HIF-1α. Animal experiment results demonstrated that compared to the sham group, the CLP group exhibited reduced nervous reflexes, decreased water content in brain tissue, as well as serum SOD and CAT activity, and increased MDA content. In addition, the CLP group exhibited disrupted tight junctions in cerebral vascular endothelium and abnormal mitochondrial morphology. The protein expression levels of Drp1, Fis1, and HIF-1α in brain tissue were increased, while those of ZO-1, occludin, claudin-5, Mfn2, and OPA1 were decreased. In contrast, the medium-dose and high-dose groups of RES showed improved neurological function, increased water content in brain tissue and SOD and CAT activity, and decreased MDA content. Cell morphology in brain tissue, tight junctions between endothelial cells, and mitochondrial structure were improved. The protein expressions of Drp1, Fis1, and HIF-1α were decreased, while those of ZO-1, occludin, claudin-5, Mfn2, and OPA1 were increased. This study suggested that RES could ameliorate cerebrovascular endothelial barrier function and maintain mitochondrial homeostasis by inhibiting oxidative stress after SAE damage, potentially through modulation of the HIF-1α signaling pathway.
Animals ; Mice ; Network Pharmacology ; Resveratrol/administration & dosage* ; Male ; Sepsis-Associated Encephalopathy/genetics* ; Signal Transduction/drug effects* ; Hypoxia-Inducible Factor 1, alpha Subunit/genetics* ; Endothelium, Vascular/metabolism* ; Molecular Docking Simulation ; Protein Interaction Maps/drug effects* ; Humans ; Sepsis/complications* ; Oxidative Stress/drug effects*

Descurainiae Semen Lepidii Semen, also known as Tinglizi, originates from Brassicaceae plants Descurainia sophia or Lepidium apetalum. The former is commonly referred to as "Southern Tinglizi(Descurainiae Semen)", while the latter is known as "Northern Tinglizi(Lepidii Semen)". To scientifically and accurately identify the origin of Tinglizi medicinal materials and traditional Chinese medicine products, this study developed a specific DNA mini-barcode based on chloroplast genome sequences. By combining the DNA mini-barcode with DNA metabarcoding technology, a method for the qualitative and quantitative identification of Tinglizi medicinal materials and Chinese patent medicines was established. In this study, chloroplast genomes of Southern Tinglizi and Northern Tinglizi and seven commonly encountered counterfeit products were downloaded from the GenBank database. Suitable polymorphic regions were identified to differentiate these species, enabling the development of the DNA mini-barcode. Using DNA metabarcoding technology, medicinal material mixtures of Southern and Northern Tinglizi, as well as the most common counterfeit product, Capsella bursa-pastoris seeds, were analyzed to validate the qualitative and quantitative capabilities of the mini-barcode and determine its minimum detection limit. Additionally, the mini-barcode was applied to Chinese patent medicines containing Tinglizi to authenticate their botanical origin. The results showed that the developed mini-barcode(psbB) exhibited high accuracy and specificity, effectively distinguishing between the two authentic origins of Tinglizi and commonly encountered counterfeit products. The analysis of mixtures demonstrated that the mini-barcode had excellent qualitative and quantitative capabilities, accurately identifying the composition of Chinese medicinal materials in mixed samples with varying proportions. Furthermore, the analysis of Chinese patent medicines revealed the presence of the adulterant species(Capsella bursa-pastoris) in addition to the authentic species(Southern and Northern Tinglizi), indicating the occurrence of adulteration in commercially available Tinglizi-containing products. This study developed a method for the qualitative and quantitative identification of multi-origin Chinese medicinal materials and related products, providing a model for research on other multi-origin Chinese medicinal materials.
DNA Barcoding, Taxonomic/methods* ; Drugs, Chinese Herbal/chemistry* ; Drug Contamination ; Genome, Chloroplast ; Medicine, Chinese Traditional

The purpose of this study was to clarify the effect of Huotan Jiedu Tongluo Decoction on atherosclerosis(AS) injury in ApoE~(-/-) mice by regulating the ferroptosis pathway. Seventy-five ApoE~(-/-) mice were randomly divided into model group, low-, medium-, and high-dose of Huotan Jiedu Tongluo Decoction groups, and evolocumab group(n=15), and 15 C57BL/6J mice were selected as the blank group. Mice in the blank group were fed with a normal diet, and those in the other groups were fed with a high-fat diet to induce AS. From the 9th week, mice in Huotan Jiedu Tongluo Decoction groups were administrated with Huotan Jiedu Tongluo Decoction at corresponding doses by gavage, and those in the blank group and the model group were given an equal volume of distilled water. Mice in the evolocumab group were treated with evolocumab 18.2 mg·kg~(-1 )by subcutaneous injection every 2 weeks. After 8 weeks of continuous intervention, oil red O staining and hematoxylin-eosin(HE) staining were employed to observe the lipid deposition and plaque formation in the aortic root. Masson staining was used to evaluate the collagen content in the aortic root. The serum levels of total cholesterol(TC), triglycerides(TG), high-density lipoprotein cholesterol(HDL-C), and low-density lipoprotein cholesterol(LDL-C) were determined by biochemical kits. The levels of Fe~(2+), superoxide dismutase(SOD), malondialdehyde(MDA), and glutathione(GSH) in the aorta were measured by colorimetry. The protein and mRNA levels of nuclear factor erythroid 2-related factor 2(Nrf2), glutathione peroxidase 4(GPX4), solute carrier family 7 member 11(SLC7A11), and acyl-CoA synthetase long chain family member 4(ACSL4) in the aorta were detected by Western blot and RT-qPCR, respectively. The expression of Nrf2, GPX4, and SLC7A11 was localized by immunofluorescence. The results showed that low-, medium-, and high-dose Huotan Jiedu Tongluo Decoction reduced the plaque formation of aortic root and increased the collagen content in AS mice. At the same time, Huotan Jiedu Tongluo Decoction improved the lipid metabolism by lowering the levels of TC, LDL-C, and TG and elevating the level of HDL-C in the serum. Huotan Jiedu Tongluo Decoction enhanced the antioxidant capacity by elevating the levels of GSH and SOD and lowering the level of MDA in the aorta and inhibiting the accumulation of Fe~(2+) in the aorta. In addition, Huotan Jiedu Tongluo Decoction up-regulated the protein and mRNA levels of Nrf2, GPX4, and SLC7A11, while down-regulating the protein and mRNA levels of ACSL4. In summary, Huotan Jiedu Tongluo Decoction can effectively alleviate AS lesions in ApoE~(-/-) mice by activating the Nrf2/GPX4 pathway, reducing lipid peroxidation, and inhibiting ferroptosis.
Animals ; Ferroptosis/drug effects* ; Atherosclerosis/metabolism* ; Drugs, Chinese Herbal/administration & dosage* ; NF-E2-Related Factor 2/genetics* ; Mice ; Mice, Inbred C57BL ; Apolipoproteins E/metabolism* ; Male ; Phospholipid Hydroperoxide Glutathione Peroxidase/genetics* ; Signal Transduction/drug effects* ; Humans ; Mice, Knockout

Prone position ventilation (PPV) has been widely used in the treatment strategy of patients with acute respiratory distress syndrome (ARDS). Patients undergoing PPV may develop facial edema and are at risk for pressure injuries due to prolonged prone positioning. In clinical practice, preventive measures such as repositioning, protective dressings, and pressure-relief cushions are commonly used to prevent pressure injuries. However, factors such as improper endotracheal tube placement, self-paid dressings, and delayed clearance of oral and nasal secretions have reduced the effectiveness of preventing facial pressure injuries. To address the above issues, a device for preventing pressure injuries on the faces of patients in the prone position was designed by healthcare workers in the nursing department of Dalian Friendship Hospital, and a National Utility Model Patent of China was obtained (ZL 2024 2 0340439.8). The device consists of a support plate and a circuit control system. The support plate is equipped with two support members. Support member 1 is directly fixed to the support plate, while support member 2 is connected to the support plate via a slide and a spiral rod, serving to support the patient's face and allowing for adjustment of the appropriate width according to the size of the patient's face. Inside the two support members, there are several telescopic rods, with the upper ends designed as spherical supports. The height and position of the telescopic components can be adjusted through a circuit control system, regularly changing the pressure distribution on the patient's face, thereby achieving the purpose of changing the pressure points on the face. The inner wall of support member 2 is equipped with a camera, allowing direct observation of the patient's facial condition through a monitor, avoiding compression of the eyes and nose, and promptly removing secretions from the mouth to keep the face clean, thereby reducing the risk of facial pressure-related injuries. The center of the two support members features a hollow slot, facilitating the placement of a tracheal tube. The circuit control system includes a random module, a time setting module, a control module, and a drive module. Parameters can be set as needed. When the shortest set time is reached, the random module and time setting module send instructions to the control module. Upon receiving the instructions from the time setting module and the random number from the random module, the control module transmits information to the drive module. The drive module, upon receiving the information, controls multiple telescopic rods to adjust their height and position, thereby changing the support points on the patient's face. The device features a simple structure and convenient operation, allowing for flexible adaptation to the patient's facial shape. It can be replaced with the patient's facial pressure area, providing an intuitive view of the patient's facial pressure situation. With automation and high safety, it helps reduce the risk of pressure-related injuries and lightens the workload of medical staff.
Humans ; Pressure Ulcer/prevention & control* ; Prone Position ; Equipment Design ; Facial Injuries/prevention & control* ; Respiration, Artificial/instrumentation* ; Respiratory Distress Syndrome/therapy*