Electroencephalography (EEG) is a non-invasive neurophysiological monitoring technique. It records the electrical activity of the cerebral cortex using electrodes placed on the scalp surface. Owing to its high safety, portability, and millisecond-level temporal resolution, EEG has been widely utilized in a variety of fields, including clinical diagnosis, brain-computer interfaces (BCIs), and cognitive neuroscience research. However, due to its microvolt-level amplitude, EEG is highly susceptible to various artifacts, including electrooculographic (EOG), electrocardiographic (ECG), electromyographic (EMG), and power line interference (PLI). These artifacts can obscure genuine neural activity and introduce spurious electrophysiological features. Consequently, they may compromise EEG signal quality, thereby reducing the reliability of downstream analyses. To address this issue, numerous EEG artifact removal methods have been developed, including both traditional denoising techniques and deep learning-based approaches. Traditional EEG denoising methods have long served as the primary solutions for artifact removal. Representative approaches include filtering, regression, and blind source separation. Although these methods have demonstrated effectiveness in specific scenarios, they suffer from several inherent limitations. Filtering assumes that artifacts and EEG signals can be separated in the frequency domain, but many artifacts, such as EOG and EMG, overlap with EEG spectra, which may lead to the loss of valuable neural information. Regression methods require high-quality artifact references to estimate and subtract contaminations, limiting their effectiveness in reference-free scenarios. Blind source separation can remove artifacts without external references, but it typically requires the number of EEG channels to exceed the number of sources, restricting its application in single- or low-channel EEG recordings. Deep learning-based EEG denoising methods address these limitations effectively. First, they learn the nonlinear mapping between contaminated and clean EEG directly from data in an end-to-end manner. This approach does not rely on assumptions about spectral separability, thereby preserving neural activity more completely. Second, the reference information is incorporated during the training phase, allowing the trained model to perform artifact removal independently without external references. Third, deep learning models can be flexibly designed to accommodate various recording setups, achieving robust denoising for both high-density and single-channel EEG. Collectively, these advantages enable deep learning-based methods to overcome the main challenges of traditional approaches, providing more accurate and reliable EEG signal recovery. The superior denoising performance of deep learning-based EEG denoising methods has attracted increasing attention in EEG artifact removal research. As a result, many deep learning-based denoising methods have been developed and successfully applied in neural engineering areas. However, a systematic review of the techniques and applications in this field is still lacking. To address this gap, this paper reviews recent advances in deep learning-based EEG denoising from four perspectives: technical principle, benchmark dataset, denoising model, and evaluation method. Representative applications in neural signal analysis and BCI decoding are also summarized. Furthermore, the advantage, existing challenge, and future research direction of deep learning-based EEG denoising are discussed. This review aims to provide valuable theoretical insights and technical guidance for researchers. It is also expected to promote further advances and broader applications of deep learning-based EEG denoising techniques.

OBJECTIVE: To identify the key features of facial and tongue images associated with anemia in female populations, establish anemia risk-screening models, and evaluate their performance. METHODS: A total of 533 female participants (anemic and healthy) were recruited from Shuguang Hospital. Facial and tongue images were collected using the TFDA-1 tongue and face diagnosis instrument. Color and texture features from various parts of facial and tongue images were extracted using Face Diagnosis Analysis System (FDAS) and Tongue Diagnosis Analysis System version 2.0 (TDAS v2.0). Least Absolute Shrinkage and Selection Operator (LASSO) regression was used for feature selection. Ten machine learning models and one deep learning model (ResNet50V2 + Conv1D) were developed and evaluated. RESULTS: Anemic women showed lower a-values, higher L- and b-values across all age groups. Texture features analysis showed that women aged 30-39 with anemia had higher angular second moment (ASM)and lower entropy (ENT) values in facial images, while those aged 40-49 had lower contrast (CON), ENT, and MEAN values in tongue images but higher ASM. Anemic women exhibited age-related trends similar to healthy women, with decreasing L-values and increasing a-, b-, and ASM-values. LASSO identified 19 key features from 62. Among classifiers, the Artificial Neural Network (ANN) model achieved the best performance [area under the curve (AUC): 0.849, accuracy: 0.781]. The ResNet50V2 model achieved comparable results [AUC: 0.846, accuracy: 0.818]. CONCLUSION Differences in facial and tongue images suggest that color and texture features can serve as potential TCM phenotype and auxiliary diagnostic indicators for female anemia.
Humans ; Female ; Tongue/diagnostic imaging* ; Adult ; Anemia/diagnosis* ; Middle Aged ; Face/diagnostic imaging* ; Young Adult ; Machine Learning

Risks of food safety induced by small molecule drug residues in animal food and environment have become an increasing public concern,so it is necessary to develop highly sensitive and easy-to-operate techniques to detect small molecules.Herein,a metasurface plasma resonance(MetaSPR)sensor chip coupled with gold nanoparticles(AuNPs)was developed for detection of enrofloxacin(ENR)based on competitive immunoassay.The detection range of the sensor for ENR was 0.025-3.2 ng/mL,and the detection limit(3σ)was 20 pg/mL.The biosensor showed excellent performance including high selectivity,good stability,ease to operate and high throughput,etc.The developed method was applied to detection of ENR residues in real samples,with recoveies of 96.0% -105.0%.The proposed sensing strategy provided new technique reference for detection of other small molecules in the field of residue analysis in food safety and environment monitoring.

Objective To verify the efficacy of a domestic human DNA quantification kit based on real-time fluorescence quantitative PCR in detecting the total human DNA concentration,male DNA concen-tration in mixed male/female DNA samples,the degree of DNA degradation and inhibitor tolerance.Methods Samples with different concentrations,different male/female ratios,different concentrations of inhibitors,and different degradation degrees were tested using the domestic human DNA quantification kit based on real-time fluorescence quantitative PCR.This kit was compared with a similar product on the market and was applied to the detection of DNA from real cases.Results This human DNA quan-tification kit can effectively detect human DNA as low as 0.001 65 ng/μL,and 6.25 pg/μL of male DNA in mixed samples with a male-to-female ratio of 1∶15 000.Even when the sample contains as high as 400 ng/μL of humic acid or 1 000 μmol/L of hemin alone,the DNA concentration can still be accurately detected.The degradation index can effectively characterize the degradation degree of the sample.This kit has been successfully applied in forensic practice.Conclusion This human DNA quan-tification kit is accurate and reliable in detection.It can accurately reflect the degradation of DNA and inhibitor tolerance.It has good performance in quantitative accuracy,determination of the male/female ratio in mixed samples,and inhibitor tolerance.It has application potential in forensic case examination.

Icaritin(ICT)is an 8-isopentenylflavonoid,which is the main effective component of the tra-ditional Chinese medicine Epimedium.Previously,we found that Icaritin inhibits the growth of glioblasto-ma(GBM)cells.Herein we aim to study the in vivo anti-GBM effectiveness of Icaritin and explore its mechanism.The results of MTT assay,flow cytometry,comet assay and cellular immunofluorescence as-say in vitro showed that ICT inhibited the proliferation of four kinds of GBM cells,U87,U251,U118 and A172,induced early apoptosis(P＜0.001)and late apoptosis(P＜0.05)in U87 cells,induced DNA damage in U87 cells,and blocked the growth of U87 cells at the G0/G1 phase(P＜0.0001)in a concen-tration-time-dependent manner.In vivo subcutaneous tumor transplantation tumor experiments showed that feeding 200 mg/kg(P＜0.01)and 400 mg/kg(P＜0.001)ICT had a significant inhibitory effect on the growth of GBM subcutaneous tumors,and had no significant toxic effects on heart,liver,spleen,lung and kidney tissues.The results of network pharmacological analysis,molecular docking and cellular thermodynamic experiments showed that there were 26 possible target proteins between ICT and GBM,a-mong which the expression of p53 in GBM tissues was significantly(P＜0.001)higher than in normal tis-sues,and the binding energy of ICT and p53 was lower;cellular thermodynamic experiments verified that ICT significantly enriched the level of p53 in the living cells of GBM,which indicated that ICT could tar-get p53.The expression of key proteins in the DNA damage repair and apoptosis-associated FOXO signa-ling pathway was detected by ICT.The results showed that the expression of ATR(P＜0.01),P53(P＜0.001),P21(P＜0.05)and γ-H2AX(P＜0.05)was up-regulated,whereas the expression of Cyc-lin E1(P＜0.01),E2F1(P＜0.05),CDK2(P＜0.01),Rb(P＜0.001),p-Rb(P＜0.0001)and WRN(P＜0.0001)expression were down-regulated.There was no significant change in the expres-sion of FOXO 1 in the FOXO pathway or a significant down-regulation of its phosphorylation level.This study demonstrated that ICT could effectively inhibit the growth of GBM cells in vivo.It targets p53 to regulate the DNA damage repair pathway and FOXO signaling pathway to induce GBM cell cycle arrest and apoptosis.

Objective:To explore the interrelationship between brain functional networks and features in functional magnetic resonance imaging(fMRI)of patients with Alzheimer's disease(AD),and to construct mixed-function networks(MFN),and apply them in machine learning classification models to improve the accuracy of AD classification.Methods:102 AD patients and 227 healthy subjects in the Alzheimer's Neuroimaging Initiative(ADNI)dataset were retrospectively analyzed.The partial correlation brain network of the blood oxygen level dependent(BOLD)signal was calculated and fused with low-frequency wave amplitude(ALFF),fractional low-frequency wave amplitude(fALFF)and local consistency(ReHo)features to construct MFN.Network topology parameters were extracted,and a variety of machine learning classification models were constructed based on MFN topological parameters,accuracy,precision,recall and area under the curve(AUC)were used to evaluate the predictive efficiency of the models.Results:By constructed MFN and calculated intra group to inter group ratio(IIGR),35 features could be obtained from ALFF,fALFF and ReHo feature topological parameter analysis,after rank sum test and FDR correction,there were statistical differences among 28 features(P＜0.05).The classification results show that,all the five classifiers have high classification performance on the test data set.The accuracy,precision and recall rates of random forest(RF),adaptive lifting algorithm(AdaBoost),guided aggregation algorithm(Bagging)and support vector machine(SVM)were all 99.7％,and the AUC values were up to 100％,99.5％,99.1％and 99.5％,respectively.The accuracy(98.5％),precision(98.5％),recall(98.5％),and AUC(99.1％)of the multi-layer perceptron(MLP)were slightly lower than other models,but remained excellent.It was worth noting that RF has the highest AUC value of all models at 100.0％,while Bagging has the lowest AUC value(99.1％)in the integrated approach.The results of performance comparison show that,MFN classification model can significantly improve the recognition and classification of AD disease,and greatly improve the performance of various indicators of the classifier.The results showed that,MFN classification model was superior to intelligent classification based fusion,DBN-based multitask learning,PVT-TSVM,unsupervised learning and clustering,SVM and SVM of degree 3 polynomial kernel function in key indicators such as accuracy(99.13％),AUC(99.42％),recall rate(99.46％)and specificity(99.42％)with plasma proteins,machine learning algorithms.It was further proved that MFN classification model has good generalization ability and robustness in AD disease classification.Conclusion:The AD classification model constructed based on brain mixed function network topology parameters and machine learning can improve the accuracy of AD classification.

Aim To establish and evaluate an alternative meth-od for detecting skin sensitization of cosmetics based on local lymph node assay using bromodeoxyuridine(BrdU)with flow cytometry(FCM).Methods(1)25％hexyl cinnamic alde-hyde(HCA)was chosen as a positive control with an acetone:olive oil(4∶1,V/V,AOO)mixture as a vehicle control for the experiment.The dorsal sides of both ears of mice were treated with test solutions on day 1,day 2,and day 3.Brdu solution was injected inter-peritoneally on day 5.On day 6,the bilateral ears and mandibular lymph nodes were excised,and the number of Brdu positive cells was measured by flow cytometry.The stim-ulation index(SI)was calculated to identify whether it was ≥3,in order to establish the method of LLNA:Brdu-FCM.(2)BrdU-FCM test was conducted using a blind method with the fif-teen reference substances listed in OECD TG429 whose skin sensitization potentials were known.The test substances were dissolved in AOO,N,N-dimethylformamide(DMF)or dimeth-yl sulfoxide(DMSO)at three different concentrations.Tests were performed the same as above.SI and EC2.7 were calculat-ed to evaluate whether the test substance was categorized as a skin sensitizer.The reliability and accuracy of the method were validated by comparing the classification of test substances with that in OECD TG429.Results The SI for 25％HCA was 3.9,showing positive in the skin sensitization test.It demonstrated that the LLNA:Brdu-FCM test method was properly implemen-ted.Nine test substances(2,4-dinitrochlorobenzene,4-pheny-lenediamine,cobalt chloride,2-mercaptobenzothiazole,hexyl-cinnamaldehyde,eugenol,phenyl benzoate,cinnamic alcohol,imidazolidinyl urea)were positive,and six test substances(methyl methacrylate,chlorobenzene,isopropanol,lactic acid,methyl salicylate,salicylic acid)were negative.The method was evaluated with sensitivity of 90％,specificity of 100％,positive prediction rate of 100％,negative prediction rate of 83％,false positive rate of 0％,false negative rate of 17％and accuracy of 93％.The LLNA:BrdU-FCM assay could correctly categorize the test substances that were skin sensitizers or non-sensitizers.Conclusion The LLNA:BrdU-FCM assay appears to be a relia-ble predictor of skin sensitization protential of chemicals,and it is expected to an alternative method for identifying skin sensitization as a supplementary in safety evaluation of cosmetic ingredient.

Objective To investigate the distribution and antimicrobial resistance profiles of common pathogens isolated from cerebrospinal fluid(CSF)in CHINET program from 2015 to 2021.Methods The bacterial strains isolated from CSF were identified in accordance with clinical microbiology practice standards.Antimicrobial susceptibility test was conducted using Kirby-Bauer method and automated systems per the unified CHINET protocol.Results A total of 14 014 bacterial strains were isolated from CSF samples from 2015 to 2021,including the strains isolated from inpatients(95.3％)and from outpatient and emergency care patients(4.7％).Overall,19.6％of the isolates were from children and 80.4％were from adults.Gram-positive and Gram-negative bacteria accounted for 68.0％and 32.0％,respectively.Coagulase negative Staphylococcus accounted for 73.0％of the total Gram-positive bacterial isolates.The prevalence of MRSA was 38.2％in children and 45.6％in adults.The prevalence of MRCNS was 67.6％in adults and 69.5％in children.A small number of vancomycin-resistant Enterococcus faecium(2.2％)and linezolid-resistant Enterococcus faecalis(3.1％)were isolated from adult patients.The resistance rates of Escherichia coli and Klebsiella pneumoniae to ceftriaxone were 52.2％and 76.4％in children,70.5％and 63.5％in adults.The prevalence of carbapenem-resistant E.coli and K.pneumoniae(CRKP)was 1.3％and 47.7％in children,6.4％and 47.9％in adults.The prevalence of carbapenem-resistant Acinetobacter baumannii(CRAB)and Pseudomonas aeruginosa(CRPA)was 74.0％and 37.1％in children,81.7％and 39.9％in adults.Conclusions The data derived from antimicrobial resistance surveillance are crucial for clinicians to make evidence-based decisions regarding antibiotic therapy.Attention should be paid to the Gram-negative bacteria,especially CRKP and CRAB in central nervous system(CNS)infections.Ongoing antimicrobial resistance surveillance is helpful for optimizing antibiotic use in CNS infections.

Objective To investigate the antimicrobial resistance of clinical isolates from elderly patients(≥65 years)in major medical institutions across China.Methods Bacterial strains were isolated from elderly patients in 52 hospitals participating in the CHINET Antimicrobial Resistance Surveillance Program during the period from 2015 to 2021.Antimicrobial susceptibility test was carried out by disk diffusion method and automated systems according to the same CHINET protocol.The data were interpreted in accordance with the breakpoints recommended by the Clinical and Laboratory Standards Institute(CLSI)in 2021.Results A total of 514 715 nonduplicate clinical isolates were collected from elderly patients in 52 hospitals from January 1,2015 to December 31,2021.The number of isolates accounted for 34.3％of the total number of clinical isolates from all patients.Overall,21.8％of the 514 715 strains were gram-positive bacteria,and 78.2％were gram-negative bacteria.Majority(90.9％)of the strains were isolated from inpatients.About 42.9％of the strains were isolated from respiratory specimens,and 22.9％were isolated from urine.More than half(60.7％)of the strains were isolated from male patients,and 39.3％isolated from females.About 51.1％of the strains were isolated from patients aged 65-＜75 years.The prevalence of methicillin-resistant strains(MRSA)was 38.8％in 32 190 strains of Staphylococcus aureus.No vancomycin-or linezolid-resistant strains were found.The resistance rate of E.faecalis to most antibiotics was significantly lower than that of Enterococcus faecium,but a few vancomycin-resistant strains(0.2％,1.5％)and linezolid-resistant strains(3.4％,0.3％)were found in E.faecalis and E.faecium.The prevalence of penicillin-susceptible S.pneumoniae(PSSP),penicillin-intermediate S.pneumoniae(PISP),and penicillin-resistant S.pneumoniae(PRSP)was 94.3％,4.0％,and 1.7％in nonmeningitis S.pneumoniae isolates.The resistance rates of Klebsiella spp.(Klebsiella pneumoniae 93.2％)to imipenem and meropenem were 20.9％and 22.3％,respectively.Other Enterobacterales species were highly sensitive to carbapenem antibiotics.Only 1.7％-7.8％of other Enterobacterales strains were resistant to carbapenems.The resistance rates of Acinetobacter spp.(Acinetobacter baumannii 90.6％)to imipenem and meropenem were 68.4％and 70.6％respectively,while 28.5％and 24.3％of P.aeruginosa strains were resistant to imipenem and meropenem,respectively.Conclusions The number of clinical isolates from elderly patients is increasing year by year,especially in the 65-＜75 age group.Respiratory tract isolates were more prevalent in male elderly patients,and urinary tract isolates were more prevalent in female elderly patients.Klebsiella isolates were increasingly resistant to multiple antimicrobial agents,especially carbapenems.Antimicrobial resistance surveillance is helpful for accurate empirical antimicrobial therapy in elderly patients.

Objective To examine the changing prevalence and antimicrobial resistance profiles of Burkholderia cepacia in 52 hospitals across China from 2015 to 2021.Methods A total of 9 261 strains of B.cepacia were collected from 52 hospitals between January 1,2015 and December 31,2021.Antimicrobial susceptibility of the strains was tested using Kirby-Bauer method or automated antimicrobial susceptibility testing systems according to a unified protocol.The results were interpreted according to the breakpoints released in the Clinical & Laboratory Standards Institute(CLSI)guidelines(2023 edition).Results A total of 9 261 strains of B.cepacia were isolated from all age groups,especially elderly patients.The proportion was 11.1％(1 032 strains)in children,significantly lower than the proportion in adults.About half(46.5％,4 310/9 261)of the strains were isolated from patients at least 60 years old and 42.3％(3 919/9 261)of the strains were isolated from young adults.Most isolates(71.1％)were isolated from sputum and respiratory secretions,followed by urine(10.7％)and blood samples(8.1％).B.cepacia isolates were highly susceptible to the five antimicrobial agents recommended in the CLSI M100 document(33rd edition,2023).B.cepacia isolates showed relatively higher resistance rates to meropenem and levofloxacin.However,the resistance rates to ceftazidime,trimethoprim-sulfamethoxazole,and minocycline remained below 8.1％.The percentage of B.cepacia strains resistant to levofloxacin was the highest compared to other antibiotics in any of the three age groups(from 12.4％in the patients＜18 years old to 20.6％in the patients aged 60 years or older).Conclusions B.cepacia is one of the clinically important non-fermenting gram-negative bacteria.Accurate and timely reporting of antimicrobial susceptibility test results and ongoing antimicrobial resistance surveillance are helpful for rational prescription of antimicrobial agents and proper prevention and control of nosocomial infections.