ABSTRACT
INTRODUCTION: Since the discovery of gonadotropin-inhibitory hormone (GnIH), it has been found to play a critical role in reproduction in vertebrates. Recently, a regulatory role of GnIH in appetite and energy metabolism has emerged, although its precise physiological mechanisms remain unknown. METHODS: Thus, the present study evaluated the effects of a single or long-term intraperitoneal GnIH treatment on the food intake, weight, and glucolipid metabolism of chickens, as well as investigating the possible neuroendocrinology factors and mechanisms involved in GnIH-induced obesity and glucolipid metabolism disorder. RESULTS: Our results show that the intraperitoneal administration of GnIH to chickens resulted in a marked body mass increase, hyperlipidemia, hyperglycemia, and glucose intolerance. Subsequently, the results of metabolomics studies and the pharmacological inhibition of the 5-HT2C receptor revealed that blocking the 5-HT2C receptor reinforced the effects of GnIH on food intake, body weight, and blood glucose and lipid levels, resulting in even worse cases of GnIH-induced hyperglycemia, hyperlipidemia, and hepatic lipid deposition. This suggests that, via the 5-HT2C receptor, peripheral 5-HT may act as a negative feedback regulator to interplay with GnIH and jointly control energy balance homeostasis in chickens. DISCUSSION: Our present study provides evidence of cross-talk between GnIH and 5-HT in food intake and energy metabolism at the in vivo pharmacological level, and it proposes a molecular basis for these interactions, suggesting that functional interactions between GnIH and 5-HT may open new avenues for understanding the mechanism of the neuroendocrine network involved in appetite and energy metabolism, as well as providing a new therapeutic strategy to prevent obesity, diabetes, and metabolic disorders.
Subject(s)
Chickens , Energy Metabolism , Feeding Behavior , Receptor, Serotonin, 5-HT2C , Serotonin , Animals , Energy Metabolism/drug effects , Receptor, Serotonin, 5-HT2C/metabolism , Serotonin/metabolism , Feeding Behavior/drug effects , Feeding Behavior/physiology , Hypothalamic Hormones/metabolism , Male , Blood Glucose/metabolism , Blood Glucose/drug effects , Hyperlipidemias/metabolism , Hyperlipidemias/chemically inducedABSTRACT
There is unclear evidence available on the associations between multiple metals and fasting blood glucose (FBG) in children, and whether they could be beneficial from physical activity. We included 283 children aged 4-12 years from two panel studies with 4-consecutive morning urinary 13 essential metals and 10 non-essential metals repeated across 3 seasons. We employed multiple informant model, linear mixed-effect model, and quantile g-computation to evaluate associations of single metal and their mixture with FBG and interactions with extra-school activity. The results showed that positive relations of multiple essential metals (aluminum, chromium, copper, iron, molybdenum (Mo), nickel, selenium (Se), strontium, zinc) and non-essential metals (arsenic (As), cadmium (Cd), rubidium, titanium (Ti), thallium) with FBG were the strongest at lag 0 (the health examination day), especially in overweight & obesity children (FDR <0.05). The strongest effect presented 1-fold increment in As was related to FBG increased 1.66% (95%CI: 0.84%, 2.48%) in overweight & obesity children. Notably, modification of extra-school activity showed significant, and the effects of multiple metals on FBG were attenuated in children taking total extra-school activity ≥1 h/day, and only one type of which, low or moderate & high intensity extra-school activity reached 20 min/day (Pint <0.05). For instance, each 1-fold increased As was associated with 1.41% increased FBG in overall children taking total extra-school activity <1 h/day, while that of 0.13% in those ≥1 h/day. Meanwhile, mixture of all, essential and non-essential metals were associated with increased FBG, a trend that decreased and became nonsignificant in children having certain extra-school activity, which were dominated by Mo, Se, Ti, Cd. And such relations were substantially beneficial from extra-school activity in overweight & obesity children. Accordingly, multiple essential and non-essential metals, both individual and in mixture, were positively related to FBG in children, which might be attenuated by regular physical activity.
Subject(s)
Blood Glucose , Exercise , Metals , Humans , Child , Child, Preschool , Female , Male , Blood Glucose/analysis , Metals/urine , Fasting , Environmental Pollutants/urineABSTRACT
Evidence on joint association of a phthalate mixture with thyroid function among children and its underlying mechanism is largely unknown. We aimed to explore the associations of 10 urinary phthalate metabolites (mPAEs), either as individuals or as a mixture, with thyroid function indicators [free thyroxine, free triiodothyronine (FT3), and thyroid-stimulating hormone (TSH)] in 144 children aged 4-12 years with up to 3 repeated visits across 3 seasons. Significant and positive associations were observed for mono-(2-ethylhexyl) phthalate (MEHP), mono-iso-butyl phthalate (MiBP), and mono-n-butyl phthalate (MnBP) with TSH, as well as monobenzyl phthalate (MBzP) with FT3 in dose-response manners. The relationship between MEHP and TSH remained robust in multiple-phthalate models. Bayesian kernel machine regression (BKMR) models revealed overall linear associations of the 10 mPAE mixture with higher TSH and FT3 levels, and MEHP and MBzP were major contributors. Meanwhile, MEHP, MiBP, and MnBP were linked to the elevation of multiple cytokines including CCL 27, CCL3, CXCL1, and IL-16. Among them, IL-16 mediated the relationships of MEHP and MiBP with TSH, and the mediated proportions were 24.16% and 24.27%, respectively. Our findings suggested that mPAEs dominated by MEHP were dose-responsively associated with elevated TSH among healthy children and mediated by IL-16.
Subject(s)
Environmental Pollutants , Phthalic Acids , Child , Humans , Environmental Exposure , Thyroid Gland/metabolism , Bayes Theorem , Interleukin-16 , Phthalic Acids/metabolism , ThyrotropinABSTRACT
The link between phthalates exposure and arterial stiffness in adults remains unclear. We aimed to investigate the associations of urinary phthalate metabolites with arterial stiffness in a longitudinal panel study involving 3 repeated visits among 127 Chinese adults. Urine samples were collected once a day for 4 consecutive days and 10 urinary phthalate metabolites were measured by gas chromatography-tandem mass spectrometry (GC-MS/MS). Brachial ankle pulse wave velocity (baPWV) and ankle-brachial index (ABI) were determined using an oscillometric device (BP-203RPEIII; Omron) in physical examinations during each visit. Linear mixed-effect (LME) models with the adaptive Least Absolute Shrinkage and Selection Operator (LASSO) method were applied to assess the associations between urinary phthalate metabolites and arterial stiffness parameters. The odds ratio (OR) for peripheral arterial disease (PAD) was estimated using generalized estimating equations. For ABI, mono-methyl phthalate (MMP) and mono-n-butyl phthalate (MBP) at lag 0 day were selected by the adaptive LASSO, whereas no phthalates were selected for baPWV. After adjusting for potential covariates and other metabolites, we found ABI reduction was associated with one-unit increase of ln-transformed urinary MBP at lag 0 day [ß = 0.013 (SE = 0.006), P = 0.003)]. Stratified analysis revealed that the inverse association was more evident in males (Pinteraction = 0.025). In addition, we observed a borderline risk of PAD in relation to MBP exposure at lag 0 day (P = 0.06). Our data suggested that environmental exposure to MBP may contribute to arterial stiffness, and the effect seems to be sex-specific.
Subject(s)
Phthalic Acids , Vascular Stiffness , Ankle Brachial Index , Female , Gas Chromatography-Mass Spectrometry , Humans , Male , Phthalic Acids/urine , Pulse Wave Analysis , Tandem Mass SpectrometryABSTRACT
Phthalates exposure has been reported to be linked with arterial stiffness. However, the biological mechanisms underlying this association remain unclear. We conducted a panel study using 338 paired urine-blood samples by repeated measurements of 123 adults across 3 seasons to assess the potential mediating role of plasma microRNAs (miRNAs) in the association of phthalates exposure with arterial stiffness. We measured 10 urinary phthalate metabolites by gas chromatography-tandem mass spectrometry (GC-MS/MS) and 5 candidate arterial stiffness-related miRNAs (miR-146a, miR-222, miR-125b, miR-126, and miR-21) in plasma by real-time PCR. Arterial stiffness parameters including brachial-ankle pulse wave velocity (baPWV) and ankle-brachial index (ABI) were determined in health examinations during each visit. Linear mixed-effect (LME) models revealed that mono-methyl phthalate (MMP), mono-iso-butyl phthalate (MiBP), mono-n-butyl phthalate (MBP), mono-n-octyl phthalate (MOP), and mono-(2-ethyl-5-carboxypentyl) phthalate (MECPP) were significantly associated with one or more of the 5 plasma miRNAs (all PFDR < 0.05). Based on weighted quantile sum (WQS) regression, we found positive associations of phthalate metabolites mixture with miR-146a, miR-125b, and miR-222, and individual MMP and MBP were the major contributors. Additionally, miR-146a was inversely related to ABI. Mediation analysis further indicated that miR-146a mediated 31.6% and 21.3% of the relationships of MMP and MiBP with ABI, respectively. Our findings suggested that certain phthalates exposure was related to plasma miRNAs alterations in a dose-response manner and miR-146a might partly mediate phthalate-associated ABI reduction.
Subject(s)
Environmental Pollutants , MicroRNAs , Phthalic Acids , Vascular Stiffness , Ankle Brachial Index , Environmental Exposure/analysis , Environmental Pollutants/analysis , Gas Chromatography-Mass Spectrometry , Phthalic Acids/analysis , Pulse Wave Analysis , Tandem Mass SpectrometryABSTRACT
Studies on associations of metals with leucocyte telomere length (LTL) were mainly limited to several most common toxic metals and single-metal effect, but the impact of other common metals and especially the overall joint associations and interactions of metal mixture with LTL are largely unknown. We included 15 plasma metals and LTL among 4906 participants from Dongfeng-Tongji cohort. Multivariable linear regression was used to estimate associations of individual metals with LTL. We also applied Bayesian kernel machine regression (BKMR) and quantile g-computation regression (Q-g) to evaluate the overall association and interactions, and identified the major contributors as well as the potential modifications by major characteristics. Multivariable linear regression found vanadium, copper, arsenic, aluminum and nickel were negatively associated with LTL, and a 2-fold change was related to 1.9%-5.1% shorter LTL; while manganese and zinc showed 3.7% and 4.0% longer LTL (all P < 0.05) in multiple-metal models. BKMR confirmed above metals and revealed a linearly inverse joint association between 15 metals and LTL. Q-g regression further indicated each quantile increase in mixture was associated with 5.2% shorter LTL (95% CI: -8.1%, -2.3%). Furthermore, manganese counteracted against aluminum and vanadium respectively (Pint<0.05). In addition, associations of vanadium, aluminum and metal mixture with LTL were more prominent in overweight participants. Our results are among the first to provide a new comprehensive view of metal mixture exposure on LTL attrition in the general population, including identifying the major components, metals interactions and the overall effects.
Subject(s)
Aluminum , Manganese , Aged , Bayes Theorem , China , Humans , Middle Aged , Telomere , Vanadium/toxicityABSTRACT
Using bacteriophages (phages) as environmental sanitizers has been recognized as a potential alternative method to remove bacterial contamination in vitro; however, very few studies are available on the application of phages for infection control in hospitals. Here, we performed a 3-year prospective intervention study using aerosolized phage cocktails as biocontrol agents against carbapenem-resistant Acinetobacter baumannii (CRAB) infection in the hospital. When a CRAB-infected patient was identified in an intensive care unit (ICU), their surrounding environment was chosen for phage aerosol decontamination. Before decontamination, 501 clinical specimens from the patients were subjected to antibiotic resistance analysis and phage typing. The optimal phage cocktails were a combination of different phage families or were constructed by next-evolutionary phage typing with the highest score for the host lysis zone to prevent the development of environmental CRAB phage resistance. The phage infection percentage of the antibiotic-resistant A. baumannii strains was 97.1%, whereas the infection percentage in the antibiotic-susceptible strains was 79.3%. During the phage decontamination periods from 2017 to 2019, the percentage of carbapenem-resistant A. baumannii in test ICUs decreased significantly from 65.3% to 55%. The rate of new acquisitions of CRAB infection over the three years was 4.4 per 1000 patient-days, which was significantly lower than that in the control wards (8.9 per 1000 patient-days) where phage decontamination had never been performed. In conclusion, our results support the potential of phage cocktails to decrease CRAB infection rates, and the aerosol generation process may make this approach more comprehensive and time-saving.
Subject(s)
Acinetobacter Infections , Acinetobacter baumannii , Bacteriophages , Cross Infection , Acinetobacter Infections/microbiology , Acinetobacter Infections/prevention & control , Aerosols , Anti-Bacterial Agents/pharmacology , Carbapenems/pharmacology , Cross Infection/microbiology , Cross Infection/prevention & control , Drug Resistance, Multiple, Bacterial , Humans , Microbial Sensitivity Tests , Prospective StudiesABSTRACT
Multicellular tumor spheroids and tumoroids are considered ideal in vitro models that reflect the features of the tumor microenvironment. Biomimetic components resembling the extracellular matrix form scaffolds to provide structure to 3-dimensional (3D) culture systems, supporting the growth of both spheroids and tumoroids. Although Matrigel has long been used to support 3D culture systems, batch variations, component complexity, and the use of components derived from tumors are complicating factors. To address these issues, we developed the ACD 3D culture system to provide better control and consistency. We evaluated spheroid and tumoroid formation using the ACD 3D culture system, including the assessment of cell viability and cancer marker expression. Under ACD 3D culture conditions, spheroids derived from cancer cell lines exhibited cancer stem cell characteristics, including a sphere-forming size and the expression of stem cell marker genes. The ACD 3D culture system was also able to support patient-derived primary cells and organoid cell cultures, displaying adequate cell growth, appropriate morphology, and resistance to oxaliplatin treatment. These spheroids could also be used for drug screening purposes. In conclusion, the ACD 3D culture system represents an efficient tool for basic cancer research and therapeutic development.
Subject(s)
Neoplasms , Spheroids, Cellular , Humans , Neoplasms/drug therapy , Neoplasms/metabolism , Cell Culture Techniques , Cell Line , Stem Cells/metabolism , Tumor MicroenvironmentABSTRACT
OBJECTIVE: Type 2 diabetes mellitus (T2DM) is known to be related to increased arterial stiffness. However, little is known about the risk of T2DM due to accelerated arterial stiffness and the underlying mechanism involved. We aimed to examine arterial stiffness, as determined by brachial-ankle pulse wave velocity (baPWV), in relation to T2DM among a community-based population and whether the association was mediated by white blood cell (WBC) counts. Approach and results: A total of 1036 Chinese adults aged 64.3 years with complete data were qualified in the present study. The dose-response association between baPWV levels, WBC counts, and risk of T2DM were explored using generalized linear models or multivariate logistic regression models. A mediation analysis was conducted to investigate the role of WBC counts on the association between baPWV and T2DM. After multivariate adjustments, we observed a dose-responsive relationship between increased baPWV and elevated risk of T2DM: comparing extreme tertiles of baPWV, the adjusted odds ratio for T2DM risk was 2.29 (95% CI, 1.32-3.98; P for trend =0.005). In addition, significant dose-dependent relationships were found across baPWV tertiles with increasing total or differential WBC counts, which in turn, were positively related to higher risk of T2DM (all P for trend <0.05). Mediation analyses indicated that total WBC count mediated 4.5% of the association between increased baPWV and elevated T2DM risk. CONCLUSIONS: Increased arterial stiffness might increase T2DM risk among middle-aged and older Chinese adults, which was partially mediated by total WBC count.
Subject(s)
Diabetes Mellitus, Type 2/physiopathology , Leukocyte Count , Vascular Stiffness , Ankle Brachial Index , Asian People , Diabetes Mellitus, Type 2/ethnology , Female , Humans , Male , Middle Aged , Obesity/physiopathology , Pulse Wave Analysis , Risk FactorsABSTRACT
Evidence available on the effects of size-fractionated particulate matters and their constituents on children's renal function is lack. We conducted a longitudinal panel study among 144 children aged 4-12 years with up to 3 repeated visits from 2018 to 2019. We estimated the effects of size-fractionated particle number counts (PNCs) and their 13 constituents on estimated glomerular filtration rate (eGFR) over different lag times with linear mixed-effects models and Bayesian kernel machine regression. We found the inverse dose-responsive associations of 3 sizes PNCs with eGFR were the strongest at lag 2 day. Compared to PNC0.5, PNC1 and PNC2.5 showed stronger and similar effects on eGFR reduction. On average, an interquartile range increase in PNC0.5, PNC1 and PNC2.5 were significantly associated with 1.70%, 2.82% and 2.76% decrease in eGFR, respectively. Girls were more susceptible to the toxicity of PNC1 and PNC2.5 exposure on eGFR. Several constituents including organic carbon (OC), Mg+, PO3- and HC2O4- in 3 sizes PNCs were robustly and consistently linked to eGFR reduction at lag 2 day. Moreover, the cumulative effects of different constituents on lower eGFR were significant, when they were all at or above a size-independent threshold (the 60th, 65th, and 70th percentiles in PNC0.5, PNC1 and PNC2.5 constituents, respectively), compared to their median value. And only OC displayed a significantly detrimental effect on eGFR when all the other constituents were fixed at 25th, 50th, and 75th percentiles. In summary, short-term exposure to PNCs were size-dependent related to reduced eGFR in dose-responsive manner among healthy children, and OC might play a more important role in PNC-induced nephrotoxicity than others.
Subject(s)
Air Pollutants/toxicity , Environmental Exposure , Kidney , Particulate Matter/toxicity , Air Pollutants/analysis , Bayes Theorem , Child , Child, Preschool , Female , Glomerular Filtration Rate/drug effects , Humans , Linear Models , Longitudinal Studies , Male , Particle Size , Particulate Matter/analysisABSTRACT
Drug detection and identification technology are of great significance in drug supervision and management. To determine the exact source of drugs, it is often necessary to directly identify multiple varieties of drugs produced by multiple manufacturers. Near-infrared spectroscopy (NIR) combined with chemometrics is generally used in these cases. However, existing NIR classification modeling methods have great limitations in dealing with a large number of categories and spectra, especially under the premise of insufficient samples, unbalanced samples, and sensitive identification error cost. Therefore, this paper proposes a NIR multi-classification modeling method based on a modified Bidirectional Generative Adversarial Networks (Bi-GAN). It makes full utilization of the powerful feature extraction ability and good sample generation quality of Bi-GAN and uses the generated samples with obvious features, an equal number between classes, and a sufficient number within classes to replace the unbalanced and insufficient real samples in the courses of spectral classification. 1721 samples of four kinds of drugs produced by 29 manufacturers were used as experimental materials, and the results demonstrate that this method is superior to other comparative methods in drug NIR classification scenarios, and the optimal accuracy rate is even more than 99% under ideal conditions.
Subject(s)
Pharmaceutical Preparations , Spectroscopy, Near-InfraredABSTRACT
Multispectral pedestrian detection, which consists of a color stream and thermal stream, is essential under conditions of insufficient illumination because the fusion of the two streams can provide complementary information for detecting pedestrians based on deep convolutional neural networks (CNNs). In this paper, we introduced and adapted a simple and efficient one-stage YOLOv4 to replace the current state-of-the-art two-stage fast-RCNN for multispectral pedestrian detection and to directly predict bounding boxes with confidence scores. To further improve the detection performance, we analyzed the existing multispectral fusion methods and proposed a novel multispectral channel feature fusion (MCFF) module for integrating the features from the color and thermal streams according to the illumination conditions. Moreover, several fusion architectures, such as Early Fusion, Halfway Fusion, Late Fusion, and Direct Fusion, were carefully designed based on the MCFF to transfer the feature information from the bottom to the top at different stages. Finally, the experimental results on the KAIST and Utokyo pedestrian benchmarks showed that Halfway Fusion was used to obtain the best performance of all architectures and the MCFF could adapt fused features in the two modalities. The log-average miss rate (MR) for the two modalities with reasonable settings were 4.91% and 23.14%, respectively.
Subject(s)
Pedestrians , Attention , Benchmarking , Humans , Lighting , Neural Networks, ComputerABSTRACT
The gender differences of immunity have been elucidated in many vertebrates and invertebrates. However, the information of this difference was still not clear in the sea cucumber Apostichopus japonicus, which is one of the most valuable aquaculture species and susceptible to diseases caused by pathogen infection. In the present study, the transcriptome of coelomocytes from female and male A. japonicus before and after spawning was obtained by RNA-sequencing technology. A total of 4,538 and 8,248 differentially expressed genes were identified between female and male A. japonicus before and after spawning, respectively, indicating that the gender differences of gene expression profiles in A. japonicus were more remarkable after spawning. Further KEGG enrichment analyses were conducted for both male and female up-regulated genes before and after spawning. The results revealed that the capacity to kill pathogens in female A. japonicus might be more powerful than that in males no matter before and after spawning; the antioxidant ability in male A. japonicus was probably stronger than that in females after spawning; the complement system in male A. japonicus might be more effective than that in females after spawning; and the apoptosis was likely to be more serious in male A. japonicus before spawning. Moreover, we speculated that the fatty acid composition might be one of the inducements for gender specific immune differences of A. japonicus. Overall, the results of our study illustrated the global gender specific immune differences of A. japonicus and contributed to understanding of the molecular mechanisms underlying sea cucumber immune regulation.
Subject(s)
Immunity, Innate/genetics , Immunocompetence/physiology , Stichopus/immunology , Transcriptome , Animals , Female , Male , Reproduction/immunology , Sequence Analysis, RNA/veterinary , Sex Characteristics , Sex Factors , Sexual Behavior, Animal , Stichopus/genetics , Stichopus/physiologyABSTRACT
The combination of near infrared spectrum and pattern recognition methods has a wide application prospect in rapid and nondestructive supervision and management of drugs. The traditional identification methods regard the smallest error rate as the goal while the imbalance of classes is ignored. This makes the positive class is overwhelming covered by the negative class and reduces its effect for the classifier, so that the classification results tend to recognize the negative class correctly, which severely affects the identification accuracy. In this paper, we mainly studied the class imbalance problems of true or false drugs via infrared spectral data of its, and then propose a balance cascading and sparse representation based classification method (BC-SRC) by combining the Balance Cascading with SRC. We sampling majority samples from the majority class for several times, which has the same size as minority samples and the majority samples we sampled can contain all the majority class samples entirely (sampling times is ceiling the result of majority samples number divide minority samples number). We can get sets of results, and then obtain the final predict labels form those results. Experiments of three databases achieved on Matlab2012a shows that the method is effective. From the experimental results, it can be seen that the method is superior to the commonly used Partial Least Squares (PLS), Extreme Learning Machine (ELM) and BP. Particularly, for the imbalanced databases, when the imbalance factor is greater than 10, the proposed method has more stable performance with higher classification accuracy than the existing ones mentioned above.
ABSTRACT
Near-infrared(NIR)As a fast and non-destructive testing technology, spectroscopy techniques is very suitable for pharmaceutical discrimination. Autoencoder network, as a hot research topic, has drawn widespread attention in machine learning research in recent years. Compared with traditional surface learning algorithm models, Autoencoder network has more powerful modeling capability as a typical deep networks model. Based on the unsupervised greedy layer-wise pre-training, autoencoder trains the network layer by layer while minimizing the error in reconstructing. Each layer is pre-trained with an unsupervised learning algorithm, learning a nonlinear transformation of the input of each layer which is the output of the previous layer. Pre-whitening process could get the inner structural features of the data more effectively. The supervised fine-tuning is followed with the unsupervised pre-training which sets the stage for a final training phase. The deep architecture is fine-tuned with respect to a supervised training criterion with gradient-based optimization. In this paper, firstly, the preprocessing step and pre-whitening transformation were used to treat near-infrared spectroscopy data of erythromycin ethylsuccinate, The pre-whitening transformation would reduce the correlation of the features, which gave each feature the same variance. Experimental results showed that the pre-whitening process had improved the classification accuracy of Sparse Denoising Autoencoder (SDAE) effectively. The SDAE with two hidden layers combined with pre-whitening was used to build the classification model for the identification of counterfeit pharmaceutical. The BP neural networks was compared with SVM algorithm for the classification accuracy and mean absolute difference (MAD). SDAE algorithm had higher classification accuracy than BP neural networks which had the same network structure with the SDAE networks, and SDAE algorithm also performed better than the SVM algorithm when the train datasets achieved a certain amount. As to the generalization performances, SDAE algorithm had less mean absolute difference of classification accuracy than SVM and BP Neural Networks. This result showed that SDAE algorithm could be effectively used to discriminate the counterfeit pharmaceutical.
Subject(s)
Algorithms , Neural Networks, Computer , Machine Learning , Spectroscopy, Near-InfraredABSTRACT
As an effective technique to identify counterfeit drugs, Near Infrared Spectroscopy has been successfully used in the drug management of grass-roots units, with classifier modeling of Pattern Recognition. Due to a major disadvantage of the characteristic overlap and complexity, the wide bandwidth and the weak absorption of the Spectroscopy signals, it seems difficult to give a satisfactory solutions for the modeling problem. To address those problems, in the present paper, a summation wavelet extreme learning machine algorithm (SWELM(CS)) combined with Cuckoo research was adopted for drug discrimination by NIRS. Specifically, Extreme Learning Machine (ELM) was selected as the classifier model because of its properties of fast learning and insensitivity, to improve the accuracy and generalization performances of the classifier model; An inverse hyperbolic sine and a Morlet-wavelet are used as dual activation functions to improve convergence speed, and a combination of activation functions makes the network more adequate to deal with dynamic systems; Due to ELM' s weights and hidden layer threshold generated randomly, it leads to network instability, so Cuckoo Search was adapted to optimize model parameters; SWELM(CS) improves stability of the classifier model. Besides, SWELM(CS) is based on the ELM algorithm for fast learning and insensitivity; the dual activation functions and proper choice of activation functions enhances the capability of the network to face low and high frequency signals simultaneously; it has high stability of classification by Cuckoo Research. This compact structure of the dual activation functions constitutes a kernel framework by extracting signal features and signal simultaneously, which can be generalized to other machine learning fields to obtain a good accuracy and generalization performances. Drug samples of near in- frared spectroscopy produced by Xian-Janssen Pharmaceutical Ltd were adopted as the main objects in this paper. Experiments for binary classification and multi-label classification were conducted, and the conclusion proved that the proposed method has more stable performance, higher classification accuracy and lower sensitivity to training samples than the existing ones, such as the BP neural network, ELM and-ELM by particle swarm optimization.
Subject(s)
Artificial Intelligence , Pharmaceutical Preparations/analysis , Spectroscopy, Near-Infrared , Algorithms , Models, Theoretical , Neural Networks, Computer , Software , Wavelet AnalysisABSTRACT
Semi-supervised medical image segmentation (SSMIS) has witnessed substantial advancements by leveraging limited labeled data and abundant unlabeled data. Nevertheless, existing state-of-the-art (SOTA) methods encounter challenges in accurately predicting labels for the unlabeled data, giving rise to disruptive noise during training and susceptibility to erroneous information overfitting. Moreover, applying perturbations to inaccurate predictions further impedes consistent learning. To address these concerns, we propose a novel cross-head mutual mean-teaching network (CMMT-Net) incorporated weak-strong data augmentations, thereby benefiting both co-training and consistency learning. More concretely, our CMMT-Net extends the cross-head co-training paradigm by introducing two auxiliary mean teacher models, which yield more accurate predictions and provide supplementary supervision. The predictions derived from weakly augmented samples generated by one mean teacher are leveraged to guide the training of another student with strongly augmented samples. Furthermore, two distinct yet synergistic data perturbations at the pixel and region levels are introduced. We propose mutual virtual adversarial training (MVAT) to smooth the decision boundary and enhance feature representations, and a cross-set CutMix strategy to generate more diverse training samples for capturing inherent structural data information. Notably, CMMT-Net simultaneously implements data, feature, and network perturbations, amplifying model diversity and generalization performance. Experimental results on three publicly available datasets indicate that our approach yields remarkable improvements over previous SOTA methods across various semi-supervised scenarios. The code is available at https://github.com/Leesoon1984/CMMT-Net.
Subject(s)
Supervised Machine Learning , Humans , Image Processing, Computer-Assisted/methods , Algorithms , Neural Networks, ComputerABSTRACT
Stroke is a leading cause of disability and fatality in the world, with ischemic stroke being the most common type. Digital Subtraction Angiography images, the gold standard in the operation process, can accurately show the contours and blood flow of cerebral vessels. The segmentation of cerebral vessels in DSA images can effectively help physicians assess the lesions. However, due to the disturbances in imaging parameters and changes in imaging scale, accurate cerebral vessel segmentation in DSA images is still a challenging task. In this paper, we propose a novel Edge Regularization Network (ERNet) to segment cerebral vessels in DSA images. Specifically, ERNet employs the erosion and dilation processes on the original binary vessel annotation to generate pseudo-ground truths of False Negative and False Positive, which serve as constraints to refine the coarse predictions based on their mapping relationship with the original vessels. In addition, we exploit a Hybrid Fusion Module based on convolution and transformers to extract local features and build long-range dependencies. Moreover, to support and advance the open research in the field of ischemic stroke, we introduce FPDSA, the first pixel-level semantic segmentation dataset for cerebral vessels. Extensive experiments on FPDSA illustrate the leading performance of our ERNet.