Search

Abstract. Background:. Risk assessment and treatment stratification for three-vessel coronary disease (TVD) remain challenging. This study aimed to investigate the prognostic value of left atrial volume index (LAVI) with the Synergy Between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery (SYNTAX) score II, and its association with the long-term prognosis after three strategies (percutaneous coronary intervention [PCI], coronary artery bypass grafting [CABG], and medical therapy [MT]) in patients with TVD. Methods:. This study was a post hoc analysis of a large, prospective cohort of patients with TVD in China, that aimed to determine the long-term outcomes after PCI, CABG, or optimal MT alone. A total of 8943 patients with TVD were consecutively enrolled between 2004 and 2011 at Fuwai Hospital. A total of 7818 patients with available baseline LAVI data were included in the study. Baseline, procedural, and follow-up data were collected. The primary endpoint was major adverse cardiac and cerebrovascular events (MACCE), which was a composite of all-cause death, myocardial infarction (MI), and stroke. Secondary endpoints included all-cause death, cardiac death, MI, revascularization, and stroke. Long-term outcomes were evaluated among LAVI quartile groups. Results:. During a median follow-up of 6.6 years, a higher LAVI was strongly associated with increased risk of MACCE (Q3: hazard ratio [HR] 1.20, 95% confidence interval [CI] 1.06-1.37, P = 0.005; Q4: HR 1.85, 95%CI 1.64-2.09, P

Objective: To investigate the potential efficacy and safety of Lutai Danshen Baishao granules (LDBG) for treating female melasma associated with kidney deficiency and blood stasis patterns. Methods: A randomized, double-blind, placebo-controlled trial was conducted at the Third Central Hospital of Tianjin, China from March to December 2023. A total of 110 female patients with melasma linked to kidney deficiency and blood stasis were enrolled and treated with either LDBG or a placebo twice daily for 60 days. Efficacy was assessed through measures such as the total melasma area, reduced melasma area, reduction rate of melasma area, melasma color score, Melasma Area and Severity Index (MASI) score, and traditional Chinese medicine (TCM) symptom score scale. Safety assessments included routine blood and biochemical tests. Results: Participants in both groups were aged 52–63 years, with no significant differences. After the 2-month intervention, the total melasma area decreased in both groups; however, a greater reduction was observed in the test group [462.50 mm2 (12.81%) vs. 100.00 mm2 (3.11%), P

Abstract Facial expression is among the most natural methods for human beings to convey their emotional information in daily life. Although the neural mechanisms of facial expression have been extensively studied employing lab-controlled images and a small number of lab-controlled video stimuli, how the human brain processes natural dynamic facial expression videos still needs to be investigated. To our knowledge, this type of data specifically on large-scale natural facial expression videos is currently missing. We describe here the natural Facial Expressions Dataset (NFED), an fMRI dataset including responses to 1,320 short (3-second) natural facial expression video clips. These video clips are annotated with three types of labels: emotion, gender, and ethnicity, along with accompanying metadata. We validate that the dataset has good quality within and across participants and, notably, can capture temporal and spatial stimuli features. NFED provides researchers with fMRI data for understanding of the visual processing of large number of natural facial expression videos.

Objective: To investigate the efficacy and safety of Liqingtong (LQT) granules in patients with dampness-heat hyperuricemia. Methods: A randomized, double-blind, placebo-controlled pilot trial was conducted at the 983rd Hospital of the Joint Logistic Support Force of the People's Liberation Army from March 15, 2023, to August 10, 2023. In total, 119 participants were enrolled in this trial, and participants were given either LQT granules or placebo for 60 days based on a health education. The primary outcome was serum uric acid (SUA) level, and the secondary outcome was the traditional Chinese medicine (TCM) symptom score, measured on days 0, 30, and 60. Safety indicators, including liver function, kidney function, blood routine, glucose, blood lipid, blood pressure, and heart rate were tested on days 0 and 60 of the trial. The data were analyzed using Prism 9 software, and the significance level was set at P

Abstract Adipose tissue development begins in the fetal period, and continues to expand after birth. Dysregulation of adipose tissue during weaning may predispose individuals to lifelong metabolic disorders. However, the developmental remodeling of adipose tissue during weaning remains largely unexplored. Here we comprehensively compare the changes in mouse subcutaneous white adipose tissue from 7 days after birth to 7 days after weaning using single-cell RNA sequencing along with other molecular and histologic assays. We characterize the developmental trajectory of preadipocytes and indicate the commitment of preadipocytes with beige potential during weaning. Meanwhile, we find immune cells unique to weaning period, whose expression of extracellular matrix proteins implies potential regulation on preadipocyte. Finally, the strongest cell-cell interaction during weaning determined by the TGFβ ligand-receptor pairs is between preadipocytes and endotheliocytes. Our results provide a detailed and unbiased cellular landscape and offer insights into the potential regulation of adipose tissue remodeling during weaning.

Abstract Mesonephric‐type (or ‐like) adenocarcinomas (MAs) of the ovary are an uncommon and aggressive histotype. They appear to arise through transdifferentiation from Müllerian lesions creating diagnostic challenges. Thus, we aimed to develop a histologic and immunohistochemical (IHC) approach to optimize the identification of MA over its histologic mimics, such as ovarian endometrioid carcinoma (EC). First, we screened 1,537 ovarian epithelial neoplasms with a four‐marker IHC panel of GATA3, TTF1, ER, and PR followed by a morphological review of EC to identify MA in retrospective cohorts. Interobserver reproducibility for the distinction of MA versus EC was assessed in 66 cases initially without and subsequently with IHC information (four‐marker panel). Expression of PAX2, CD10, and calretinin was evaluated separately, and survival analyses were performed. We identified 23 MAs from which 22 were among 385 cases initially reported as EC (5.7%) and 1 as clear cell carcinoma. The interobserver reproducibility increased from fair to substantial (κ = 0.376–0.727) with the integration of the four‐marker IHC panel. PAX2 was the single most sensitive and specific marker to distinguish MA from EC and could be used as a first‐line marker together with ER/PR and GATA3/TTF1. Patients with MA had significantly increased risk of earlier death from disease (hazard ratio = 3.08; 95% CI, 1.62–5.85; p

Abstract Cardiovascular disease, primarily caused by atherosclerotic plaque formation, is a significant health concern. The early detection of these plaques is crucial for targeted therapies and reducing the risk of cardiovascular diseases. This study presents PlaqueNet, a solution for segmenting coronary artery plaques from coronary computed tomography angiography (CCTA) images. For feature extraction, the advanced residual net module was utilized, which integrates a deepwise residual optimization module into network branches, enhances feature extraction capabilities, avoiding information loss, and addresses gradient issues during training. To improve segmentation accuracy, a depthwise atrous spatial pyramid pooling based on bicubic efficient channel attention (DASPP-BICECA) module is introduced. The BICECA component amplifies the local feature sensitivity, whereas the DASPP component expands the network’s information-gathering scope, resulting in elevated segmentation accuracy. Additionally, BINet, a module for joint network loss evaluation, is proposed. It optimizes the segmentation model without affecting the segmentation results. When combined with the DASPP-BICECA module, BINet enhances overall efficiency. The CCTA segmentation algorithm proposed in this study outperformed the other three comparative algorithms, achieving an intersection over Union of 87.37%, Dice of 93.26%, accuracy of 93.12%, mean intersection over Union of 93.68%, mean Dice of 96.63%, and mean pixel accuracy value of 96.55%.

IntroductionIn light of the public health burden of the COVID-19 pandemic, boosting the safety and immunogenicity of COVID-19 vaccines is of great concern. Numerous Traditional Chinese medicine (TCM) preparations have shown to beneficially modulate immunity. Based on pilot experiments in mice that showed that supplementation with Huoxiang Suling Shuanghua Decoction (HSSD) significantly enhances serum anti-RBD IgG titers after inoculation with recombinant SARS-CoV-2 S-RBD protein, we conducted this randomized, double-blind, placebo-controlled clinical trial aimed to evaluate the potential immunogenicity boosting effect of oral HSSD after a third homologous immunization with Sinovac’s CoronaVac SARS-CoV-2 (CVS) inactivated vaccine.MethodsA total of 70 participants were randomly assigned (1:1 ratio) to receive a third dose of CVS vaccination and either oral placebo or oral HSSD for 7 days. Safety aspects were assessed by recording local and systemic adverse events, and by blood and urine biochemistry and liver and kidney function tests. Main outcomes evaluated included serum anti-RBD IgG titer, T lymphocyte subsets, serum IgG and IgM levels, complement components (C3 and C4), and serum cytokines (IL-6 and IFN-γ). In addition, metabolomics technology was used to analyze differential metabolite expression after supplementation with HSSD.ResultsFollowing a third CVS vaccination, significantly increased serum anti-RBD IgG titer, reduced serum IL-6 levels, increased serum IgG, IgM, and C3 and C4 levels, and improved cellular immunity, evidenced by reduce balance deviations in the distribution of lymphocyte subsets, was observed in the HSSD group compared with the placebo group. No serious adverse events were recorded in either group. Serum metabolomics results suggested that the mechanisms by which HSSD boosted the immunogenicity of the CVS vaccine are related to differential regulation of purine metabolism, vitamin B6 metabolism, folate biosynthesis, arginine and proline metabolism, and steroid hormone biosynthesis.ConclusionOral HSSD boosts the immunogenicity of the CVS vaccine in young and adult individuals. This trial provides clinical reference for evaluation of TCM immunomodulators to improve the immune response to COVID-19 vaccines.

Objective: To establish a quantitative analysis of multi-components by single marker (QAMS) method for the determination of lovastatin and lovastatin acid for the quality control of Hongqu Fuling tablets. Methods: Lovastatin as a control, the RP-HPLC method was used to establish the relationship between the correction factor of lovastatin and lovastatin acid, and to simultaneously quantify both lovastatin and lovastatin acid in the Hongqu Fuling tablets by QAMS. Results: The negative sample had no interference, the specificity of the sample detection method was good, the linear relationship between lovastatin and lovastatin acid was good within the corresponding linear range, R2>0.9999, the precision and repeatability RSD values were less than 3%, and the recovery was 98.43%～103.18%. The relative correction factor of lovastatin acid and lovastatin was 0.9000, and the repeatability was good. The average content of lovastatin acid was 0.3563 mg/g determined by QAMS, and the average content of lovastatin acid determined by external standard method was 0.3668 mg/g, and the average content of lovastatin was 1.1113 mg/g, which indicated that the constructed QAMS method could be used for the multi-component quality evaluation research of Hongqu Fuling Tablets. Conclusion: The QAMS method established in this study was stable, accurate and specific, and would provide a scientific basis for the quality control and in-depth development of Hongqu Fuling tablets.

In this paper, the effect of micro-dimple textures (produced by a laser) on the tribological properties of bearings is investigated. This study offers guidelines to reduce the friction torque of the bearing pair and addresses the problem of difficult start-ups after shutdowns. Micro-pits with different texture diameters and depths were machined on the surface of journal bearings. Then, the impact of several different texture parameters on the tribological performance of the bearing pairs was studied using an orthogonal experimental design. Subsequently, the surface morphology of the bearings before and after the friction and wear test was observed using scanning electron microscopy (SEM) and energy-dispersive spectrometry (EDS). These observations were then used to determine the type/state of friction and wear, which also improves our understanding of how texture affects the service life of bearings. The results indicate that the bearings’ micro-pit surface hardness follows an approximate parabolic spatial distribution that decreases along the micro-pit wall. Furthermore, the laser processing of surface textures was found to cause hardening in certain areas, and the chemical composition of elemental carbon and oxygen at the inner surface of processed bearings increased by 31.1% and 7.9%, respectively. Moreover, abrasive wear was identified as the primary form of wear. The textured surface’s antifriction mechanism primarily functioned to trap particles, which acted as a secondary lubrication source and altered the lubrication states by serving as a medium for supplied lubricants. The results confirm that a suitable selection of texture parameters can not only effectively reduce the friction coefficient without shortening the service life of the bearing pair but also facilitate the smooth start-up of the rotor–bearing system.

ObjectiveIntermittent energy restriction (IER) is an effective weight loss strategy. However, little is known about the dynamic effects of IER on the brain-gut-microbiome axis.MethodsIn this study, a total of 25 obese individuals successfully lost weight after a 2-month IER intervention. FMRI was used to determine the activity of brain regions. Metagenomic sequencing was performed to identify differentially abundant gut microbes and pathways in from fecal samples.ResultsOur results showed that IER longitudinally reduced the activity of obese-related brain regions at different timepoints, including the inferior frontal orbital gyrus in the cognitive control circuit, the putamen in the emotion and learning circuit, and the anterior cingulate cortex in the sensory circuit. IER longitudinally reduced E. coli abundance across multiple timepoints while elevating the abundance of obesity-related Faecalibacterium prausnitzii, Parabacteroides distasonis, and Bacterokles uniformis. Correlation analysis revealed longitudinally correlations between gut bacteria abundance alterations and brain activity changes.ConclusionsThere was dynamical alteration of BGM axis (the communication of E. coli with specific brain regions) during the weight loss under the IER.

Cardamine circaeoides Hook.f. & Thomson (CC), a herb of the genus Cardamine (family Brassicaceae), has a rich historical usage in China for both culinary and medicinal purposes. It is distinguished by its remarkable ability to hyperaccumulate selenium (Se). CC has demonstrated efficacy in the prevention of metabolic disorders. However, investigations into the effects of CC on asymptomatic hyperuricemia remain scarce. The objective of this study is to elucidate the mechanism by which CC aqueous extract (CCE) exerts its anti-hyperuricemic effects on asymptomatic hyperuricemic rats induced by potassium oxonate (PO) by integrating metabolomics and network pharmacological analysis. Asymptomatic hyperuricemia was induced by feeding rats with PO (1000 mg/kg) and CCE (0.75, 1.5, or 3 g/kg) once daily for 30 days. Various parameters, including body weight, uric acid (UA) levels, histopathology of renal tissue, and inflammatory factors (IL-1β, IL-6, IL-8, and TNF-α) were assessed. Subsequently, metabolomic analysis of kidney tissues was conducted to explore the effects of CCE on renal metabolites and the related pathways. Furthermore, network pharmacology was employed to explicate the mechanism of action of CCE components identified through UPLC-Q-TOF-MS analysis. Finally, metabolomic and network-pharmacology analyses were performed to predict crucial genes dysregulated in the disease model and rescued by CCE, which were then subjected to verification by RT-qPCR. The findings revealed that CCE significantly inhibited the UA levels from the 21st day to the 30th day. Moreover, CCE exhibited significant inhibition of IL-1β, IL-6, IL-8, and TNF-α levels in renal tissues. The dysregulation of 18 metabolites and the tyrosine, pyrimidine, cysteine, methionine, sphingolipid, and histidine metabolism pathways was prevented by CCE treatment. A joint analysis of targets predicted using the network pharmacology approach and the differential metabolites found in metabolics predicted 8 genes as potential targets of CCE, and 3 of them (PNP gene, JUN gene, and ADA gene) were verified at the mRNA level by RT-qPCR. We conclude that CCE has anti-hyperuricemia effects and alleviates renal inflammation in a rat model of hyperuricemia, and these efficacies are associated with the reversal of increased ADA, PNP, and JUN mRNA expression in renal tissues.

BackgroundIntermittent energy restriction (IER) is an effective weight loss strategy. However, the accompanying changes in spontaneous neural activity are unclear, and the relationship among anthropometric measurements, biochemical indicators, and adipokines remains ambiguous.MethodsThirty-five obese adults were recruited and received a 2-month IER intervention. Data were collected from anthropometric measurements, blood samples, and resting-state functional magnetic resonance imaging at four time points. The regional homogeneity (ReHo) method was used to explore the effects of the IER intervention. The relationships between the ReHo values of altered brain regions and changes in anthropometric measurements, biochemical indicators, and adipokines (leptin and adiponectin) were analyzed.ResultsResults showed that IER significantly improved anthropometric measurements, biochemical indicators, and adipokine levels in the successful weight loss group. The IER intervention for weight loss was associated with a significant increase in ReHo in the bilateral lingual gyrus, left calcarine, and left postcentral gyrus and a significant decrease in the right middle temporal gyrus and right cerebellum (VIII). Follow-up analyses showed that the increase in ReHo values in the right LG had a significant positive correlation with a reduction in Three-factor Eating Questionnaire (TFEQ)-disinhibition and a significant negative correlation with an increase in TFEQ-cognitive control. Furthermore, the increase in ReHo values in the left calcarine had a significant positive correlation with the reduction in TFEQ-disinhibition. However, no significant difference in ReHo was observed in the failed weight loss group.ConclusionOur study provides objective evidence that the IER intervention reshaped the ReHo of some brain regions in obese individuals, accompanied with improved anthropometric measurements, biochemical indicators, and adipokines. These results illustrated that the IER intervention for weight loss may act by decreasing the motivational drive to eat, reducing reward responses to food cues, and repairing damaged food-related self-control processes. These findings enhance our understanding of the neurobiological basis of IER for weight loss in obesity.

Deep neural networks (DNNs) have been shown to be susceptible to critical vulnerabilities when attacked by adversarial samples. This has prompted the development of attack and defense strategies similar to those used in cyberspace security. The dependence of such strategies on attack and defense mechanisms makes the associated algorithms on both sides appear as closely processes, with the defense method being particularly passive in these processes. Inspired by the dynamic defense approach proposed in cyberspace to address endless arm races, this article defines ensemble quantity, network structure, and smoothing parameters as variable ensemble attributes and proposes a stochastic ensemble strategy based on heterogeneous and redundant sub-models. The proposed method introduces the diversity and randomness characteristic of deep neural networks to alter the fixed correspondence gradient between input and output. The unpredictability and diversity of the gradients make it more difficult for attackers to directly implement white-box attacks, helping to address the extreme transferability and vulnerability of ensemble models under white-box attacks. Experimental comparison of ASR-vs.-distortion curves with different attack scenarios under CIFAR10 preliminarily demonstrates the effectiveness of the proposed method that even the highest-capacity attacker cannot easily outperform the attack success rate associated with the ensemble smoothed model, especially for untargeted attacks.

Objective: To systematically explore the effect and mechanism of melastomatis dodecandri herba (Melastoma dodecandrum Lour.) in the treatment of hepatitis based on network pharmacology. Method: We evaluated the hepatoprotective effects of M. dodecandrum in concanavalin A (Con A)-induced hepatitis in mice by assessing survival rate, histological analysis, serum transaminases, and related cytokines. Then the mechanism of action was predicted by a network pharmacology-based strategy. Based on the results, we measured the hepatic expression of related genes at mRNA level and proteins related to the phosphoinositide 3-kinase (PI3K)/protein kinase B (Akt) and nuclear factor-kappa B (NF-кB) pathways. Results: Our study results clearly demonstrated that M. dodecandrum pretreatment significantly alleviated liver injury. This was demonstrated by an increase in survival rate, decreased severity of liver damage, and reduced serum transaminase levels compared with those in the Con A group. Moreover, M. dodecandrum significantly reduced the serum levels of tumor necrosis factor-α, interleukin-6, and interferon-γ and increased the liver levels of superoxide dismutase, which indicated that M. dodecandrum exhibits anti-inflammatory and antioxidant activities. On the basis of network pharmacology, 50 nodes were selected as major hubs based on their topological importance. Pathway enrichment analyses indicated that the putative targets of M. dodecandrum mostly participate in various pathways associated with the anti-inflammation response, which implies the underlying mechanism by which M. dodecandrum acts on hepatitis. Real-time fluorescent quantitative PCR analysis showed that M. dodecandrum downregulates the mRNA expression of interleukin-6, Toll-like receptor 7, interleukin-1 receptor-associated kinase-4, NF-кB and tumor necrosis factor-α in liver tissues. Western blotting showed that M. dodecandrum pretreatment protected against inflammation through activating the PI3K-Akt pathway by upregulating phosphorylated Akt (p-Akt) expression and suppressing NF-кB activation by inhibiting the phosphorylation of IKK, IκBα, and p65. Conclusion: The present work demonstrated the hepatoprotective effects of M. dodecandrum by regulating the PI3K/Akt and NF-кB pathways in Con A-induced mice, which provide insights into the treatment of hepatitis using M. dodecandrum.

The current influenza vaccines are unable to provide effective protection in many cases, like influenza viruses strain antigenic drift or shift, and the influenza continues to cause significant annual morbidity and mortality. Improving the immune response to influenza vaccination is an unmet need. Traditional Chinese medicine (TCM) and its active ingredients are commonly known to have immunomodulatory properties. We therefore compared influenza vaccination alone or formulated with Astragali Radix (Huangqi in Chinese), and several representative ingredients of TCM, including lentinan (polysaccharide), panax notoginseng saponins (saponin), breviscapine (flavone), andrographolide (terpenoid), and a Chinese herbal compound (kangai) for their potential to enhance immune responses to influenza vaccine in mice. We found that all these TCM-adjuvants were able to increase hemagglutination inhibition (HAI) antibody titers, splenocyte proliferation, splenic T cell differentiation, bone marrow dendritic cell maturity, and both Th1 and Th2 cytokine secretion of influenza vaccine to varying degrees, and that had the characteristics of no excessive inflammatory responses and bidirectional regulation simultaneously. Taken together, our findings show that Astragali Radix exerts a more comprehensive effect on vaccine immunity, on both innate and adaptive immunity. The effects of lentinan and andrographolide on adaptive immunity were more significant, while the effects of breviscapine on innate immunity were stronger, and the other two TCM adjuvants were weaker. As the first report of a comprehensive evaluation of TCM adjuvants in influenza vaccines, the results suggest that TCM and their active ingredients are good candidates for enhancing the immune response of influenza vaccines, and that suitable TCMs can be selected based on the adjuvant requirements of different vaccines.

Deep learning-based automatic modulation recognition networks are susceptible to adversarial attacks, posing significant performance vulnerabilities. In response, we introduce a defense framework enriched by tailored autoencoder (AE) techniques. Our design features a detection AE that harnesses reconstruction errors and convolutional neural networks to discern deep features, employing thresholds from reconstruction error and Kullback–Leibler divergence to identify adversarial samples and their origin mechanisms. Additionally, a restoration AE with a multi-layered structure effectively restores adversarial samples generated via optimization methods, ensuring accurate classification. Tested rigorously on the RML2016.10a dataset, our framework proves robust against adversarial threats, presenting a versatile defense solution compatible with various deep learning models.

ObjectiveTo investigate the active ingredients, underlying anti-influenza virus effects, and mechanisms of Huoxiang Suling Shuanghua Decoction (HSSD).Materials and methodsThe therapeutic effect of HSSD were confirmed through the survival rate experiment of H1N1-infected mice. Then, the HSSD solution and the ingredients absorbed into the blood after treatment with HSSD in rats were identified by UPLC/Q-TOF MS, while the main contents of ingredients were detected by high performance liquid chromatography (HPLC). Next, a systems pharmacology approach incorporating target prediction, gene ontology (GO) enrichment, kyoto encyclopedia of genes and genomes (KEGG) pathway analysis, and molecular docking were performed to screen out the active compounds and critical pathways of HSSD in treating influenza. According to prediction results, real-time quantitative polymerase chain reaction (RT-qPCR) and immunohistochemistry assay were used to detect the mRNA and protein expression levels of critical targets in H1N1-infected mice lungs.ResultsHuoxiang Suling Shuanghua Decoction improved the survival rate of H1N1-infected mice and prolonged the mice’s lifespan. Besides, HSSD exerts an antivirus effect by decreasing the levels of hemagglutinin (HA) and nucleoprotein (NP) to inhibit the replication and proliferation of H1N1, reducing the lung pathological state, inhibiting the cell apoptosis in the lung, and regulating the abnormal responses of peripheral blood, including GRA, LYM, white blood cell (WBC), PLT, and hemoglobin (HGB). Then, 87 compounds in the HSSD solution and 20 ingredients absorbed into the blood after treatment with HSSD were identified. Based on this, combined with the network analysis and previous research on antivirus, 16 compounds were screened out as the active components. Moreover, 16 potential targets were predicted by network pharmacology analysis. Next, molecular docking results showed stable binding modes between compounds and targets. Furthermore, experimental validation results indicated that HSSD regulates the contents of Immunoglobulin A (IgA), Immunoglobulin M (IgM), and Immunoglobulin G (IgG) in serum, modulating the levels of IFN-γ, IL-6, IL-10, MCP-1, MIP-1α, and IP-10 in the lung tissue, and significantly decreasing the mRNA and protein expressions of TLR4, CD14, MyD88, NF-κB p65, HIF1 α, VEGF, IL17A, and IL6 in the lung tissue.ConclusionHuoxiang Suling Shuanghua Decoction exerts an anti-influenza effect by affecting the expressions of mRNA and protein including TLR4, CD14, MyD88, NF-kB p65, HIF-1α, VEGF, IL17A, IL6, and inhibiting the accumulation of inflammation. Our study provided experimental pieces of evidence about the practical application of HSSD in treating influenza.

Recognizing highly occluded objects is believed to arise from the interaction between the brain’s vision and cognition-controlling areas, although supporting neuroimaging data are currently limited. To explore the neural mechanism during this activity, we conducted an occlusion object recognition experiment using functional magnetic resonance imaging (fMRI). During magnet resonance examinations, 66 subjects engaged in object recognition tasks with three different occlusion degrees. Generalized linear model (GLM) analysis showed that the activation degree of the occipital lobe (inferior occipital gyrus, middle occipital gyrus, and occipital fusiform gyrus) and dorsal anterior cingulate cortex (dACC) was related to the occlusion degree of the objects. Multivariate pattern analysis (MVPA) further unearthed a considerable surge in classification precision when dACC activation was incorporated as a feature. This suggested the combined role of dACC and the occipital lobe in occluded object recognition tasks. Moreover, psychophysiological interaction (PPI) analysis disclosed that functional connectivity (FC) between the dACC and the occipital lobe was enhanced with increased occlusion, highlighting the necessity of FC between these two brain regions in effectively identifying exceedingly occluded objects. In conclusion, these findings contribute to understanding the neural mechanisms of highly occluded object recognition, augmenting our appreciation of how the brain manages incomplete visual data.

Mental health has become a new challenge in cancer treatment, with a high prevalence of depression in patients with cancer. Albiflorin (AF) and paeoniflorinn (PF) are isomers extracted from the root of Paeoniae Radix Alba (Baishao in Chinese), belonging to the monoterpene glycosides, and multiple studies have been conducted on their antidepression and anti-cancer effects. However, the effects of AF and PF on cancer-related depression are unclear. Therefore, the current study aims to investigate whether the two isomers are able to exert antidepressant-like effects and understand the underlying mechanisms in a rat model, established by combining irradiation with chronic restraint stress and solitary confinement. Our results demonstrate a significant regulation of AF and PF in the pharmacodynamic index, including the peripheral blood, organ index, behavioral traits, and HPA axis, relative to control rats. In serum and cerebral cortex metabonomics analysis, AF and PF showed a significantly restorative trend in abnormal biomarkers and regulating ether lipid metabolism, alanine, aspartate, glutamate metabolism, tryptophan metabolism, carnitine metabolism, arachidonic acid metabolism, arginine and proline metabolism pathway. Eight potential biomarkers were further screened by means of receiver operating characteristic (ROC) analysis. The data indicate that AF and PF could effectively ameliorate a depression-like state in the model rats, and the mechanism may be associated with the regulation of the neuroendocrine immune system and disrupted metabolic pathways. Further experiments are warranted to comprehensively evaluate the antidepressant effects of AF and PF in cancer-related depression. This study provides a better insight into the action mechanisms of antidepression of TCM, and provides a new perspective for the therapy of cancer-related depression.

Abstract Background In the present work, we investigated the effects of aqueous extract of Maca (AEM) on energy metabolism and immunoregulation in spleen-deficient mice. Method We established a cyclophosphamide-induced spleen-deficiency model with ginseng, a herb that strengthens splenic function, as a control. Sixty male Kunming mice were randomly divided among 5 groups: normal, model, ginseng control (1.5 g/kg), AEM high dose (1.5 g/kg), and AEM low dose (0.75 g/kg). All animals, except those in the normal group, were injected with cyclophosphamide to induce spleen deficiency. Furthermore, we investigated differences in the thermotropic behaviors of mice using the Animal Thermotropism Behavior Surveillance System to detect energy metabolism-related assays and immune regulation assays. Results Mice given AEM exhibited tropism in response to hot plate exposure. AEM inhibited loss of body weight and immune organ atrophy caused by cyclophosphamide, increased the cAMP/cGMP ratio in blood, and enhanced the activities of Na+–K+-ATPase, Ca2+–Mg2+-ATPase, lactate dehydrogenase, and hepatic glycogen. AEM significantly reversed declining white blood cells and platelet counts, and increased the hemoglobin content within peripheral blood cells. AEM improved the protein levels of IFN-γ, TNF-β, IL-2, and IL-4 in the spleen. Conclusions Maca possesses the Traditional Chinese Medicine (TCM) property of warm and appears to strengthen spleen function.

Electroencephalography (EEG)-based emotion computing has become one of the research hotspots of human-computer interaction (HCI). However, it is difficult to effectively learn the interactions between brain regions in emotional states by using traditional convolutional neural networks because there is information transmission between neurons, which constitutes the brain network structure. In this paper, we proposed a novel model combining graph convolutional network and convolutional neural network, namely MDGCN-SRCNN, aiming to fully extract features of channel connectivity in different receptive fields and deep layer abstract features to distinguish different emotions. Particularly, we add style-based recalibration module to CNN to extract deep layer features, which can better select features that are highly related to emotion. We conducted two individual experiments on SEED data set and SEED-IV data set, respectively, and the experiments proved the effectiveness of MDGCN-SRCNN model. The recognition accuracy on SEED and SEED-IV is 95.08 and 85.52%, respectively. Our model has better performance than other state-of-art methods. In addition, by visualizing the distribution of different layers features, we prove that the combination of shallow layer and deep layer features can effectively improve the recognition performance. Finally, we verified the important brain regions and the connection relationships between channels for emotion generation by analyzing the connection weights between channels after model learning.

With the continuous development of deep-learning technology, ever more advanced face-swapping methods are being proposed. Recently, face-swapping methods based on generative adversarial networks (GANs) have realized many-to-many face exchanges with few samples, which advances the development of this field. However, the images generated by previous GAN-based methods often show instability. The fundamental reason is that the GAN in these frameworks is difficult to converge to the distribution of face space in training completely. To solve this problem, we propose a novel face-swapping method based on pretrained StyleGAN generator with a stronger ability of high-quality face image generation. The critical issue is how to control StyleGAN to generate swapped images accurately. We design the control strategy of the generator based on the idea of encoding and decoding and propose an encoder called ShapeEditor to complete this task. ShapeEditor is a two-step encoder used to generate a set of coding vectors that integrate the identity and attribute of the input faces. In the first step, we extract the identity vector of the source image and the attribute vector of the target image; in the second step, we map the concatenation of the identity vector and attribute vector onto the potential internal space of StyleGAN. Extensive experiments on the test dataset show that the results of the proposed method are not only superior in clarity and authenticity than other state-of-the-art methods but also sufficiently integrate identity and attribute.

One of the greatest limitations in the field of EEG-based emotion recognition is the lack of training samples, which makes it difficult to establish effective models for emotion recognition. Inspired by the excellent achievements of generative models in image processing, we propose a data augmentation model named VAE-D2GAN for EEG-based emotion recognition using a generative adversarial network. EEG features representing different emotions are extracted as topological maps of differential entropy (DE) under five classical frequency bands. The proposed model is designed to learn the distributions of these features for real EEG signals and generate artificial samples for training. The variational auto-encoder (VAE) architecture can learn the spatial distribution of the actual data through a latent vector, and is introduced into the dual discriminator GAN to improve the diversity of the generated artificial samples. To evaluate the performance of this model, we conduct a systematic test on two public emotion EEG datasets, the SEED and the SEED-IV. The obtained recognition accuracy of the method using data augmentation shows as 92.5 and 82.3%, respectively, on the SEED and SEED-IV datasets, which is 1.5 and 3.5% higher than that of methods without using data augmentation. The experimental results show that the artificial samples generated by our model can effectively enhance the performance of the EEG-based emotion recognition.

Deep neural networks (DNNs) are proven vulnerable to attack against adversarial examples. Black-box transfer attacks pose a massive threat to AI applications without accessing target models. At present, the most effective black-box attack methods mainly adopt data enhancement methods, such as input transformation. Previous data enhancement frameworks only work on input transformations that satisfy accuracy or loss invariance. However, it does not work for other transformations that do not meet the above conditions, such as the transformation which will lose information. To solve this problem, we propose a new noise data enhancement framework (NDEF), which only transforms adversarial perturbation to avoid the above issues effectively. In addition, we introduce random erasing under this framework to prevent the over-fitting of adversarial examples. Experimental results show that the black-box attack success rate of our method Random Erasing Iterative Fast Gradient Sign Method (REI-FGSM) is 4.2% higher than DI-FGSM in six models on average and 6.6% higher than DI-FGSM in three defense models. REI-FGSM can combine with other methods to achieve excellent performance. The attack performance of SI-FGSM can be improved by 22.9% on average when combined with REI-FGSM. Besides, our combined version with DI-TI-MI-FGSM, i.e., DI-TI-MI-REI-FGSM can achieve an average attack success rate of 97.0% against three ensemble adversarial training models, which is greater than the current gradient iterative attack method. We also introduce Gaussian blur to prove the compatibility of our framework.

The graph neural network (GNN) has been widely used for graph data representation. However, the existing researches only consider the ideal balanced dataset, and the imbalanced dataset is rarely considered. Traditional methods such as resampling, reweighting, and synthetic samples that deal with imbalanced datasets are no longer applicable in GNN. This study proposes an ensemble model called Boosting-GNN, which uses GNNs as the base classifiers during boosting. In Boosting-GNN, higher weights are set for the training samples that are not correctly classified by the previous classifiers, thus achieving higher classification accuracy and better reliability. Besides, transfer learning is used to reduce computational cost and increase fitting ability. Experimental results indicate that the proposed Boosting-GNN model achieves better performance than graph convolutional network (GCN), GraphSAGE, graph attention network (GAT), simplifying graph convolutional networks (SGC), multi-scale graph convolution networks (N-GCN), and most advanced reweighting and resampling methods on synthetic imbalanced datasets, with an average performance improvement of 4.5%.

Research on visual encoding models for functional magnetic resonance imaging derived from deep neural networks, especially CNN (e.g., VGG16), has been developed. However, CNNs typically use smaller kernel sizes (e.g., 3 × 3) for feature extraction in visual encoding models. Although the receptive field size of CNN can be enlarged by increasing the network depth or subsampling, it is limited by the small size of the convolution kernel, leading to an insufficient receptive field size. In biological research, the size of the neuronal population receptive field of high-level visual encoding regions is usually three to four times that of low-level visual encoding regions. Thus, CNNs with a larger receptive field size align with the biological findings. The RepLKNet model directly expands the convolution kernel size to obtain a larger-scale receptive field. Therefore, this paper proposes a mixed model to replace CNN for feature extraction in visual encoding models. The proposed model mixes RepLKNet and VGG so that the mixed model has a receptive field of different sizes to extract more feature information from the image. The experimental results indicate that the mixed model achieves better encoding performance in multiple regions of the visual cortex than the traditional convolutional model. Also, a larger-scale receptive field should be considered in building visual encoding models so that the convolution network can play a more significant role in visual representations.

Visual encoding models based on deep neural networks (DNN) show good performance in predicting brain activity in low-level visual areas. However, due to the amount of neural data limitation, DNN-based visual encoding models are difficult to fit for high-level visual areas, resulting in insufficient encoding performance. The ventral stream suggests that higher visual areas receive information from lower visual areas, which is not fully reflected in the current encoding models. In the present study, we propose a novel visual encoding model framework which uses the hierarchy of representations in the ventral stream to improve the model’s performance in high-level visual areas. Under the framework, we propose two categories of hierarchical encoding models from the voxel and the feature perspectives to realize the hierarchical representations. From the voxel perspective, we first constructed an encoding model for the low-level visual area (V1 or V2) and extracted the voxel space predicted by the model. Then we use the extracted voxel space of the low-level visual area to predict the voxel space of the high-level visual area (V4 or LO) via constructing a voxel-to-voxel model. From the feature perspective, the feature space of the first model is extracted to predict the voxel space of the high-level visual area. The experimental results show that two categories of hierarchical encoding models effectively improve the encoding performance in V4 and LO. In addition, the proportion of the best-encoded voxels for different models in V4 and LO show that our proposed models have obvious advantages in prediction accuracy. We find that the hierarchy of representations in the ventral stream has a positive effect on improving the performance of the existing model in high-level visual areas.

Computational visual encoding models play a key role in understanding the stimulus–response characteristics of neuronal populations in the brain visual cortex. However, building such models typically faces challenges in the effective construction of non-linear feature spaces to fit the neuronal responses. In this work, we propose the GaborNet visual encoding (GaborNet-VE) model, a novel end-to-end encoding model for the visual ventral stream. This model comprises a Gabor convolutional layer, two regular convolutional layers, and a fully connected layer. The key design principle for the GaborNet-VE model is to replace regular convolutional kernels in the first convolutional layer with Gabor kernels with learnable parameters. One GaborNet-VE model efficiently and simultaneously encodes all voxels in one region of interest of functional magnetic resonance imaging data. The experimental results show that the proposed model achieves state-of-the-art prediction performance for the primary visual cortex. Moreover, the visualizations demonstrate the regularity of the region of interest fitting to the visual features and the estimated receptive fields. These results suggest that the lightweight region-based GaborNet-VE model based on combining handcrafted and deep learning features exhibits good expressiveness and biological interpretability.

Emotion recognition plays an important part in human-computer interaction (HCI). Currently, the main challenge in electroencephalogram (EEG)-based emotion recognition is the non-stationarity of EEG signals, which causes performance of the trained model decreasing over time. In this paper, we propose a two-level domain adaptation neural network (TDANN) to construct a transfer model for EEG-based emotion recognition. Specifically, deep features from the topological graph, which preserve topological information from EEG signals, are extracted using a deep neural network. These features are then passed through TDANN for two-level domain confusion. The first level uses the maximum mean discrepancy (MMD) to reduce the distribution discrepancy of deep features between source domain and target domain, and the second uses the domain adversarial neural network (DANN) to force the deep features closer to their corresponding class centers. We evaluated the domain-transfer performance of the model on both our self-built data set and the public data set SEED. In the cross-day transfer experiment, the ability to accurately discriminate joy from other emotions was high: sadness (84%), anger (87.04%), and fear (85.32%) on the self-built data set. The accuracy reached 74.93% on the SEED data set. In the cross-subject transfer experiment, the ability to accurately discriminate joy from other emotions was equally high: sadness (83.79%), anger (84.13%), and fear (81.72%) on the self-built data set. The average accuracy reached 87.9% on the SEED data set, which was higher than WGAN-DA. The experimental results demonstrate that the proposed TDANN can effectively handle the domain transfer problem in EEG-based emotion recognition.

Dual-energy computed tomography (DECT) has attracted the attention of clinical researchers because of its outstanding capabilities to identify and decompose materials. Considering that material decomposition is unstable, reconstructed material images experience severe noise magnification resulting from the measurement data. To alleviate this problem, we propose a direct material reconstruction method by establishing a novel constrained reconstruction model based on total variation (TV) and block matching 3D (BM3D) frame regularization. TV regularization preserves the sparsity in the gradient domain of material maps and helps to preserve the image edges. BM3D frame is applied to depict the similarities among various patches in reconstructed images by grouping similar 2D image blocks into 3D data arrays combined with sparse transforms. To solve the program efficiently, a practical algorithm based on alternating direction method is developed. A modified strategy of block matching denoising is designed by incorporating the polychromatic reconstructed image into the problem solution. Digital and real data phantom studies are performed to validate the performance of the proposed method. The proposed method reduces the standard deviation on the selected region of interests by an average of 95.02% and 89.03% for the digital phantom and 95.21% and 84.19% for the real data compared with the extended simultaneous algebraic reconstruction technique and TV-based method, respectively. The reconstruction results demonstrate that the proposed method has promising capabilities in direct material reconstruction and superiority over its counterparts.

Recently, visual encoding and decoding based on functional magnetic resonance imaging (fMRI) has had many achievements with the rapid development of deep network computation. In the human vision system, when people process the perceived visual content, visual information flows from primary visual cortices to high-level visual cortices and also vice versa based on the bottom-up and top-down manners, respectively. Inspired by the bidirectional information flows, we proposed a bidirectional recurrent neural network (BRNN)-based method to decode the corresponding categories from fMRI data. The forward and backward directions in the BRNN module characterized the bottom-up and top-down manners, respectively. The proposed method regarded the selected voxels in each visual area (V1, V2, V3, V4, and LO) as one node of the space sequence and fed it into the BRNN module, then combined the output of the BRNN module to decode categories with the subsequent fully connected softmax layer. This new method can use the hierarchical information representations and bidirectional information flows in human visual cortices more efficiently. Experiments demonstrated that our method could improve the accuracy of the three-level category decoding. Comparative analysis validated and revealed that correlative representations of categories were included in visual cortices because of the bidirectional information flows, in addition to the hierarchical, distributed, and complementary representations that accorded with previous studies.

Real-time functional magnetic resonance imaging neurofeedback (rtfMRI-NF) is a prospective tool to enhance the emotion regulation capability of participants and to alleviate their emotional disorders. The hippocampus is a key brain region in the emotional brain network and plays a significant role in social cognition and emotion processing in the brain. However, few studies have focused on the emotion NF of the hippocampus. This study investigated the feasibility of NF training of healthy participants to self-regulate the activation of the hippocampus and assessed the effect of rtfMRI-NF on the hippocampus before and after training. Twenty-six right-handed healthy volunteers were randomly assigned to the experimental group receiving hippocampal rtfMRI-NF (n = 13) and the control group (CG) receiving rtfMRI-NF from the intraparietal sulcus rtfMRI-NF (n = 13) and completed a total of four NF runs. The hippocampus and the intraparietal sulcus were defined based on the Montreal Neurological Institute (MNI) standard template, and NF signal was measured as a percent signal change relative to the baseline obtained by averaging the fMRI signal for the preceding 20 s long rest block. NF signal (percent signal change) was updated every 2 s and was displayed on the screen. The amplitude of low-frequency fluctuation and regional homogeneity values was calculated to evaluate the effects of NF on spontaneous neural activity in resting-state fMRI. A standard general linear model (GLM) analysis was separately conducted for each fMRI NF run. Results showed that the activation of hippocampus increased after four NF training runs. The hippocampal activity of the experiment group participants was higher than that of the CG. They also showed elevated hippocampal activity and the greater amygdala–hippocampus connectivity. The anterior temporal lobe, parahippocampal gyrus, hippocampus, and amygdala of brain regions associated with emotional processing were activated during training. We presented a proof-of-concept study using rtfMRI-NF for hippocampus up-regulation in the recall of positive autobiographical memories. The current study may provide a new method to regulate our emotions and can potentially be applied to the clinical treatment of emotional disorders.

Visual encoding models are important computational models for understanding how information is processed along the visual stream. Many improved visual encoding models have been developed from the perspective of the model architecture and the learning objective, but these are limited to the supervised learning method. From the view of unsupervised learning mechanisms, this paper utilized a pre-trained neural network to construct a visual encoding model based on contrastive self-supervised learning for the ventral visual stream measured by functional magnetic resonance imaging (fMRI). We first extracted features using the ResNet50 model pre-trained in contrastive self-supervised learning (ResNet50-CSL model), trained a linear regression model for each voxel, and finally calculated the prediction accuracy of different voxels. Compared with the ResNet50 model pre-trained in a supervised classification task, the ResNet50-CSL model achieved an equal or even relatively better encoding performance in multiple visual cortical areas. Moreover, the ResNet50-CSL model performs hierarchical representation of input visual stimuli, which is similar to the human visual cortex in its hierarchical information processing. Our experimental results suggest that the encoding model based on contrastive self-supervised learning is a strong computational model to compete with supervised models, and contrastive self-supervised learning proves an effective learning method to extract human brain-like representations.

Representation invariance plays a significant role in the performance of deep convolutional neural networks (CNNs) and human visual information processing in various complicated image-based tasks. However, there has been abounding confusion concerning the representation invariance mechanisms of the two sophisticated systems. To investigate their relationship under common conditions, we proposed a representation invariance analysis approach based on data augmentation technology. Firstly, the original image library was expanded by data augmentation. The representation invariances of CNNs and the ventral visual stream were then studied by comparing the similarities of the corresponding layer features of CNNs and the prediction performance of visual encoding models based on functional magnetic resonance imaging (fMRI) before and after data augmentation. Our experimental results suggest that the architecture of CNNs, combinations of convolutional and fully-connected layers, developed representation invariance of CNNs. Remarkably, we found representation invariance belongs to all successive stages of the ventral visual stream. Hence, the internal correlation between CNNs and the human visual system in representation invariance was revealed. Our study promotes the advancement of invariant representation of computer vision and deeper comprehension of the representation invariance mechanism of human visual information processing.

In neuroscience, all kinds of computation models were designed to answer the open question of how sensory stimuli are encoded by neurons and conversely, how sensory stimuli can be decoded from neuronal activities. Especially, functional Magnetic Resonance Imaging (fMRI) studies have made many great achievements with the rapid development of deep network computation. However, comparing with the goal of decoding orientation, position and object category from human fMRI in visual cortex, accurate reconstruction of image stimuli is a still challenging work. Current prevailing methods were composed of two independent steps, (1) decoding intermediate features from human fMRI and (2) reconstruction using the decoded intermediate features. The new concept of ‘capsule’ and ‘capsule’ based neural network were proposed recently. The ‘capsule’ represented a kind of structure containing a group of neurons to perform better feature representation. Especially, the high-level capsule’s features in the capsule network (CapsNet) contains various features of image stimuli such as semantic class, orientation, location, scale and so on, and these features can better represent the processed information inherited in the fMRI data collected in visual cortex. In this paper, a novel CapsNet architecture based visual reconstruction (CNAVR) computation model is developed to reconstruct image stimuli from human fMRI. The CNAVR is composed of linear encoding computation from capsule’s features to fMRI data and inverse reconstruction computation. In the first part, we trained the CapsNet model to obtain the non-linear mappings from images to high-level capsule’s features, and from high-level capsule’s features to images again in an end-to-end manner. In the second part, we trained the non-linear mapping from fMRI data of selected voxels to high-level capsule’s features. For a new image stimulus, we can use the method to predict the corresponding high-level capsule’s features using fMRI data, and reconstruct image stimuli with the trained reconstruction part in the CapsNet. We evaluated the proposed CNAVR method on the open dataset of handwritten digital images, and exceeded about 10% than the accuracy of all existing state-of-the-art methods on the structural similarity index (SSIM). In addition, we explained the selected voxels in specific interpretable image features to prove the effectivity and generalization of the CNAVR method.

In recent years, research on decoding brain activity based on functional magnetic resonance imaging (fMRI) has made eye-catching achievements. However, constraint-free natural image reconstruction from brain activity remains a challenge, as specifying brain activity for all possible images is impractical. The problem was often simplified by using semantic prior information or just reconstructing simple images, including digitals and letters. Without semantic prior information, we present a novel method to reconstruct natural images from the fMRI signals of human visual cortex based on the computation model of convolutional neural network (CNN). First, we extracted the unit output of viewed natural images in each layer of a pre-trained CNN as CNN features. Second, we transformed image reconstruction from fMRI signals into the problem of CNN feature visualization by training a sparse linear regression to map from the fMRI patterns to CNN features. By iteratively optimization to find the matched image, whose CNN unit features become most similar to those predicted from the brain activity, we finally achieved the promising results for the challenging constraint-free natural image reconstruction. The semantic prior information of the stimuli was not used when training decoding model, and any category of images (not constraint by the training set) could be reconstructed theoretically. We found that the reconstructed images resembled the natural stimuli, especially in position and shape. The experimental results suggest that hierarchical visual features may be an effective tool to express the human visual processing.

Density prediction is of great significance for molecular design of energetic materials, since detonation velocity linearly with density and detonation pressure increases with the density squared. However, the accuracy and generalization of former reported prediction models need further improvement, because most of them are derived from small data sets and few molecular descriptors. As shown in this paper, for molecules presenting brick-like shape or containing more hydrogen-bond donors the predicted densities have large negative deviations from experimental values. Thus, a molecular morphology descriptor η and a hydrogen-bond descriptor Hb are introduced as correction items to build 3 new QSPR models. Besides, 3694 nitro compounds are adopted as data set by this work. The accuracies are obviously improved, and the generalizations are verified by an independent test set. At the level of B3PW91/6-31G(d,p), the effective ratios (ERs) of the 3 Equations, for Δρ < 5%, are 92.7%, 91.8%, and 93.3%; for Δρ < 2%, the values are 53.5%, 51.3%, and 54.7%. At the level of B3LYP/6-31G**, for Δρ < 5%, the values are 92.3%, 91.4% and 92.9%; for Δρ < 2%, the values are 53.7%, 51.4% and 53.2%.

Decoding human brain activities, especially reconstructing human visual stimuli via functional magnetic resonance imaging (fMRI), has gained increasing attention in recent years. However, the high dimensionality and small quantity of fMRI data impose restrictions on satisfactory reconstruction, especially for the reconstruction method with deep learning requiring huge amounts of labelled samples. When compared with the deep learning method, humans can recognize a new image because our human visual system is naturally capable of extracting features from any object and comparing them. Inspired by this visual mechanism, we introduced the mechanism of comparison into deep learning method to realize better visual reconstruction by making full use of each sample and the relationship of the sample pair by learning to compare. In this way, we proposed a Siamese reconstruction network (SRN) method. By using the SRN, we improved upon the satisfying results on two fMRI recording datasets, providing 72.5% accuracy on the digit dataset and 44.6% accuracy on the character dataset. Essentially, this manner can increase the training data about from n samples to 2n sample pairs, which takes full advantage of the limited quantity of training samples. The SRN learns to converge sample pairs of the same class or disperse sample pairs of different class in feature space.

Limited-angle computed tomography (CT) image reconstruction is a challenging problem in the field of CT imaging. In some special applications, limited by the geometric space and mechanical structure of the imaging system, projections can only be collected with a scanning range of less than 90°. We call this kind of serious limited-angle problem the ultra-limited-angle problem, which is difficult to effectively alleviate by traditional iterative reconstruction algorithms. With the development of deep learning, the generative adversarial network (GAN) performs well in image inpainting tasks and can add effective image information to restore missing parts of an image. In this study, given the characteristic of GAN to generate missing information, the sinogram-inpainting-GAN (SI-GAN) is proposed to restore missing sinogram data to suppress the singularity of the truncated sinogram for ultra-limited-angle reconstruction. We propose the U-Net generator and patch-design discriminator in SI-GAN to make the network suitable for standard medical CT images. Furthermore, we propose a joint projection domain and image domain loss function, in which the weighted image domain loss can be added by the back-projection operation. Then, by inputting a paired limited-angle/180° sinogram into the network for training, we can obtain the trained model, which has extracted the continuity feature of sinogram data. Finally, the classic CT reconstruction method is used to reconstruct the images after obtaining the estimated sinograms. The simulation studies and actual data experiments indicate that the proposed method performed well to reduce the serious artifacts caused by ultra-limited-angle scanning.