Association between nonalcoholic steatohepatitis and high serum ferritin levels in type 2 diabetes mellitus.

OBJECTIVE: The aim of this study was to assess the role of elevated serum ferritin levels in the onset, pathological progression and prognosis of nonalcoholic fatty liver disease. Nonalcoholic fatty liver disease has been rapidly increasing worldwide. Despite extensive research on the pathogenesis of nonalcoholic fatty liver disease, a lack of sufficient clinical research on the relationship between nonalcoholic fatty liver disease and serum ferritin levels remains. METHODS: We analysed 968 patients with type 2 diabetes mellitus who underwent liver ultrasound examination and had their serum ferritin levels measured. The presence of nonalcoholic fatty liver disease and advanced liver fibrosis was determined through abdominal ultrasound examination and the nonalcoholic fatty liver disease fibrosis score. RESULTS: Compared to that in the non-nonalcoholic fatty liver disease group, the presence of hyperferritinemia was significantly more common in the nonalcoholic fatty liver disease group (83.3 vs. 56.3%, p=0.005). When patients with nonalcoholic fatty liver disease were stratified by the nonalcoholic fatty liver disease fibrosis score, those with advanced liver fibrosis exhibited a higher prevalence of hyperferritinemia (56.3, 78.9, and 88.9% for none, simple steatosis, and advanced fibrosis, respectively; p for trend=0.002). In multivariate logistic regression, liver fibrosis was independently associated with hyperferritinemia (odds ratio [OR] 1.45; 95% confidence interval [CI] 1.18-2.02; p=0.014), and this association remained significant in male patients after adjusting for other risk factors (OR 2.66; 95% CI 1.43-5.48; p=0.026). CONCLUSION: Identifying nonalcoholic fatty liver disease patients at a risk of developing nonalcoholic steatohepatitis and advanced fibrosis is crucial for implementing timely interventions and improving patient outcomes. This study highlights the potential utility of serum ferritin levels as a serum biomarker for identifying nonalcoholic steatohepatitis patients and those at a risk of late-stage fibrosis, particularly in male patients with nonalcoholic fatty liver disease.

Hierarchical Carbon Nanocages as Superior Supports for Photothermal CO2 Catalysis.

The exploitation of hierarchical carbon nanocages with superior light-to-heat conversion efficiency, together with their distinct structural, morphological, and electronic properties, in photothermal applications could provide effective solutions to long-standing challenges in diverse areas. Here, we demonstrate the discovery of pristine and nitrogen-doped hierarchical carbon nanocages as superior supports for highly loaded, small-sized Ru particles toward enhanced photothermal CO2 catalysis. A record CO production rate of 3.1 mol·gRu-1·h-1 with above 90% selectivity in flow reactors was reached for hierarchical nitrogen-doped carbon-nanocage-supported Ru clusters under 2.4 W·cm-2 illumination without external heating. Detailed studies reveal that the enhanced performance originates from the strong broadband sunlight absorption and efficient light-to-heat conversion of nanocage supports as well as the excellent intrinsic catalytic reactivity of sub-2 nm Ru particles. Our study reveals the great potential of hierarchical carbon nanocages in photothermal catalysis to reduce the fossil fuel consumption of various industrial chemical processes and stimulates interest in their exploitation for other demanding photothermal applications.

Effects of low-intensity pulsed ultrasound on the microorganisms of expressed prostatic secretion in patients with IIIB prostatitis.

To detect and analyze the changes of microorganisms in expressed prostatic secretion (EPS) of patients with IIIB prostatitis before and after low-intensity pulsed ultrasound (LIPUS) treatment, and to explore the mechanism of LIPUS in the treatment of chronic prostatitis (CP). 25 patients (study power was estimated using a Dirichlet-multinomial approach and reached 96.5% at α = 0.05 using a sample size of 25) with IIIB prostatitis who were effective in LIPUS treatment were divided into two groups before and after LIPUS treatment. High throughput second-generation sequencing technique was used to detect and analyze the relative abundance of bacterial 16 s ribosomal variable regions in EPS before and after treatment. The data were analyzed by bioinformatics software and database, and differences with P < 0.05 were considered statistically significant. Beta diversity analysis showed that there was a significant difference between groups (P = 0.046). LEfSe detected four kinds of characteristic microorganisms in the EPS of patients with IIIB prostatitis before and after LIPUS treatment. After multiple comparisons among groups by DESeq2 method, six different microorganisms were found. LIPUS may improve patients' clinical symptoms by changing the flora structure of EPS, stabilizing and affecting resident bacteria or opportunistic pathogens.

[Micro-energy medicine in the treatment of erectile dysfunction: An update].

Erectile dysfunction (ED) is one of the most common sexual disorders in males, which seriously affects the health of the patient and well-being of the family. The therapeutic strategy of ED is an individualized comprehensive treatment based on phosphodiesterase inhibitors. At present, as a new option for the treatment of ED, micro-energy medicine has attracted more and more attention in its therapeutic effects and advantages. This article presents an overview of the progress in the studies of micro-energy medicine in the treatment of ED.

Quality and Quantity of School Lunch in Nanjing: Based on Data from the Sunshine Restaurant Supervision Platform.

School lunch plays an important role in children's healthy growth. Previous investigations revealed many problems with school lunches, including unreasonable dietary structure and insufficient micronutrients. This study aimed to assess the dietary structure and nutritional quality of lunches in Nanjing primary and middle schools. A stratified cluster random sampling method was used to select 44 schools that supply lunch in 12 districts in Nanjing, with two primary and two middle schools in each district. Twenty-four primary and twenty middle schools were selected. The Mann-Whitney U test was used to explore the influencing factors. Findings revealed a serious shortage of milk and fruit in school lunches; supply of eggs, fish, shrimp, and shellfish was less than half of the recommended quantity; livestock and poultry supply exceeded the recommended level by over four times. Energy and nutrition intake were suboptimal. Provision of energy, carbohydrates, vitamins (A, B1, B2, and C), calcium, and iron in urban primary schools was significantly higher than that in non-urban primary schools. The same pattern of significantly higher nutrients was equally seen in urban middle schools compared with non-urban middle schools, indicating that food supply was affected by regional economies. Therefore, it is urgent to improve the quality of lunches, with a particular focus on those in non-urban areas.

Research progress on the relationship between chronic prostatitis/chronic pelvic pain syndrome and the microbiota of the reproductive system.

Chronic prostatitis/chronic pelvic pain syndrome (CP/CPPS) is a common pelvic pain syndrome in males, seriously affecting patients' quality of life. For a long time, CP/CPPS has been considered a complex and variable disease, and its pathogenesis remains incompletely understood. Currently, CP/CPPS is believed to be a group of diseases characterized by pelvic pain or discomfort, urinary abnormalities, and other symptoms, each with its unique etiology, clinical characteristics, and outcomes, likely resulting from the action of pathogens or (and) certain non-infectious factors. Traditionally, CP/CPPS was thought to be unrelated to bacterial infections. However, in recent years, with the development of microbiology and the advancement of high-throughput sequencing technology, an increasing number of studies have suggested that microorganisms in the reproductive system may play an important role in the pathogenesis of CP/CPPS. The unique characteristics of CP/CPPS, such as its refractory nature and tendency to recur, may be closely related to the microbiota and their biological functions in the reproductive system. The relationship between CP/CPPS and reproductive system microorganisms is one of the current hot topics in microbiology and urology, receiving considerable attention from scholars in recent years and making a series of new advances. Through this review, we will comprehensively explore the relationship between CP/CPPS and reproductive system microorganisms, and look forward to future research directions, aiming to provide new ideas and methods for clinical diagnosis and treatment, thereby improving the treatment outcomes and quality of life of CP/CPPS patients.

CKB Promotes Mitochondrial ATP Production by Suppressing Permeability Transition Pore.

Creatine kinases are essential for maintaining cellular energy balance by facilitating the reversible transfer of a phosphoryl group from ATP to creatine, however, their role in mitochondrial ATP production remains unknown. This study shows creatine kinases, including CKMT1A, CKMT1B, and CKB, are highly expressed in cells relying on the mitochondrial F1F0 ATP synthase for survival. Interestingly, silencing CKB, but not CKMT1A or CKMT1B, leads to a loss of sensitivity to the inhibition of F1F0 ATP synthase in these cells. Mechanistically, CKB promotes mitochondrial ATP but reduces glycolytic ATP production by suppressing mitochondrial calcium (mCa2+) levels, thereby preventing the activation of mitochondrial permeability transition pore (mPTP) and ensuring efficient mitochondrial ATP generation. Further, CKB achieves this regulation by suppressing mCa2+ levels through the inhibition of AKT activity. Notably, the CKB-AKT signaling axis boosts mitochondrial ATP production in cancer cells growing in a mouse tumor model. Moreover, this study also uncovers a decline in CKB expression in peripheral blood mononuclear cells with aging, accompanied by an increase in AKT signaling in these cells. These findings thus shed light on a novel signaling pathway involving CKB that directly regulates mitochondrial ATP production, potentially playing a role in both pathological and physiological conditions.

Mechanism insights into the histopathological changes of polypropylene microplastics induced gut and liver in zebrafish.

Microplastics (MPs), emerging as significant pollutants, have been consistently detected in aquatic environments, with the Yangtze River experiencing a particularly severe level of microplastic pollution, exceeding all other watersheds in China. Polypropylene (PP), the plastic most abundantly found in the middle and lower reaches of the Yangtze River Basin, has less comprehensive research results into its toxic effects. Consequently, the present investigation employed zebrafish as a model organism to delve into the toxicological impacts of polypropylene microplastics (PP-MPs) with a diameter of 5 µm across varying concentrations (300 mg/L and 600 mg/L). Using histopathological, microbiota profiling, and transcriptomic approaches, we systematically evaluated the impact of PP-MPs exposure on the intestine and liver of zebrafish. Histopathological analysis revealed that exposure to PP-MPs resulted in thinner intestinal walls, damaged intestinal mucosa, and hepatic cellular damage. Intestinal microbiota profiling demonstrated that, the richness, uniformity, diversity, and homogeneity of gut microbes significantly increased after the PP-MPs exposure at high concentration. These alterations were accompanied by shifts in the relative abundance of microbiota associated with intestinal pathologies, suggesting a profound impact on the intestinal microbial community structure. Concurrently, hepatic transcriptome analysis and RT-qPCR indicated that the downregulation of pathways and genes associated with cell proliferation regulation and DNA damage repair mechanisms contributed to hepatic cellular damage, ultimately exerting adverse effects on the liver. Correlation analysis between the intestinal microbiota and liver transcriptome profiles further highlighted significant associations between intestinal microbiota and the downregulated hepatic pathways. Collectively, these results provide novel insights into the subacute toxicological mechanisms of PP-MPs in aquatic organisms and highlight the need for further research on the ecological and health risks associated with PP-MPs pollution.

A feasible interlayer strategy for simultaneous light and heat management in photothermal catalysis.

Photothermal conversion represents one crucial approach for solar energy harvesting and its exploitation as a sustainable alternative to fossil fuels; however, an efficient, cost-effective, and generalized approach to enhance the photothermal conversion processes is still missing. Herein, we develop a feasible and efficient photothermal conversion strategy that achieves simultaneous light and heat management using supported metal clusters and WSe2 interlayer toward enhanced CO2 hydrogenation photothermal catalysis. The interlayer can simultaneously reduce heat loss in the catalytic layer and improve light absorption, leading to an 8-fold higher CO2 conversion rate than the controls. The optical and thermal performance of WSe2 interlayered catalysts on different substrates was quantified using Raman spectroscopy. This work demonstrates a feasible and generalized approach for effective light and heat management in solar harvesting. It also provides important design guidelines for efficient photothermal converters that facilitate the remediation of the energy and environmental crises faced by humans.

Tuning Adsorbate-Mediated Strong Metal-Support Interaction by Oxygen Vacancy: A Case Study in Ru/TiO2.

The adsorbate-mediated strong metal-support interaction (A-SMSI) offers a reversible means of altering the selectivity of supported metal catalysts, thereby providing a powerful tool for facile modulation of catalytic performance. However, the fundamental understanding of A-SMSI remains inadequate and methods for tuning A-SMSI are still in their nascent stages, impeding its stabilization under reaction conditions. Here, we report that the initial concentration of oxygen vacancy in oxide supports plays a key role in tuning the A-SMSI between Ru nanoparticles and defected titania (TiO2-x). Based on this new understanding, we demonstrate the in situ formation of A-SMSI under reaction conditions, obviating the typically required CO2-rich pretreatment. The as-formed A-SMSI layer exhibits remarkable stability at various temperatures, enabling excellent activity, selectivity and long-term stability in catalyzing the reverse water gas-shift reaction. This study deepens the understanding of the A-SMSI and the ability to stabilize A-SMSI under reaction conditions represents a key step for practical catalytic applications.

Lactobacillus reuteri JCM 1112 ameliorates chronic acrylamide-induced glucose metabolism disorder via the bile acid-TGR5-GLP-1 axis and modulates intestinal oxidative stress in mice.

Acrylamide (AA) is a toxic food contaminant that has been reported to cause glucose metabolism disorders (GMD) at high doses. However, it is unclear whether chronic low-dose AA can induce GMD and whether probiotics can alleviate AA-induced GMD. Here, C57BL/6N mice were orally administered with 5 mg per kg bw AA for 10 weeks, followed by another 3 weeks of glucagon-like peptide-1 (GLP-1) analogue (dulaglutide) treatment. Chronic low-dose AA exposure increased the blood glucose level and decreased serum insulin and GLP-1 levels, whereas dulaglutide treatment decreased the blood glucose level and increased the serum insulin level in AA-exposed mice. Then, mice were administered with AA or AA + INT-777 (Takeda G-protein-coupled receptor 5 (TGR5) agonist) for 10 weeks. INT-777 treatment reversed AA-induced downregulation of ileal TGR5 and proglucagon (PG) gene expression and decreased the serum GLP-1 level. These findings indicated that chronic low-dose AA induced GMD via inhibiting the TGR5-GLP-1 axis. Finally, mice were administered with AA for 10 weeks, followed by another 3 weeks of Lactobacillus reuteri JCM 1112 supplementation. L. reuteri supplementation significantly increased serum glucose, insulin and GLP-1 levels, upregulated ileal TGR5 and PG gene expression, and effectively restored the imbalance of bile acid (BA) metabolism in AA-exposed mice, demonstrating that L. reuteri ameliorates chronic AA-induced GMD via the BA-TGR5-GLP-1 axis. In addition, L. reuteri significantly enhanced ileal superoxide dismutase and catalase activities and total antioxidant capacity, thereby preventing chronic AA-induced oxidative stress. Our research provides new insights into the GMD toxicity of chronic low-dose AA and confirms the role of probiotics in alleviating AA-induced GMD.

Non-Photochemical Origin of Selectivity Difference between Light and Dark Catalytic Conditions.

The interest in introducing light into heterogeneous catalysis is driven not only by the urgent need of replacing fossil energy but also by the promise of controlling product selectivity by light. The product selectivity differences observed in recent studies between light and dark reactions are often attributed to photochemical effects. Here, we report the discovery of a non-photochemical origin of selectivity difference, at essentially the same CO2 conversion rate, between photothermal and thermal CO2 hydrogenation reactions over a Ru/TiO2-x catalyst. While the presence of the photochemical effect from ultraviolet light is confirmed, it merely enhances the catalytic activity. Systematic investigation reveals that the gradual formation of an adsorbate-mediated strong metal-support interaction under catalytic conditions is responsible for the variation in the catalytic selectivity. We demonstrate that differences in product selectivity under light/dark reactions do not necessarily originate from photochemical effects. Our study refines the basis for determining photochemical effects and highlights the importance of excluding non-photochemical effects in mechanistic studies of light-controlled product selectivity.

Hydrogen Bubbles: Harmonizing Local Hydrogen Transfer for Efficient Plastic Hydro-Depolymerization.

Hydro-depolymerization presents a promising avenue for transforming plastic waste into high-value hydrocarbons, offering significant potential for value-added recycling. However, a major challenge in this method arises from kinetic limitations due to insufficient hydrogen concentration near the active sites, requiring optimal catalytic performance only at higher hydrogen pressures. In this study, we address this hurdle by developing "hydrogen bubble catalysts" featuring Ru nanoparticles within mesoporous SBA-15 channels (Ru/SBA). The distinctive feature of Ru/SBA catalysts lies in their capacity for physical hydrogen storage and chemically reversible hydrogen spillover, ensuring a timely and ample hydrogen supply. Under identical reaction conditions, the catalytic activity of Ru/SBA surpassed that of Ru/SiO2 (no hydrogen storage capacity) by over 4-fold. This substantial enhancement in catalytic performance provides significant opportunities for near atmospheric pressure hydro-depolymerization of plastic waste.

Layered Double Hydroxide Derivatives for Polyolefin Upcycling.

While Ru-catalyzed hydrogenolysis holds significant promise in converting waste polyolefins into value-added alkane fuels, a major constraint is the high cost of noble metal catalysts. In this work, we propose, for the first time, that Co-based catalysts derived from CoAl-layered double hydroxide (LDH) are alternatives for efficient polyolefin hydrogenolysis. Leveraging the chemical flexibility of the LDH platform, we reveal that metallic Co species serve as highly efficient active sites for polyolefin hydrogenolysis. Furthermore, we introduced Ni into the Co framework to tackle the issue of restricted hydrogenation ability associated with contiguous Co-Co sites. In-situ analysis indicates that the integration of Ni induces electron transfer and facilitates hydrogen spillover. This dual effect synergistically enhances the hydrogenation/desorption of olefin intermediates, resulting in a significant reduction in the yield of low-value CH4 from 27.1 to 12.6%. Through leveraging the unique properties of LDH, we have developed efficient and cost-effective catalysts for the sustainable recycling and valorization of waste polyolefin materials.

Immobilization mechanisms of Sr(II), Ni(II), and Cd(II) on glomalin-related soil protein in mangrove sediments at the microscopic scale.

Glomalin-related soil protein (GRSP) is a significant component in the sequestration of heavy metal in soils, but its mechanisms for metal adsorption are poorly known. This study combined spectroscopic data with molecular docking simulations to reveal metal adsorption onto GRSP's surface functional groups at the molecular level. The EXAFS combined with FTIR and XPS analyses indicated that the adsorption of Cd(II), Sr(II), and Ni(II) by GRSP occurred mainly through the coordination of -OH and -COOH groups with the metal. The -COOH and -OH groups bound to the metal as electron donors and the electron density of the oxygen atom decreased, suggesting that electrostatic attraction might be involved in the adsorption process. Two-dimensional correlation spectroscopy revealed that preferential adsorption occurred on GRSP for the metal in sequential order of -COOH groups followed by -OH groups. The presence of the Ni-C shell in the Ni EXAFS spectrum suggested that Ni formed organometallic complexes with the GRSP surface. However, Sr-C and Cd-C were absent in the second shell of the Sr and Cd spectra, which was attributed to the adsorption of Sr and Cd ions with large hydration ion radius by GRSP to form outer-sphere complexes. Through molecular docking simulations, negatively charged residues such as ASP151 and ASP472 in GRSP were found to provide electrostatic attraction and ligand combination for the metal adsorption, which was consistent with the spectroscopic analyses. Overall, these findings provided new insights into the interaction mechanisms between GRSP and metals, which will help deepen our understanding of the ecological functions of GRSP in metal sequestration.

M-type refractive index profile erbium-doped fiber for high-efficiency multicore EDFA.

Cladding-pumped multicore erbium-doped fiber is an important element for future spatial division multiplexing (SDM) amplification. We propose an M-type erbium-doped multicore fiber to achieve high-efficiency SDM amplification. The performance of cladding-pumped erbium-doped fiber with a central refractive index depression has been investigated, and the M-type fiber has better amplification performance than conventional fibers by reducing the signal mode overlap with the doped region. The experiment results show that the M-type 4-core erbium-doped fiber has a gain improvement of 2.8 dB compared with conventional 4-core fiber. The pump conversion efficiency (PCE) has been enhanced from 4.47% to 8.01%. For a 7.0 W pump power at 976 nm, the M-type fiber exhibits an average gain of 20.0 dB and an average noise fiber of 6.8 dB at the L-band. The core-to-core gain variation is less than 1.6 dB.

A novel deep-learning model based on τ-shaped convolutional network (τNet) with long short-term memory (LSTM) for physiological fatigue detection from EEG and EOG signals.

In recent years, fatigue driving has become the main cause of traffic accidents, leading to increased attention towards fatigue detection systems. However, the pooling and strided convolutional operations in fatigue detection algorithm based on traditional deep learning methods may led to the loss of some useful information. This paper proposed a novel τ -shaped convolutional network ( τ Net ) aiming to address this issue. Unlike traditional network structures, τ Net incorporates the operations of upsampling features and concatenating high- and low-level features, enabling full utilization of useful information. Moreover, considering that the fatigue state is a mental state involving temporal evolution, we proposed the novel long short-term memory (LSTM)- τ -shaped convolutional network (LSTM- τ Net ), a parallel structure composed of LSTM and τ Net for fatigue detection, where τ Net extracts time-invariant features with location information, and LSTM extracts long temporal dependencies. We compared LSTM- τ Net with six competing methods based on two datasets. Results showed that the proposed algorithm achieved higher classification accuracy than the other methods, with 94.25% on EEG data (binary classification) and 82.19% on EOG data (triple classification). Additionally, the proposed algorithm exhibits low computational cost, good training stability, and robustness against insufficient training. Therefore, it is promising for further implementation of fatigue online detection systems.

Readiness for hospital discharge and its association with post-discharge outcomes among oesophageal cancer patients after oesophagectomy: A prospective observational study.

AIM: To examine the level and influencing factors of discharge readiness among patients with oesophageal cancer following oesophagectomy and to explore its association with post-discharge outcomes (post-discharge coping difficulty and unplanned readmission). BACKGROUND: Oesophageal cancer is common and usually treated via oesophagectomy in China. The assessment of patient's discharge readiness gradually attracts attention as patients tend to be discharged more quickly. DESIGN: Prospective observational study. The STROBE statement was followed. METHODS: In total, 154 participants with oesophageal cancer after oesophagectomy were recruited in a tertiary cancer centre in Southern China from July 2019 to January 2020. The participants completed a demographic and disease-related questionnaire, the Quality of Discharge Teaching Scale and Readiness for Hospital Discharge Scale before discharge. Post-discharge outcomes were investigated on the 21st day (post-discharge coping difficulty) and 30th day (unplanned readmission) after discharge separately. Multiple linear regressions were used for statistical analysis. RESULTS: The mean scores of discharge readiness and quality of discharge teaching were (154.02 ± 31.58) and (138.20 ± 24.20) respectively. The quality of discharge teaching, self-care ability, dysphagia and primary caregiver mainly influenced patient's discharge readiness and explained 63.0% of the variance. The low discharge readiness could predict more risk of post-discharge coping difficulty (r = -0.729, p < 0.01) and unplanned readmission (t = -2.721, p < 0.01). CONCLUSIONS: Discharge readiness among patients with oesophageal cancer following oesophagectomy is influenced by various factors, especially the quality of discharge teaching. A high discharge readiness corresponds to good post-discharge outcomes. IMPLICATIONS FOR THE PROFESSION AND PATIENT CARE: Healthcare professionals should improve the discharge readiness by constructing high-quality discharge teaching, cultivating patients' self-care ability, mobilizing family participation and alleviating dysphagia to decrease adverse post-discharge outcomes among patients with oesophageal cancer. PATIENTS OR PUBLIC CONTRIBUTION: Patients with oesophageal cancer after oesophagectomy who met the inclusion criteria were recruited.

A generative adaptive convolutional neural network with attention mechanism for driver fatigue detection with class-imbalanced and insufficient data.

Over the past few years, fatigue driving has emerged as one of the main causes of traffic accidents, necessitating the development of driver fatigue detection systems. However, many existing methods involves tedious manual parameter tunings, a process that is both time-consuming and results in task-specific models. On the other hand, most of the researches on fatigue recognition are based on class-balanced and sufficient data, and effectively "mine" meaningful information from class-imbalanced and insufficient data for fatigue recognition is still a challenge. In this paper, we proposed two novel models, the attention-based residual adaptive multiscale fully convolutional network-long short term memory network (ARMFCN-LSTM), and the Generative ARMFCN-LSTM (GARMFCN-LSTM) aiming to address this issue. ARMFCN-LSTM excels at automatically extracting multiscale representations through adaptive multiscale temporal convolutions, while capturing temporal dependency features through LSTM. GARMFCN-LSTM integrates Wasserstein GAN with gradient penalty (WGAN-GP) into ARMFCN-LSTM to improve driver fatigue detection performance by alleviating data scarcity and addressing class imbalances. Experimental results show that ARMFCN-LSTM achieves the highest classification accuracy of 95.84% in driver fatigue detection on the class-balanced EEG dataset (binary classification), and GARMFCN-LSTM attained an improved classification accuracy of 84.70% on the class-imbalanced EOG dataset (triple classification), surpassing the competing methods. Therefore, the proposed models are promising for further implementations in online driver fatigue detection systems.