Bilirubin influences the predictive effect of body mass index on hospital mortality in critically ill patients.

Introduction: Body mass index (BMI) can predict mortality in critically ill patients. Moreover, mortality is related to increased bilirubin levels. Thus, herein, we aimed to investigate the effect of bilirubin levels on the usefulness of BMI in predicting mortality in critically ill patients. Methods: Data were extracted from the Medical Information Mart for Intensive Care (MIMIC IV) database. Patients were divided into two groups according to their total bilirubin levels within 24 h. Cox proportional hazard regression models were applied to obtain adjusted hazard ratios and 95 % confidence intervals for the correlation between BMI categories and hospital mortality. The dose-response relationship was flexibly modeled using a restricted cubic spline (RCS) with three knots. Results: Of the 14376 patients included, 3.4 % were underweight, 29.3 % were of normal body weight, 32.2 % were overweight, and 35.1 % were obese. For patients with total bilirubin levels <2 mg/dL, hospital mortality was significantly lower in patients with obesity than in normal body weight patients (p < 0.05). However, the opposite results were observed for patients with total bilirubin levels ≥2 mg/dL. The Cox proportional hazard regression models suggested that the risk of death was lower in patients with overweightness and obesity than in normal body weight patients when the total bilirubin levels were <2 mg/dL, but not in the other case (total bilirubin levels ≥2 mg/dL). RCS analyses showed that, for patients with total bilirubin levels <2 mg/dL, the risk of death gradually decreased with increasing BMI. Conversely, for patients with total bilirubin levels ≥2 mg/dL, this risk did not decrease with increasing BMI until reaching obesity, after which it increased rapidly. Conclusion: BMI predicted the risk of death differently in critically ill patients with different bilirubin levels.

Is Telemedicine Suitable for Patients with Chronic Inflammatory Skin Conditions? A Systematic Review.

Telemedicine, the provision of remote healthcare, has gained prominence, accelerated by the COVID-19 pandemic. It has the potential to replace routine in-person follow-up visits for patients with chronic inflammatory skin conditions. However, it remains unclear whether telemedicine can effectively substitute in-person consultations for this patient group. This systematic review assessed the effectiveness and safety of telemedicine compared with traditional in-person care for chronic inflammatory skin diseases. A comprehensive search in various databases identified 11 articles, including 5 randomized controlled trials (RCTs) and 1 clinical controlled trial (CCT). These studies evaluated telemedicine's impact on patients with psoriasis and atopic dermatitis, with varying methods like video consultations and digital platforms. The findings tentatively suggest that telemedicine does not seem to be inferior compared with in-person care, particularly in terms of condition severity and quality of life for patients with chronic inflammatory skin diseases. However, these results should be interpreted with caution due to the inherent uncertainties in the evidence. There are indications that telemedicine can offer benefits such as cost-effectiveness, time savings, and reduced travel distances, but it is important to recognize these findings as preliminary, necessitating further validation through more extensive research.

A novel pencil graphite electrode modified with an iron-based conductive metal-organic framework exhibited good ability in simultaneous sensing bisphenol A and bisphenol S.

Bisphenol A (BPA) and its substitute bisphenol S (BPS) are desirable materials widely used in manufacturing plastic products but can pose carcinogenic risks to humans. A new conductive iron-based metal-organic framework (Fe-HHTP)-modified pencil graphite electrode (PGE) for electrochemically sensing BPA and BPS was prepared and fully characterized by SEM, TEM, FT-IR, XRD, and XPS. Results showed that the optimal conditions for preparing Fe-HHTP/PGE were a pH of 6.5, a Fe-HHTP concentration of 2 mg·mL-1, a deposition potential of 0 V, and a deposition time of 100 s. The Fe-HHTP/PGE prepared under such conditions harbored a significant electrocatalytic activity with a detection limit of 0.8 nM for BPA and 1.7 nM for BPS (S/N = 3). Correspondingly, the electrochemical response current was linearly correlated to BPA and BPS, ranging from 0.01 to 100 µM. Fe-HHTP/PGE also obtained satisfactory recoveries by 93.8-102.1% and 96.0-101.3% for detecting BPA and BPS in plastic food packaging samples. Our work has provided a novel electrochemical tool to simultaneously detect BPA and BPS in food packaging samples and environmental matrixes.

Preliminary findings on diagnostic performance of computed tomography perfusion images for intracranial arterial stenosis: a retrospective study.

OBJECTIVES: Computed tomographic perfusion (CTP) can play an auxiliary role in the selection of patients with acute ischemic stroke for endovascular treatment. However, data on CTP in non-stroke patients with intracranial arterial stenosis are scarce. We aimed to investigate images in patients with asymptomatic intracranial arterial stenosis to determine the detection accuracy and interpretation time of large/medium-artery stenosis or occlusion when combining computed tomographic angiography (CTA) and CTP images. METHODS: We retrospectively reviewed 39 patients with asymptomatic intracranial arterial stenosis from our hospital database from January 2021 to August 2023 who underwent head CTP, head CTA, and digital subtraction angiography (DSA). Head CTA images were generated from the CTP data, and the diagnostic performance for each artery was assessed. Two readers independently interpreted the CTA images before and after CTP, and the results were analyzed. RESULTS: After adding CTP maps, the accuracy (area under the curve) of diagnosing internal carotid artery (R1: 0.847 vs. 0.907, R2: 0.776 vs. 0.887), middle cerebral artery (R1: 0.934 vs. 0.933, R2: 0.927 vs. 0.981), anterior cerebral artery (R1: 0.625 vs. 0.750, R2: 0.609 vs. 0.750), vertebral artery (R1: 0.743 vs. 0.764, R2: 0.748 vs. 0.846), and posterior cerebral artery (R1: 0.390 vs. 0.575, R2: 0.390 vs. 0.585) occlusions increased for both readers (p < 0.05). Mean interpretation time (R1: 72.4 ± 6.1 s vs. 67.7 ± 6.4 s, R2: 77.7 ± 3.8 s vs. 72.6 ± 4.7 s) decreased when using a combination of both images both readers (p < 0.001). CONCLUSIONS: The addition of CTP images improved the accuracy of interpreting CTA images and reduced the interpretation time in asymptomatic intracranial arterial stenosis. These findings support the use of CTP imaging in patients with asymptomatic intracranial arterial stenosis.

A comparative study of GNN and MLP based machine learning for the diagnosis of Alzheimer's Disease involving data synthesis.

Alzheimer's Disease (AD) is a neurodegenerative disease that commonly occurs in older people. It is characterized by both cognitive and functional impairment. However, as AD has an unclear pathological cause, it can be hard to diagnose with confidence. This is even more so in the early stage of Mild Cognitive Impairment (MCI). This paper proposes a U-Net based Generative Adversarial Network (GAN) to synthesize fluorodeoxyglucose - positron emission tomography (FDG-PET) from magnetic resonance imaging - T1 weighted imaging (MRI-T1WI) for further usage in AD diagnosis including its early-stage MCI. The experiments have displayed promising results with Structural Similarity Index Measure (SSIM) reaching 0.9714. Furthermore, three types of classifiers are developed, i.e., one Multi-Layer Perceptron (MLP) based classifier, two Graph Neural Network (GNN) based classifiers where one is for graph classification and the other is for node classification. 10-fold cross-validation has been conducted on all trials of experiments for classifier comparison. The performance of these three types of classifiers has been compared with the different input modalities setting and data fusion strategies. The results have shown that GNN based node classifier surpasses the other two types of classifiers, and has achieved the state-of-the-art (SOTA) performance with the best accuracy at 90.18% for 3-class classification, namely AD, MCI and normal control (NC) with the synthesized fluorodeoxyglucose - positron emission tomography (FDG-PET) features fused at the input level. Moreover, involving synthesized FDG-PET as part of the input with proper data fusion strategies has also proved to enhance all three types of classifiers' performance. This work provides support for the notion that machine learning-derived image analysis may be a useful approach to improving the diagnosis of AD.

4-Octyl itaconate inhibits poly(I:C)-induced interferon-ß secretion in mouse bone marrow-derived macrophages partially by activating Nrf2.

Viruses have become a major threat to human health. Interferon-ß (IFN-ß) has a key role in the antivirus process, as it can increase the expression of antivirus-associated genes. Itaconate and its derivatives can regulate the immune response, secretion of inflammatory factors, and pyroptosis of macrophages. The effect of itaconate on IFN-ß secretion of double-stranded RNA-induced macrophages are not well known. A derivative of itaconate, 4-octoyl itaconate (4-OI), was used to treat mouse bone marrow-derived macrophages (BMDM) induced with 100 µg/mL poly(I:C). The IFN-ß concentration was detected through ELISA, and IFN-ß mRNA expression was detected through quantitative PCR. High-throughput transcriptome sequencing was used to analyze changes in the BMDM transcriptome after 4-OI treatment. The Nrf2 expression was knocked down with siRNA.4-OI inhibited poly(I:C)-induced IFN-ß secretion and mRNA expression in BMDM. Results of transcriptome sequencing revealed that 4-OI downregulated 1047 genes and upregulated 822 genes. GO and KEGG enrichment of differently expressed genes revealed that many downregulated genes were related to the anti-virus process, whereas many upregulated genes were related to metabolism. The Nrf2 inhibitor ML385 and Nrf2 siRNA could partially reverse the inhibitory effect of 4-OI. In conclusion, 4-octyl itaconate could inhibit the poly(I:C)-induced interferon-ß secretion in BMDM partially by regulating Nrf2.

Fractional Tensor Recurrent Unit (fTRU): A Stable Forecasting Model With Long Memory.

The tensor recurrent model is a family of nonlinear dynamical systems, of which the recurrence relation consists of a p -fold (called degree- p ) tensor product. Despite such models frequently appearing in advanced recurrent neural networks (RNNs), to this date, there are limited studies on their long memory properties and stability in sequence tasks. In this article, we propose a fractional tensor recurrent model, where the tensor degree p is extended from the discrete domain to the continuous domain, so it is effectively learnable from various datasets. Theoretically, we prove that a large degree p is essential to achieve the long memory effect in a tensor recurrent model, yet it could lead to unstable dynamical behaviors. Hence, our new model, named fractional tensor recurrent unit (fTRU), is expected to seek the saddle point between long memory property and model stability during the training. We experimentally show that the proposed model achieves competitive performance with a long memory and stable manners in several forecasting tasks compared to various advanced RNNs.

A Systematic Review of Application Progress on Machine Learning-Based Natural Language Processing in Breast Cancer over the Past 5 Years.

Artificial intelligence (AI) has been steadily developing in the medical field in the past few years, and AI-based applications have advanced cancer diagnosis. Breast cancer has a massive amount of data in oncology. There has been a high level of research enthusiasm to apply AI techniques to assist in breast cancer diagnosis and improve doctors' efficiency. However, the wise utilization of tedious breast cancer-related medical care is still challenging. Over the past few years, AI-based NLP applications have been increasingly proposed in breast cancer. In this systematic review, we conduct the review using preferred reporting items for systematic reviews and meta-analyses (PRISMA) and investigate the recent five years of literature in natural language processing (NLP)-based AI applications. This systematic review aims to uncover the recent trends in this area, close the research gap, and help doctors better understand the NLP application pipeline. We first conduct an initial literature search of 202 publications from Scopus, Web of Science, PubMed, Google Scholar, and the Association for Computational Linguistics (ACL) Anthology. Then, we screen the literature based on inclusion and exclusion criteria. Next, we categorize and analyze the advantages and disadvantages of the different machine learning models. We also discuss the current challenges, such as the lack of a public dataset. Furthermore, we suggest some promising future directions, including semi-supervised learning, active learning, and transfer learning.

Natural Language Processing Applications for Computer-Aided Diagnosis in Oncology.

In the era of big data, text-based medical data, such as electronic health records (EHR) and electronic medical records (EMR), are growing rapidly. EHR and EMR are collected from patients to record their basic information, lab tests, vital signs, clinical notes, and reports. EHR and EMR contain the helpful information to assist oncologists in computer-aided diagnosis and decision making. However, it is time consuming for doctors to extract the valuable information they need and analyze the information from the EHR and EMR data. Recently, more and more research works have applied natural language processing (NLP) techniques, i.e., rule-based, machine learning-based, and deep learning-based techniques, on the EHR and EMR data for computer-aided diagnosis in oncology. The objective of this review is to narratively review the recent progress in the area of NLP applications for computer-aided diagnosis in oncology. Moreover, we intend to reduce the research gap between artificial intelligence (AI) experts and clinical specialists to design better NLP applications. We originally identified 295 articles from the three electronic databases: PubMed, Google Scholar, and ACL Anthology; then, we removed the duplicated papers and manually screened the irrelevant papers based on the content of the abstract; finally, we included a total of 23 articles after the screening process of the literature review. Furthermore, we provided an in-depth analysis and categorized these studies into seven cancer types: breast cancer, lung cancer, liver cancer, prostate cancer, pancreatic cancer, colorectal cancer, and brain tumors. Additionally, we identified the current limitations of NLP applications on supporting the clinical practices and we suggest some promising future research directions in this paper.

Integrated analyses of brain and platelet omics reveal their common altered and driven molecules in Alzheimer's disease.

Platelets may serve as a perfect peripheral source for exploring diagnostic biomarkers for Alzheimer's disease (AD); however, the molecular linkage between platelet and the brain is missing. To find the common altered and driving molecules in both brain and the platelet, we performed an integrated analysis of our platelet omics and brain omics reported in the literature, and analyzed their correlations with AD-specific pathology and cognitive impairment. By integrating the gene and protein expression profiles from 269 AD patients, we deduced 239 differentially expressed proteins (DEPs) appeared in both brain and the platelet, and 70.3% of them had consistent changes. Further analysis demonstrated that the altered brain and peripheral regulations were pinpointed into 10 imbalanced pathways. We also found that 117 DEPs, including ADAM10, were closely associated to the AD-specific ß-amyloid and tau pathologies; and the changes of IDH3B and RTN1 had a potential diagnostic value for cognitive impairment analyzed by machine learning. Finally, we identified that HMOX2 and SERPINA3 could serve as driving molecules in neurodegeneration, and they were increased and decreased in AD patients, respectively. Together, this integrated brain and platelet omics provides a valuable resource for establishing efficient peripheral diagnostic biomarkers and potential therapeutic targets for AD.

Efficiency of Multifunctional Antibacterial Hydrogels for Chronic Wound Healing in Diabetes: A Comprehensive Review.

Diabetic chronic wounds or amputation, which are complications of diabetes mellitus (DM), are a cause of great suffering for diabetics. In addition to the lack of oxygen, elevated reactive oxygen species (ROS) and reduced vascularization, microbial invasion is also a critical factor that induces non-healing chronic diabetic wounds, ie, wounds still remaining in the stage of inflammation, after which the wound tissue begins to age and becomes necrotic. To clear up the infection, alleviate the inflammation in the wound and prevent necrosis, many kinds of hydrogel have been fabricated to eliminate infections with pathogens. The unique properties of hydrogels make them ideally suited to wound dressings because they provide a moist environment for wound healing and act as a barrier against bacteria. This review article will mainly cover the recent developments and innovations of antibacterial hydrogels for diabetic chronic wound healing.

Alzheimer-like tau accumulation in dentate gyrus mossy cells induces spatial cognitive deficits by disrupting multiple memory-related signaling and inhibiting local neural circuit.

Abnormal tau accumulation and spatial memory loss constitute characteristic pathology and symptoms of Alzheimer disease (AD). Yet, the intrinsic connections and the mechanism between them are not fully understood. In the current study, we observed a prominent accumulation of the AD-like hyperphosphorylated and truncated tau (hTau N368) proteins in hippocampal dentate gyrus (DG) mossy cells of 3xTg-AD mice. Further investigation demonstrated that the ventral DG (vDG) mossy cell-specific overexpressing hTau for 3 months induced spatial cognitive deficits, while expressing hTau N368 for only 1 month caused remarkable spatial cognitive impairment with more prominent tau pathologies. By in vivo electrophysiological and optic fiber recording, we observed that the vDG mossy cell-specific overexpression of hTau N368 disrupted theta oscillations with local neural network inactivation in the dorsal DG subset, suggesting impairment of the ventral to dorsal neural circuit. The mossy cell-specific transcriptomic data revealed that multiple AD-associated signaling pathways were disrupted by hTau N368, including reduction of synapse-associated proteins, inhibition of AKT and activation of glycogen synthase kinase-3ß. Importantly, chemogenetic activating mossy cells efficiently attenuated the hTau N368-induced spatial cognitive deficits. Together, our findings indicate that the mossy cell pathological tau accumulation could induce the AD-like spatial memory deficit by inhibiting the local neural network activity, which not only reveals new pathogenesis underlying the mossy cell-related spatial memory loss but also provides a mouse model of Mossy cell-specific hTau accumulation for drug development in AD and the related tauopathies.

An intelligent early warning system of analyzing Twitter data using machine learning on COVID-19 surveillance in the US.

The World Health Organization (WHO) declared on 11th March 2020 the spread of the coronavirus disease 2019 (COVID-19) a pandemic. The traditional infectious disease surveillance had failed to alert public health authorities to intervene in time and mitigate and control the COVID-19 before it became a pandemic. Compared with traditional public health surveillance, harnessing the rich data from social media, including Twitter, has been considered a useful tool and can overcome the limitations of the traditional surveillance system. This paper proposes an intelligent COVID-19 early warning system using Twitter data with novel machine learning methods. We use the natural language processing (NLP) pre-training technique, i.e., fine-tuning BERT as a Twitter classification method. Moreover, we implement a COVID-19 forecasting model through a Twitter-based linear regression model to detect early signs of the COVID-19 outbreak. Furthermore, we develop an expert system, an early warning web application based on the proposed methods. The experimental results suggest that it is feasible to use Twitter data to provide COVID-19 surveillance and prediction in the US to support health departments' decision-making.

Applications of Explainable Artificial Intelligence in Diagnosis and Surgery.

In recent years, artificial intelligence (AI) has shown great promise in medicine. However, explainability issues make AI applications in clinical usages difficult. Some research has been conducted into explainable artificial intelligence (XAI) to overcome the limitation of the black-box nature of AI methods. Compared with AI techniques such as deep learning, XAI can provide both decision-making and explanations of the model. In this review, we conducted a survey of the recent trends in medical diagnosis and surgical applications using XAI. We have searched articles published between 2019 and 2021 from PubMed, IEEE Xplore, Association for Computing Machinery, and Google Scholar. We included articles which met the selection criteria in the review and then extracted and analyzed relevant information from the studies. Additionally, we provide an experimental showcase on breast cancer diagnosis, and illustrate how XAI can be applied in medical XAI applications. Finally, we summarize the XAI methods utilized in the medical XAI applications, the challenges that the researchers have met, and discuss the future research directions. The survey result indicates that medical XAI is a promising research direction, and this study aims to serve as a reference to medical experts and AI scientists when designing medical XAI applications.

TiO2 Hollow Spheres With Flower-Like SnO2 Shell as Anodes for Lithium-Ion Batteries.

SnO2 is a promising anode material for lithium-ion batteries due to its high theoretical specific capacity and low operation voltage. However, its poor cycling performance hinders its commercial application. In order to improve the cycling stability of SnO2 electrodes, novel flower-like SnO2/TiO2 hollow spheres were prepared by facile hydrothermal method using carbon spheres as templates. Their flower-like shell and mesoporous structure highlighted a large specific surface area and excellent ion migration performance. Their TiO2 hollow sphere matrix and 2D SnO2 nano-flakes ensured good cycle stability. The electrochemical measurements indicated that novel flower-like SnO2/TiO2 hollow spheres delivered a high specific capacity, low irreversible capacity loss and superior rate performance. After 1,000 cycles at current densities of 200 mA g-1, the capacity of the flower-like SnO2/TiO2 hollow spheres was still maintained at 720 mAh g-1. Their rate capacity reached 486 mAh g-1 when the current densities gradually increase to 2,000 mA g-1.

Silicon Nanowire Field-Effect Transistor as Label-Free Detection of Hepatitis B Virus Proteins with Opposite Net Charges.

The prevalence of hepatitis B virus (HBV) is a global healthcare threat, particularly chronic hepatitis B (CHB) that might lead to hepatocellular carcinoma (HCC) should not be neglected. Although many types of HBV diagnosis detection methods are available, some technical challenges, such as the high cost or lack of practical feasibility, need to be overcome. In this study, the polycrystalline silicon nanowire field-effect transistors (pSiNWFETs) were fabricated through commercial process technology and then chemically functionalized for sensing hepatitis B virus surface antigen (HBsAg) and hepatitis B virus X protein (HBx) at the femto-molar level. These two proteins have been suggested to be related to the HCC development, while the former is also the hallmark for HBV diagnosis, and the latter is an RNA-binding protein. Interestingly, these two proteins carried opposite net charges, which could serve as complementary candidates for evaluating the charge-based sensing mechanism in the pSiNWFET. The measurements on the threshold voltage shifts of pSiNWFETs showed a consistent correspondence to the polarity of the charges on the proteins studied. We believe that this report can pave the way towards developing an approachable tool for biomedical applications.

Identification of novel ACAN mutations in two Chinese families and genotype-phenotype correlation in patients with 74 pathogenic ACAN variations.

BACKGROUND: ACAN (OMIM 155760) is located on chromosome 15q26 and encodes the production of aggrecan. Aggrecan is a large chondroitin sulfate proteoglycan with a molecular weight of 254 kDa and contains 2530 amino acids. It is a critical structural component of the extracellular matrix of cartilage, including growth plate, articular, and intervertebral disk cartilage. It plays a key role in bone development. METHODS: Here, we describe two pedigrees with loss-of-function variants in ACAN. Whole exome sequencing was performed for the probands from each family. We illustrate the clinical variability associated with ACAN variants. RESULTS: The proband of pedigree A manifested short stature, relative macrocephaly, mild flat nasal bridge, low-set ears, short neck, and short thumbs. The proband of pedigree B had short height, abnormal vertebral development, and central precocious puberty. By trio-based whole exome sequencing and in silico analyses, we identified two de novo heterozygous variants of ACAN: NM_013227.4: c.116dupT, p.Arg40Glufs*51 and NM_013227.4: c.2367delC, p.Ser790Glnfs*20 (accession number: AC103982.10). CONCLUSION: The clinical manifestations of ACAN gene variants are diverse. ACAN gene variants are important genetic factors for short stature and should be considered as the differential diagnosis of children with idiopathic short stature (ISS).

Molecular dynamics and in silico mutagenesis on the reversible inhibitor-bound SARS-CoV-2 main protease complexes reveal the role of lateral pocket in enhancing the ligand affinity.

The 2019 novel coronavirus pandemic caused by SARS-CoV-2 remains a serious health threat to humans and there is an urgent need to develop therapeutics against this deadly virus. Recent scientific evidences have suggested that the main protease (Mpro) enzyme in SARS-CoV-2 can be an ideal drug target due to its crucial role in the viral replication and transcription processes. Therefore, there are ongoing research efforts to identify drug candidates against SARS-CoV-2 Mpro that resulted in hundreds of X-ray crystal structures of ligand-bound Mpro complexes in the Protein Data Bank (PDB) describing the interactions of different fragment chemotypes within different sites of the Mpro. In this work, we performed rigorous molecular dynamics (MD) simulation of 62 reversible ligand-Mpro complexes in the PDB to gain mechanistic insights about their interactions at the atomic level. Using a total of over 3 µs long MD trajectories, we characterized different pockets in the apo Mpro structure, and analyzed the dynamic interactions and binding affinity of ligands within those pockets. Our results identified the key residues that stabilize the ligands in the catalytic sites and other pockets of Mpro. Our analyses unraveled the role of a lateral pocket in the catalytic site in Mpro that is critical for enhancing the ligand binding to the enzyme. We also highlighted the important contribution from HIS163 in the lateral pocket towards ligand binding and affinity against Mpro through computational mutation analyses. Further, we revealed the effects of explicit water molecules and Mpro dimerization in the ligand association with the target. Thus, comprehensive molecular-level insights gained from this work can be useful to identify or design potent small molecule inhibitors against SARS-CoV-2 Mpro.

The Role and Impact of Deep Learning Methods in Computer-Aided Diagnosis Using Gastrointestinal Endoscopy.

At present, the application of artificial intelligence (AI) based on deep learning in the medical field has become more extensive and suitable for clinical practice compared with traditional machine learning. The application of traditional machine learning approaches to clinical practice is very challenging because medical data are usually uncharacteristic. However, deep learning methods with self-learning abilities can effectively make use of excellent computing abilities to learn intricate and abstract features. Thus, they are promising for the classification and detection of lesions through gastrointestinal endoscopy using a computer-aided diagnosis (CAD) system based on deep learning. This study aimed to address the research development of a CAD system based on deep learning in order to assist doctors in classifying and detecting lesions in the stomach, intestines, and esophagus. It also summarized the limitations of the current methods and finally presented a prospect for future research.

Linking reservoir ecosystems research to the sustainable development goals.

Reservoirs account for about 10% of the freshwater stored in lakes worldwide. These reservoirs are home to 'reservoir ecosystems', that is, the aquatic and non-aquatic interactive ecosystems associated with artificial lakes where water is stored, typically behind a dam, for human purposes. While reservoir ecosystems provide various ecosystem services for sustainable development, their significance in research and policy has not been well understood and not well defined in the 2030 United Nation's (UN) Agenda for Sustainable Development. To advance understanding of reservoir ecosystems and their impact on policy, here we provide an overview of research on reservoir ecosystems and link it to UN SDGs and their Targets. Based on 5280 articles published in the last three decades, we applied network visualization to construct a framework for research addressing reservoir ecosystems. The framework covers four major themes: (1) ecosystem structure and function, (2) environmental pollution and stress effects, (3) climate impacts and ecological feedbacks, and (4) ecosystem services and management. We have found that sustainable reservoir ecosystems synergistically support 121 Targets of UN SDGs (71% of all). Reservoir ecosystems have both negative and positive implications for 15 targets (9%) and negative trade-offs for only 3 targets (2%). Thirty SDG Targets (18%) are unrelated to sustainable reservoir ecosystems. The synergies and trade-offs exist in three fields, securing basic material needs (SDGs 2, 6, 7, 14 and 15), pursuing common human well-being (SDGs 1, 3, 4, 5, 8 and 10), and coordinating sustainable governance policies (SDGs 9, 11, 12, 13, 16 and 17). Exploring these linkages allows better integration of reservoir ecosystems into the UN SDGs framework and guides sustainable management of reservoir ecosystems for sustainable development.