Dynamic and Real-Time Object Detection Based on Deep Learning for Home Service Robots.

Home service robots operating indoors, such as inside houses and offices, require the real-time and accurate identification and location of target objects to perform service tasks efficiently. However, images captured by visual sensors while in motion states usually contain varying degrees of blurriness, presenting a significant challenge for object detection. In particular, daily life scenes contain small objects like fruits and tableware, which are often occluded, further complicating object recognition and positioning. A dynamic and real-time object detection algorithm is proposed for home service robots. This is composed of an image deblurring algorithm and an object detection algorithm. To improve the clarity of motion-blurred images, the DA-Multi-DCGAN algorithm is proposed. It comprises an embedded dynamic adjustment mechanism and a multimodal multiscale fusion structure based on robot motion and surrounding environmental information, enabling the deblurring processing of images that are captured under different motion states. Compared with DeblurGAN, DA-Multi-DCGAN had a 5.07 improvement in Peak Signal-to-Noise Ratio (PSNR) and a 0.022 improvement in Structural Similarity (SSIM). An AT-LI-YOLO method is proposed for small and occluded object detection. Based on depthwise separable convolution, this method highlights key areas and integrates salient features by embedding the attention module in the AT-Resblock to improve the sensitivity and detection precision of small objects and partially occluded objects. It also employs a lightweight network unit Lightblock to reduce the network's parameters and computational complexity, which improves its computational efficiency. Compared with YOLOv3, the mean average precision (mAP) of AT-LI-YOLO increased by 3.19%, and the detection precision of small objects, such as apples and oranges and partially occluded objects, increased by 19.12% and 29.52%, respectively. Moreover, the model inference efficiency had a 7 ms reduction in processing time. Based on the typical home activities of older people and children, the dataset Grasp-17 was established for the training and testing of the proposed method. Using the TensorRT neural network inference engine of the developed service robot prototype, the proposed dynamic and real-time object detection algorithm required 29 ms, which meets the real-time requirement of smooth vision.

The association between circulating neutrophil extracellular trap related biomarkers and retinal vein occlusion incidence: A case-control pilot study.

Retinal vein occlusion (RVO) is the second most common retinal vascular disorders and causes visual damage in a large population. Neutrophil extracellular traps (NETs) formation (NETosis) is an important cause of vascular diseases, however, the association between NETs related biomarkers and RVO development remained unclear. In this pilot study, a total of 77 RVO cases and 48 controls were included between Jan 2020 and July 2020. Besides, the circulating levels of three NETs related markers, cell-free DNA (cfDNA), myeloperoxidase (MPO)-DNA and citrullinated histone H3 (H3Cit), were detected in all the participants and thus the association between NETosis and RVO incidence was analyzed. Advanced assays were conducted to investigate the inflammation and thrombosis related biomarkers in RVO cases with higher or lower NETs biomarkers. When the results were considered, it was found that NETs biomarkers, including cfDNA, MPO-DNA and H3Cit, were increased in the RVO cases comparing with the controls (P < 0.05). Through the receiver operating characteristic (ROC) analyses, we found that circulating NETs related biomarkers demonstrated potential diagnostic effects for RVO and the AUCs of plasma cfDNA, MPO-DNA and H3Cit were 0.859, 0.871 and 0.928, respectively (P < 0.001). Through analyzing the correlations between circulating NETs markers and RVO stages and durations, inflammatory markers as well as thrombotic indexes, it was found that NETs were related with the RVO subtypes, inflammatory status and thrombus formation. In conclusion, the plasma NETs remnants are significantly increased in RVO cases. Besides, advanced studies demonstrate that inflammation as well as thrombus formation might be involved in this association.

Optimization-Based Online Initialization and Calibration of Monocular Visual-Inertial Odometry Considering Spatial-Temporal Constraints.

The online system state initialization and simultaneous spatial-temporal calibration are critical for monocular Visual-Inertial Odometry (VIO) since these parameters are either not well provided or even unknown. Although impressive performance has been achieved, most of the existing methods are designed for filter-based VIOs. For the optimization-based VIOs, there is not much online spatial-temporal calibration method in the literature. In this paper, we propose an optimization-based online initialization and spatial-temporal calibration method for VIO. The method does not need any prior knowledge about spatial and temporal configurations. It estimates the initial states of metric-scale, velocity, gravity, Inertial Measurement Unit (IMU) biases, and calibrates the coordinate transformation and time offsets between the camera and IMU sensors. The work routine of the method is as follows. First, it uses a time offset model and two short-term motion interpolation algorithms to align and interpolate the camera and IMU measurement data. Then, the aligned and interpolated results are sent to an incremental estimator to estimate the initial states and the spatial-temporal parameters. After that, a bundle adjustment is additionally included to improve the accuracy of the estimated results. Experiments using both synthetic and public datasets are performed to examine the performance of the proposed method. The results show that both the initial states and the spatial-temporal parameters can be well estimated. The method outperforms other contemporary methods used for comparison.

Comprehensive Interactome Analysis Reveals that STT3B Is Required for N-Glycosylation of Lassa Virus Glycoprotein.

Lassa virus (LASV) is the causative agent of a fatal hemorrhagic fever in humans. The glycoprotein (GP) of LASV mediates viral entry into host cells, and correct processing and modification of GP by host factors is a prerequisite for virus replication. Here, using an affinity purification-coupled mass spectrometry (AP-MS) strategy, 591 host proteins were identified as interactors of LASV GP. Gene ontology analysis was performed to functionally annotate these proteins, and the oligosaccharyltransferase (OST) complex was highly enriched. Functional studies conducted by using CRISPR-Cas9-mediated knockouts showed that STT3A and STT3B, the two catalytically active isoforms of the OST complex, are essential for the propagation of the recombinant arenavirus rLCMV/LASV glycoprotein precursor, mainly via affecting virus infectivity. Knockout of STT3B, but not STT3A, caused hypoglycosylation of LASV GP, indicating a preferential requirement of LASV for the STT3B-OST isoform. Furthermore, double knockout of magnesium transporter 1 (MAGT1) and tumor suppressor candidate 3 (TUSC3), two specific subunits of STT3B-OST, also caused hypoglycosylation of LASV GP and affected virus propagation. Site-directed mutagenesis analysis revealed that the oxidoreductase CXXC active-site motif of MAGT1 or TUSC3 is essential for the glycosylation of LASV GP. NGI-1, a small-molecule OST inhibitor, can effectively reduce virus infectivity without affecting cell viability. The STT3B-dependent N-glycosylation of GP is conserved among other arenaviruses, including both the Old World and New World groups. Our study provided a systematic view of LASV GP-host interactions and revealed the preferential requirement of STT3B for LASV GP N-glycosylation.IMPORTANCE Glycoproteins play vital roles in the arenavirus life cycle by facilitating virus entry and participating in the virus budding process. N-glycosylation of GPs is responsible for their proper functioning; however, little is known about the host factors on which the virus depends for this process. In this study, a comprehensive LASV GP interactome was characterized, and further study revealed that STT3B-dependent N-glycosylation was preferentially required by arenavirus GPs and critical for virus infectivity. The two specific thioredoxin subunits of STT3B-OST MAGT1 and TUSC3 were found to be essential for the N-glycosylation of viral GP. NGI-1, a small-molecule inhibitor of OST, also showed a robust inhibitory effect on arenavirus. Our study provides new insights into LASV GP-host interactions and extends the potential targets for the development of novel therapeutics against Lassa fever in the future.

A Subcellular Quantitative Proteomic Analysis of Herpes Simplex Virus Type 1-Infected HEK 293T Cells.

Herpes simplex virus type 1 (HSV-1) is widespread double-stranded DNA (dsDNA) virus that establishes life-long latency and causes diverse severe symptoms. The mechanisms of HSV-1 infection and HSV-1's interactions with various host cells have been studied and reviewed extensively. Type I interferons were secreted by host cells upon HSV infection and play a vital role in controlling virus proliferation. A few studies, however, have focused on HSV-1 infection without the presence of interferon (IFN) signaling. In this study, HEK 293T cells with low toll-like receptor (TLR) and stimulator of interferon genes protein (STING) expression were infected with HSV-1 and subjected to a quantitative proteomic analysis. By using a subcellular fractionation strategy and high-performance mass spectrometry, a total of 6607 host proteins were quantified, of which 498 proteins were differentially regulated. A bioinformatics analysis indicated that multiple signaling pathways might be involved in HSV-1 infection. A further functional study indicated the role of Interferon-induced transmembrane protein 3 (IFITM3), Coiled-coil-helix-coiled-coil-helix domain-containing protein 2 (CHCHD2), and Tripartite motif-containing protein 27 (TRIM27) in inhibiting viral DNA replication and proliferation. Our data provide a global view of host responses to HSV-1 infection in HEK 293T cells and identify the proteins involved in the HSV-1 infection process.

Global quantitative proteomic analysis profiles host protein expression in response to Sendai virus infection.

Sendai virus (SeV) is an enveloped nonsegmented negative-strand RNA virus that belongs to the genus Respirovirus of the Paramyxoviridae family. As a model pathogen, SeV has been extensively studied to define the basic biochemical and molecular biologic properties of the paramyxoviruses. In addition, SeV-infected host cells were widely employed to uncover the mechanism of innate immune response. To identify proteins involved in the SeV infection process or the SeV-induced innate immune response process, system-wide evaluations of SeV-host interactions have been performed. cDNA microarray, siRNA screening and phosphoproteomic analysis suggested that multiple signaling pathways are involved in SeV infection process. Here, to study SeV-host interaction, a global quantitative proteomic analysis was performed on SeV-infected HEK 293T cells. A total of 4699 host proteins were quantified, with 742 proteins being differentially regulated. Bioinformatics analysis indicated that regulated proteins were mainly involved in "interferon type I (IFN-I) signaling pathway" and "defense response to virus," suggesting that these processes play roles in SeV infection. Further RNAi-based functional studies indicated that the regulated proteins, tripartite motif (TRIM24) and TRIM27, affect SeV-induced IFN-I production. Our data provided a comprehensive view of host cell response to SeV and identified host proteins involved in the SeV infection process or the SeV-induced innate immune response process.

Activation of the RLR/MAVS Signaling Pathway by the L Protein of Mopeia Virus.

The family Arenaviridae includes several important human pathogens that can cause severe hemorrhagic fever and greatly threaten public health. As a major component of the innate immune system, the RLR/MAVS signaling pathway is involved in recognizing viral components and initiating antiviral activity. It has been reported that arenavirus infection can suppress the innate immune response, and NP and Z proteins of pathogenic arenaviruses can disrupt RLR/MAVS signaling, thus inhibiting production of type I interferon (IFN-I). However, recent studies have shown elevated IFN-I levels in certain arenavirus-infected cells. The mechanism by which arenavirus infection induces IFN-I responses remains unclear. In this study, we determined that the L polymerase (Lp) of Mopeia virus (MOPV), an Old World (OW) arenavirus, can activate the RLR/MAVS pathway and thus induce the production of IFN-I. This activation is associated with the RNA-dependent RNA polymerase activity of Lp. This study provides a foundation for further studies of interactions between arenaviruses and the innate immune system and for the elucidation of arenavirus pathogenesis. IMPORTANCE: Distinct innate immune responses are observed when hosts are infected with different arenaviruses. It has been widely accepted that NP and certain Z proteins of arenaviruses inhibit the RLR/MAVS signaling pathway. The viral components responsible for the activation of the RLR/MAVS signaling pathway remain to be determined. In the current study, we demonstrate for the first time that the Lp of MOPV, an OW arenavirus, can activate the RLR/MAVS signaling pathway and thus induce the production of IFN-I. Based on our results, we proposed that dynamic interactions exist among Lp-produced RNA, NP, and the RLR/MAVS signaling pathway, and the outcome of these interactions may determine the final IFN-I response pattern: elevated or reduced. Our study provides a possible explanation for how IFN-I can become activated during arenavirus infection and may help us gain insights into the interactions that form between different arenavirus components and the innate immune system.

Strong coupling between mid-infrared localized plasmons and phonons.

We numerically and experimentally demonstrate strong coupling between the mid-infrared localized surface plasmon resonances supported by plasmonic metamaterials and the phonon vibrational resonances of polymethyl methacrylate (PMMA) molecules. The plasmonic resonances are tuned across the phonon resonance of PMMA molecules at 52 THz to observe the strong coupling, which manifests itself as an anti-crossing feature with two newly formed plasmon-phonon modes. It is also shown that the forbidden energy gap due to mode splitting is proportional to the overlapped optical power between the plasmonic resonance mode and the PMMA molecules, providing an effective approach for manipulating the coupling strength of light-matter interaction.

Numerical study of hyperlenses for three-dimensional imaging and lithography.

The development of nanostructured metamaterials and the ability to engineer material dispersion has led to impressive advances in the diverse field of nanophotonics. Of interest to many is the enhanced ability to control, illuminate, and image with light on subwavelength scales. In this letter, we numerically demonstrate a hyperlens with unprecedented radial-resolution at 5 nm scale for both imaging and lithography applications. Both processes are shown to have accuracy that surpasses the Abbe diffraction limit in the radial direction, which has potential applications for 3D imaging and lithography. Design optimization is discussed with regards to several important hyperlens parameters.

Broadband and wide angle infrared wire-grid polarizer.

An infrared polarizer consisting of metal-insulator-metal (MIM) gratings is designed with transmittance exceeding 85% and polarization extinction ratio (ER) higher than 70 dB in the wavelength range from 1.5 to 8 µm. Moreover, the polarizer retains excellent performance even when the incident angle increases to as large as 80°. The MIM gratings support magnetic polariton (MP) in the near-infrared regime and operate non-resonantly in the mid-infrared regime, both of which result in the advantages of high extinction ratio, broadband, and wide angle. The proposed structure can find applications in polarizer, beam-splitter, filter, and isolator in the infrared range.

Wide field super-resolution surface imaging through plasmonic structured illumination microscopy.

We experimentally demonstrate a wide field surface plasmon (SP) assisted super-resolution imaging technique, plasmonic structured illumination microscopy (PSIM), by combining tunable SP interference (SPI) with structured illumination microscopy (SIM). By replacing the laser interference fringes in conventional SIM with SPI patterns, PSIM exhibits greatly enhanced resolving power thanks to the unique properties of SP waves. This PSIM technique is a wide field, surface super-resolution imaging technique with potential applications in the field of high-speed biomedical imaging.

Body Extension by Using Two Mobile Manipulators.

This paper presents a remotely operated robotic system that includes two mobile manipulators to extend the functional capabilities of a human body. Compared with previous tele-operation or robotic body extension systems, using two mobile manipulators helps with enlarging the workspace and allowing manipulation of large or long objects. The system comprises a joystick for controlling the mobile base and robotic gripper, and a motion capture system for controlling the arm poses. They together enable tele-operated dual-arm and large-space manipulation. In the experiments, a human tele-operator controls the two mobile robots to perform tasks such as handover, long object manipulation, and cooperative manipulation. The results demonstrated the effectiveness of the proposed system, resulting in extending the human body to a large space while keeping the benefits of having two limbs.

PDIA4 Is a Host Factor Important for Lymphocytic Choriomeningitis Virus Infection.

Mammalian arenaviruses are rodent-borne zoonotic viruses, some of which can cause fatal hemorrhagic diseases in humans. The first discovered arenavirus, lymphocytic choriomeningitis virus (LCMV), has a worldwide distribution and can be fatal for transplant recipients. However, no FDA-approved drugs or vaccines are currently available. In this study, using a quantitative proteomic analysis, we identified a variety of host factors that could be needed for LCMV infection, among which we found that protein disulfide isomerase A4 (PDIA4), a downstream factor of endoplasmic reticulum stress (ERS), is important for LCMV infection. Biochemical analysis revealed that LCMV glycoprotein was the main viral component accounting for PDIA4 upregulation. The inhibition of ATF6-mediated ERS could prevent the upregulation of PDIA4 that was stimulated by LCMV infection. We further found that PDIA4 can affect the LCMV viral RNA synthesis processes and release. In summary, we conclude that PDIA4 could be a new target for antiviral drugs against LCMV.

Synchronization Rather Than Finite-Time Synchronization Results of Fractional-Order Multi-Weighted Complex Networks.

This article investigates the synchronization of fractional-order multi-weighted complex networks (FMWCNs) with order α ∈ (0,1) . A useful fractional-order inequality t0C Dtα V(x(t)) ≤ -µV(x(t)) is extended to a more general form t0C Dtα V(x(t)) ≤ -µVγ(x(t)),γ ∈ (0,1] , which plays a pivotal role in studies of synchronization for FMWCNs. However, the inequality t0C Dtα V(x(t)) ≤ -µVγ(x(t)),γ ∈ (0,1) has been applied to achieve the finite-time synchronization for fractional-order systems in the absence of rigorous mathematical proofs. Based on reduction to absurdity in this article, we prove that it cannot be used to obtain finite-time synchronization results under bounded nonzero initial value conditions. Moreover, by using feedback control strategy and Lyapunov direct approach, some sufficient conditions are presented in the forms of linear matrix inequalities (LMIs) to ensure the synchronization for FMWCNs in the sense of a widely accepted definition of synchronization. Meanwhile, these proposed sufficient results cannot guarantee the finite-time synchronization of FMWCNs. Finally, two chaotic systems are given to verify the feasibility of the theoretical results.

Visualizing lymphocytic choriomeningitis virus infection in cells and living mice.

Mammarenavirus are a large family of enveloped negative-strand RNA viruses that include several agents responsible for severe hemorrhagic fevers. Until now, no FDA-licensed drug has been admitted for treating an arenavirus infection, and only few effective anti-arenavirus drugs have been tested in vivo. In this work, we designed a recombinant reporter arenavirus lymphocytic choriomeningitis virus that stably expressed nanoluciferase (LCMV-Nluc). The LCMV-Nluc was proved to share similar biological properties with wild-type LCMV and the Nluc intensity reliably reflected viral replication both in vitro and in vivo. Replication of the Nluc-encoding virus in living mice can be visualized by real-time bioluminescent imaging, and bioluminescence can be detected in a variety of organs of infected mice. This work provides a novel approach that enables real-time study of the arenavirus infection and is a convenient and valuable tool for screening of compounds that are active against arenaviruses in vitro and in living mice.

An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks.

Complex contact-rich insertion is a ubiquitous robotic manipulation skill and usually involves nonlinear and low-clearance insertion trajectories as well as varying force requirements. A hybrid trajectory and force learning framework can be utilized to generate high-quality trajectories by imitation learning and find suitable force control policies efficiently by reinforcement learning. However, with the mentioned approach, many human demonstrations are necessary to learn several tasks even when those tasks require topologically similar trajectories. Therefore, to reduce human repetitive teaching efforts for new tasks, we present an adaptive imitation framework for robot manipulation. The main contribution of this work is the development of a framework that introduces dynamic movement primitives into a hybrid trajectory and force learning framework to learn a specific class of complex contact-rich insertion tasks based on the trajectory profile of a single task instance belonging to the task class. Through experimental evaluations, we validate that the proposed framework is sample efficient, safer, and generalizes better at learning complex contact-rich insertion tasks on both simulation environments and on real hardware.

Protective Role of microRNA-200a in Diabetic Retinopathy Through Downregulation of PDLIM1.

BACKGROUND: Diabetic retinopathy (DR) is a most common microvascular complication and regarded as the leading cause of blindness in the working age population. The involvement of miR-200a in various disorders has become recognized, and the objective of this study was to identify the protective effect of miR-200a in the development of DR. METHODS: The contents of miR-200a and its potential target gene, PDZ and LIM domain protein 1 (PDLIM1), were detected in both in-vivo and in-vitro DR models. Retinal leakage and inflammatory factor concentrations were detected after vitreous injections of miR-200a/PDLIM1 vectors in mice. The cellular viability, apoptosis and cellular migration were investigated using trypan blue staining, flow cytometry and transwell assay with human retinal microvascular endothelial cells (HRMECs). Besides, the prediction and confirmation of miR-200a targeting PDLIM1 were conducted with bioinformation analyses and dual-luciferase reporter assay. RESULTS: Lower miR-200a and higher PDLIM1 levels were detected in both in-vivo and in-vitro DR models. Besides, it was found that miR-200a treatment would significantly inhibit retinal permeability and inflammatory factors. Through targeting PDLIM1, it was found that miR-200a could improve cellular viability, remit apoptotic status and reduce cellular migration significantly in high glucose-treated HRMECs. CONCLUSION: Our results demonstrated that miR-200a could be used as a potential therapy target through down-regulating PDLIM1 in DR.

Activation of the STAT3 Signaling Pathway by the RNA-Dependent RNA Polymerase Protein of Arenavirus.

Arenaviruses cause chronic and asymptomatic infections in their natural host, rodents, and several arenaviruses cause severe hemorrhagic fever that has a high mortality in infected humans, seriously threatening public health. There are currently no FDA-licensed drugs available against arenaviruses; therefore, it is important to develop novel antiviral strategies to combat them, which would be facilitated by a detailed understanding of the interactions between the viruses and their hosts. To this end, we performed a transcriptomic analysis on cells infected with arenavirus lymphocytic choriomeningitis virus (LCMV), a neglected human pathogen with clinical significance, and found that the signal transducer and activator of transcription 3 (STAT3) signaling pathway was activated. A further investigation indicated that STAT3 could be activated by the RNA-dependent RNA polymerase L protein (Lp) of LCMV. Our functional analysis found that STAT3 cannot affect LCMV multiplication in A549 cells. We also found that STAT3 was activated by the Lp of Mopeia virus and Junin virus, suggesting that this activation may be conserved across certain arenaviruses. Our study explored the interactions between arenaviruses and STAT3, which may help us to better understand the molecular and cell biology of arenaviruses.

Asiatic acid attenuates diabetic retinopathy through TLR4/MyD88/NF-κB p65 mediated modulation of microglia polarization.

AIM: This study aimed to evaluate the effects of Asiatic acid (AA), a naturally occurring compound of pentacyclic triterpenoid, on the pathological processes of diabetic retinopathy (DR). METHODS: SD rats were induced to develop early DR by intraperitoneal injection of STZ (60 mg/kg). Four weeks after injection, the diabetic rats were orally administrated with 37.5 mg/kg or 75 mg/kg AA every day for four weeks. The integrity of blood-retinal barrier (BRB) was measured by Evans blue staining. The polarization of microglia was determined by real-time PCR, western blot, and ELISA assays. The inner BRB (iBRB) or outer BRB (oBRB) breakdown was induced in human retinal endothelial cells or APRE19 cells through co-culture with high glucose and LPS-stimulated microglia BV2 cells. The damage to the iBRB and oBRB was measured using transendothelial/transepithelial electrical resistance (TEER/TER) and FITC-conjugated dextran cell permeability assays. KEY FINDINGS: Results demonstrated that AA alleviated BRB breakdown, as evidenced by decreased protein expression of occludin, claudin-5, and ZO-1. Furthermore, AA treatment suppressed inflammation and M1 polarization, while it increased M2 polarization in the retina of DR rats. In vitro, the iBRB or oBRB breakdown was alleviated by AA. LPS-induced M1-polarization of BV2 cells under high glucose condition was also repressed through AA administration. Finally, we demonstrated that AA weakened the TLR4/MyD88/NF-κB p65 signaling pathway both in vivo and in vitro. SIGNIFICANCE: AA ameliorated early DR by regulating microglia polarization via the TLR4/MyD88/NF-κB p65 pathway. These data indicate that AA is a potential candidate for DR treatment.

Obstructive Sleep Apnea is Related with the Risk of Retinal Vein Occlusion.

BACKGROUND: Retinal vein occlusion (RVO) was a vision-threatening retinal vascular disorder, however, the relationship between obstructive sleep apnea (OSA) and RVO risk remained unclear. METHODS: A total of 45 RVO cases and 45 controls between April 2018 and April 2020 were included. All the participants underwent full-night polysomnography (PSG) and thus detected the severity of OSA. Besides, the relationship between the apnea-hypopnea index (AHI) and oxidative and inflammatory biomarkers, including 8-hydroxy-2 deoxyguanosine (8-OHdG), C-reactive protein (CRP), interleukin 1 beta (IL1ß), interleukin 6 (IL6) and tumor necrosis factor alpha (TNFα) were detected. The incidences of macular edema (ME) and neovascular glaucoma (NVG) were detected in a three-months follow-up. RESULTS: In this case-control study, it was found that OSA incidence was increased in the RVO cases comparing with the cataract controls. Advanced analyses about the RVO subtypes demonstrated that incidence of OSA was higher in the central RVO (CRVO) cases comparing with branch RVO (BRVO) cases. Plasma samples from OSA cases demonstrated relatively higher concentrations of oxidative stress parameters and inflammatory biomarkers, including 8-OHdG, CRP, IL1ß, and IL6, in the RVO cases. Significant linear correlations between AHI and oxidative/inflammatory biomarkers were detected, and advanced analyses on the OSA subtypes demonstrated that these biomarkers were significantly higher in cases with later stages of OSA. In a three months follow-up, an impaired visual activity improvement rate and increased ME incidence in the OSA group among all the RVO cases were detected. CONCLUSION: OSA was related with an increased incidence of RVO. Besides, OSA would lead to increased oxidative and inflammatory biomarkers concentrations in the RVO cases. OSA could be used as a harmful prognostic factor of visual activity improvement and ME incidences. These findings highlighted the role of OSA in the development of RVO.