Acquiring high-resolution light fields (LFs) is expensive. LF angular superresolution aims to synthesize the required number of views from a given sparse set of spatially high-resolution images. Existing methods struggle with sparsely sampled LFs captured with large baselines. Some methods rely on depth estimation and view reprojection, and are sensitive to textureless and occluded regions. Other non-depth based methods suffer from aliasing or blurring effects due to the large disparity. In addition, most methods require specific models for different interpolation rates, which reduces their flexibility in practice. In this paper, we propose a learning framework that overcomes these challenges by exploiting the global and local structures of LFs. Our framework includes aggregation across both the angular and spatial dimensions to fully exploit the input data and a novel bilateral upsampling module that upsamples each epipolar plane image while better preserving its local parallax structure. Furthermore, our method predicts the weights of the interpolation filters based on both subpixel offset and range difference, allowing angular superresolution at different rates with a single model. We show that our non-depth based method outperforms the state-of-the-art methods in terms of handling large disparities and flexibility on both real-world and synthetic LF images.

Multimodal brain magnetic resonance (MR) imaging is indispensable in neuroscience and neurology. However, due to the accessibility of MRI scanners and their lengthy acquisition time, multimodal MR images are not commonly available. Current MR image synthesis approaches are typically trained on independent datasets for specific tasks, leading to suboptimal performance when applied to novel datasets and tasks. Here, we present TUMSyn, a Text-guided Universal MR image Synthesis generalist model, which can flexibly generate brain MR images with demanded imaging metadata from routinely acquired scans guided by text prompts. To ensure TUMSyn's image synthesis precision, versatility, and generalizability, we first construct a brain MR database comprising 31,407 3D images with 7 MRI modalities from 13 centers. We then pre-train an MRI-specific text encoder using contrastive learning to effectively control MR image synthesis based on text prompts. Extensive experiments on diverse datasets and physician assessments indicate that TUMSyn can generate clinically meaningful MR images with specified imaging metadata in supervised and zero-shot scenarios. Therefore, TUMSyn can be utilized along with acquired MR scan(s) to facilitate large-scale MRI-based screening and diagnosis of brain diseases.

BACKGROUND:Major depressive disorder (MDD) with atypical features, namely depression with atypical features (AFD), is one of the most common clinical specifiers of MDD, closely associated with bipolar disorder (BD). However, there is still a lack of clinical guidelines for the diagnosis, treatment, and prognosis of AFD. Our study mainly focuses on three issues about how to identify AFD, what is the appropriate individualized treatment for AFD, and what are the predictive biomarkers of conversion to BD.METHODS:The Study of Individualized Diagnosis and Treatment for Depression with Atypical Features (iDoT-AFD) is a multicenter, prospective, open-label study consisting of a 12-week randomized controlled trial (RCT) and a continued follow-up until 4 years or reaching the study endpoint. It is enrolling 480 patients with AFD (120 per treatment arm), 100 patients with BD, and 100 healthy controls (HC). Multivariate dimension information is collected including clinical features, cognitive function, kynurenine pathway metabolomics, and multimodal magnetic resonance imaging (MRI) data. Firstly, multivariate informatics analyses are performed to recognize patients with AFD from participants including the first-episode and recurrent atypical depression, patients with BD, and patients with HC. Secondly, patients with atypical depression are randomly allocated to one of the four treatment groups including "single application of selective serotonin reuptake inhibitor (SSRI) or serotonin-noradrenaline reuptake inhibitor (SNRI)", "SSRI/SNRI combined with mood stabilizer," "SSRI/SNRI combined with quetiapine (≥ 150 mg/day)," or "treatment as usual (TAU)" and then followed up 12 weeks to find out the optimized treatment strategies. Thirdly, patients with atypical depression are followed up until 4 years or switching to BD, to explore the risk factors of conversion from atypical depression to BD and eventually build the risk warning model of conversion to BD.DISCUSSION:The first enrolment was in August 2019. The iDoT-AFD study explores the clinical and biological markers for the diagnosis, treatment, and prognosis of AFD and further provides evidence for clinical guidelines of AFD.TRIAL REGISTRATION:ClinicalTrials.gov NCT04209166. Registered on December 19, 2019.

Cervical cancer is one of the primary factors that endanger women's health, and Thin-prep cytologic test (TCT) has been widely applied for early screening. Automatic whole slide image (WSI) classification is highly demanded, as it can significantly reduce the workload of pathologists. Current methods are mainly based on suspicious lesion patch extraction and classification, which ignore the intrinsic relationships between suspicious patches and neglect the other patches apart from the suspicious patches, and therefore limit their robustness and generalizability. Here we propose a novel method to solve the problem, which is based on graph attention network (GAT) and supervised contrastive learning. First, for each WSI, we extract and rank a large number of representative patches based on suspicious cell detection. Then, we select the top-K and bottom-K suspicious patches to construct two graphs seperately. Next, we introduce GAT to aggregate the features from each node, and use supervised contrastive learning to obtain valuable representations of graphs. Specifically, we design a novel contrastive loss so that the latent distances between two graphs are enlarged for positive WSIs and reduced for negative WSIs. Experimental results show that the proposed GAT method outperforms conventional methods, and also demonstrate the effectiveness of supervised contrastive learning.

Artificial intelligence (AI) as an emerging technology is gaining momentum in medical imaging. Recently, deep learning-based AI techniques have been actively investigated in medical imaging, and its potential applications range from data acquisition and image reconstruction to image analysis and understanding. In this review, we focus on the use of deep learning in image reconstruction for advanced medical imaging modalities including magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET). Particularly, recent deep learning-based methods for image reconstruction will be emphasized, in accordance with their methodology designs and performances in handling volumetric imaging data. It is expected that this review can help relevant researchers understand how to adapt AI for medical imaging and which advantages can be achieved with the assistance of AI.

Purpose: Parkinson’s disease (PD) diagnosis algorithms based on quantitative susceptibility mapping (QSM) and image algorithms rely on substantia nigra (SN) labeling. However, the difference between SN labels from different experts (or segmentation algorithms) will have a negative impact on downstream diagnostic tasks, such as the decrease of the accuracy of the algorithm or different diagnostic results for the same sample. In this article, we quantify the accuracy of the algorithm on different label sets and then improve the convolutional neural network (CNN) model to obtain a high-precision and highly robust diagnosis algorithm.Methods: The logistic regression model and CNN model were first compared for classification between PD patients and healthy controls (HC), given different sets of SN labeling. Then, based on the CNN model with better performance, we further proposed a novel “gated pooling” operation and integrated it with deep learning to attain a joint framework for image segmentation and classification.Results: The experimental results show that, with different sets of SN labeling that mimic different experts, the CNN model can maintain a stable classification accuracy at around 86.4%, while the conventional logistic regression model yields a large fluctuation ranging from 78.9 to 67.9%. Furthermore, the “gated pooling” operation, after being integrated for joint image segmentation and classification, can improve the diagnosis accuracy to 86.9% consistently, which is statistically better than the baseline.Conclusion: The CNN model, compared with the conventional logistic regression model using radiomics features, has better stability in PD diagnosis. Furthermore, the joint end-to-end CNN model is shown to be suitable for PD diagnosis from the perspectives of accuracy, stability, and convenience in actual use.

According to the health monitoring and early warning evaluation of wind turbines,the modal parameter iden tification based on the ambient load excitation and the computed order analysis methods were applied in the online modal analysis of wind turbine gearbox system.An online modal parameter identification and fault diagnosis system for wind turbines was designed and developed.Comparison of test results between experimental modal and online modal parameter identification,and online modal identification parameters due to the different ambient load excitation condition,showed that methods for online modal parameter identification were feasible,real-time,stable and reliable.By now,a large amount of data for online modal measurement,dynamic characteristics due to varied ambient load excitation and fault diagnosis of wind turbines has been provided.

In non-local patch-based (NLPB) labeling, a target voxel can fuse its label from the manual labels of the atlas voxels in accordance to the patch-based voxel similarities. Although state-of-the-art NLPB method mainly focuses on labeling a single target image by many atlases, we propose a novel semi-supervised strategy to address the realistic case of only a few atlases yet many unlabeled targets. Specifically, we create an -graph of voxels, such that each target voxel can fuse its label from not only atlas voxels but also other target voxels. Meanwhile, each atlas voxel can utilize the feedbacks from the graph to check whether its expert labeling needs to be corrected. The -graph is built by applying (dual-layer) sparsity learning to all target and atlas voxels represented by their surrounding patches. By embedding the voxel labels to the graph, the target voxels can jointly compute their labels. In the experiment …

miR-21 has been reported to be a strong anti-apoptosis factor as shown that it suppresses OGD-induced apoptotic cell death and protects against ischemic neuronal death. However, the underlying mechanisms of miR-21 regulating cerebral ischemic/reperfusion (I/R) injury are not entirely understood. The

With the development of nerve stereotactic technology, brain stereotactic surgery has become an effective method for the current treatment of Parkinson's disease. The accurate localization of the target nuclei is the key issue of the treatment. In this paper, we constructed a dependence tree model to identify the target nuclei further. Theory of fuzzy connectedness was used in the segmentation. Experimental results show it was more desirable and suitable to the clinical applications.

利用荧光标记显微成像研究树突棘的形态结构是神经学的重要研究手段之一,现阶段在低信噪比图像中对树突棘检测并进行形态学分析主要依赖人工参与,使得分析缺乏客观参照系且极其耗费人力.提出了一种基于断点匹配搜索和分段采样曲线拟合的算法,能够自动检测神经树突的边缘,实现树突与突棘的分离,并计算出繁杂图像的树突与突棘的大小、个数、密度等参量.分析结果表明,此方法为树突棘图像的分析提供了高效、准确的分析工具.

The objective of this study was to investigate the compositional profiles and microbial shifts of oral microbiota during head-and-neck radiotherapy. Bioinformatic analysis based on 16S rRNA gene pyrosequencing was performed to assess the diversity and variation of oral microbiota of irradiated patients. Eight patients with head and neck cancers were involved in this study. For each patient, supragingival plaque samples were collected at seven time points before and during radiotherapy. A total of 147,232 qualified sequences were obtained through pyrosequencing and bioinformatic analysis, representing 3,460 species level operational taxonomic units (OTUs) and 140 genus level taxa. Temporal variations were observed across different time points and supported by cluster analysis based on weighted UniFrac metrics. Moreover, the low evenness of oral microbial communities in relative abundance was revealed by Lorenz curves. This study contributed to a better understanding of the detailed characterization of oral bacterial diversity of irradiated patients.

Radiotherapy is the primary treatment modality used for patients with head-and-neck cancers, but inevitably causes microorganism-related oral complications. This study aims to explore the dynamic core microbiome of oral microbiota in supragingival plaque during the course of head-and-neck radiotherapy. Eight subjects aged 26 to 70 were recruited. Dental plaque samples were collected (over seven sampling time points for each patient) before and during radiotherapy. The V1-V3 hypervariable regions of bacterial 16S rRNA genes were amplified, and the high-throughput pyrosequencing was performed. A total of 140 genera belonging to 13 phyla were found. Four phyla (Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria) and 11 genera (Streptococcus, Actinomyces, Veillonella, Capnocytophaga, Derxia, Neisseria, Rothia, Prevotella, Granulicatella, Luteococcus, and Gemella) were found in all subjects, supporting the concept of a core microbiome. Temporal variation of these major cores in relative abundance were observed, as well as a negative correlation between the number of OTUs and radiation dose. Moreover, an optimized conceptual framework was proposed for defining a dynamic core microbiome in extreme conditions such as radiotherapy. This study presents a theoretical foundation for exploring a core microbiome of communities from time series data, and may help predict community responses to perturbation as caused by exposure to ionizing radiation.

AI-2-mediated quorum sensing has been identified in various bacteria, including both Gram-negative and Gram-positive species, and numerous phenotypes have been reported to be regulated by this mechanism, using the luxS-mutant strain. But the AI-2 production process confused this regulatory function; some considered this regulation as the result of a metabolic change, which refers to an important metabolic cycle named activated methyl cycle (AMC), caused by luxS-mutant simultaneously with the defect of AI-2. Herein we hypothesized that the quorum sensing system-not the metabolic aspect-is responsible for such a regulatory function. In this study, we constructed plasmids infused with sahH and induced protein expression in the luxS-mutant strain to make the quorum-sensing system and metabolic system independent. The biofilm-related genes were investigated by real-time polymerase chain reaction (PCR), and the results demonstrated that the quorum-sensing completed strain restored the gene expression of the defective strain, but the metabolically completed one did not. This evidence supported our hypothesis that the autoinducer-2-mediated, quorum-sensing system, not the AMC, was responsible for luxS mutant regulation.

Objective To compare the biofilm early formation ability of Pseudomonas aeruginosa PAO1,Streptococcus mutans UA159 and Escherichia coli MG1655 through a biofilm quantitative analysis named BioFilm Ring Test.Method Based on th eimmobilization of magnetic beads by adherent cells,an assay was developed for the kinetic quantification of biofilm formation in this study.Result There were no significant difference between Pseudomonas aeruginosa PAO1 and Escherichia coli MG1655 in bacterial growth rate.But the biofilm formation speed of Pseudomonas aeruginosa PAO1 was much quicker than that of Escherichia coli MG1655.There were no significant difference between Escherichia coli MG1655 and Streptococcus mutans UA159 in biofilm formation speed.But the bacterial growth rate of Escherichia coli MG1655 was much higher than that of Streptococcus mutans UA159.Conclusion Each bacterial strain has its own special pattern of biofilm formation.The BioFilm Ring Test may be used as a rapid,reproducible and easy-handling method to study the kinetics of bacterial biofilm early formation.

In this paper, we evaluate different non-rigid image registration methodologies in the context of atlas-based brain image segmentation. Three non-rigid voxel-based registration regularization schemes (viscous fluid, elastic and curvature-based registration) combined with the mutual information similarity measure are compared. We conduct large-scale atlas-based segmentation experiments on a set of 20 anatomically labelled MR brain images in order to find the optimal parameter settings for each scheme. The performance of the optimal registration schemes is evaluated in their capability of accurately segmenting 49 different brain sub-structures of varying size and shape.

In this paper, we evaluate different schemes for constructing a mean shape anatomical atlas for atlas-based segmentation of MR brain images. Each atlas is constructed and validated using a database of 20 images for which detailed manual delineations of 49 different subcortical structures are available. Atlas construction and atlas based segmentation are performed by non-rigid intensity-based registration using a viscous fluid deformation model with parameters that were optimally tuned for this particular task. The segmentation performance of each atlas scheme is evaluated on the same database using a leave-one-out approach and measured by the volume overlap of corresponding regions in the ground-truth manual segmentation and the warped atlas label image.