SUnet: A multi-organ segmentation network based on multiple attention

Organ segmentation in abdominal or thoracic computed tomography (CT) images plays a crucial role in medical diagnosis as it enables doctors to locate and evaluate organ abnormalities quickly, thereby guiding surgical planning, and aiding treatment decision-making. This paper proposes a novel and efficient medical image segmentation method called SUnet for multi-organ segmentation in the abdomen and thorax. SUnet is a fully attention-based neural network. Firstly, an efficient spatial reduction attention (ESRA) module is introduced not only to extract image features better, but also to reduce overall model parameters, and to alleviate overfitting. Secondly, SUnet's multiple attention-based feature fusion module enables effective cross-scale feature integration. Additionally, an enhanced attention gate (EAG) module is considered by using grouped convolution and residual connections, providing richer semantic features. We evaluate the performance of the proposed model on synapse multiple organ segmentation dataset and automated cardiac diagnostic challenge dataset. SUnet achieves an average Dice of 84.29% and 92.25% on these two datasets, respectively, outperforming other models of similar complexity and size, and achieving state-of-the-art results.

Early childhood household instability, adolescent structural neural network architecture, and young adulthood depression: A 21-year longitudinal study

Unstable and unpredictable environments are linked to risk for psychopathology, but the underlying neural mechanisms that explain how instability relate to subsequent mental health concerns remain unclear. In particular, few studies have focused on the association between instability and white matter structures despite white matter playing a crucial role for neural development. In a longitudinal sample recruited from a population-based study (N = 237), household instability (residential moves, changes in household composition, caregiver transitions in the first 5 years) was examined in association with adolescent structural network organization (network integration, segregation, and robustness of white matter connectomes; Mage = 15.87) and young adulthood anxiety and depression (six years later). Results indicate that greater instability related to greater global network efficiency, and this association remained after accounting for other types of adversity (e.g., harsh parenting, neglect, food insecurity). Moreover, instability predicted increased depressive symptoms via increased network efficiency even after controlling for previous levels of symptoms. Exploratory analyses showed that structural connectivity involving the left fronto-lateral and temporal regions were most strongly related to instability. Findings suggest that structural network efficiency relating to household instability may be a neural mechanism of risk for later depression and highlight the ways in which instability modulates neural development.

Network architecture dependent mechanical response in temperature responsive collagen-PNIPAM composites

Collagen is the most abundant protein in the mammalian extracellular matrix. In-vitro collagen-based materials with specific mechanical properties are important for various bio-medical and tissue-engineering applications. Here, we study the reversible mechanical switching behaviour of a bio-compatible composite formed by collagen networks seeded with thermo-responsive poly(N-isopropylacrylamide) (PNIPAM) microgel particles, by exploiting the swelling/de-swelling of the particles across the lower critical solution temperature (LCST). Interestingly, we find that the shear modulus of the system reversibly enhances whenever the diameter of the microgel particles is changed from that corresponding to the polymerization temperature of the composite, irrespective of swelling or, de-swelling. However, the degree of such enhancement significantly depends on the temperature-dependent collagen network architecture quantified by the mesh size of the network. Furthermore, confocal imaging of the composite during the temperature switching reveals that the reversible clustering of microgel particles above LCST plays a crucial role in the observed switching response.

Herbivore-induced pollinator limitation increases community stability of mutualism-antagonism continuum

Plants connect both pollinators and herbivores, and motivate the exploration of community structure in ecological networks merging antagonistic and mutualistic interactions. Evidence has shown that the two opposite plant-animal interactions are not independent from each other, in particular, herbivores can affect plant-pollinator pairwise interactions. Here, we explored effects of herbivore-induced pollinator limitation on community stability (including temporal stability and composition stability) of the mutualism-antagonism continuum. Our model demonstrated that pollinator limitation can boost up both temporal stability (i.e., the proportion of stable communities) and composition stability (i.e., species persistence), while the positive effects also depend on the strength of antagonistic and mutualistic interactions. Specifically, a community with higher temporal stability has a higher composition stability. Meanwhile, the correlations between network architecture and composition stability are also affected by pollinator limitation. Therefore, our results highlight that pollinator limitation can enhance community stability and may alter network architecture-composition stability relationship, and further advance the interplay between multiple types of species interactions within ecological networks.

Caenorhabditis elegans Connectomes of both Sexes as Image Classifiers

Connectome, the complete wiring diagram of the nervous system of an organism, is the biological substrate of the mind. While biological neural networks are crucial to the understanding of neural computation mechanisms, recent artificial neural networks (ANNs) have been developed independently from the study of real neural networks. Computational scientists are searching for various ANN architectures to improve machine learning since the architectures are associated with the accuracy of ANNs. A recent study used the hermaphrodite Caenorhabditis elegans (C. elegans) connectome for image classification tasks, where the edge directions were changed to construct a directed acyclic graph (DAG). In this study, we used the whole-animal connectomes of C. elegans hermaphrodite and male to construct a DAG that preserves the chief information flow in the connectomes and trained them for image classification of MNIST and fashion-MNIST datasets. The connectome-inspired neural networks exhibited over 99.5% and 92.6% of accuracy for MNIST and fashion-MNIST datasets, respectively, which increased from the previous study. Together, we conclude that realistic biological neural networks provide the basis of a plausible ANN architecture. This study suggests that biological networks can provide new inspiration to improve artificial intelligences (AIs).

Dataset on the small- and large deformation mechanical properties of emulsion-filled gelatin hydrogels as a model particle-filled composite food gel

In this article we present data related to the original research articles ‘Effect of matrix architecture on the elastic behavior of an emulsion-filled polymer gel’ (Gravelle et al., 2021) and ‘The influence of network architecture on the large deformation and fracture behavior of emulsion-filled gelatin gels’ (Gravelle and Marangoni, 2021). The small deformation elastic (Young's) modulus and large deformation fracture behavior of emulsion-filled composite gelatin gels are reported as a function of filler volume fraction (ϕ_f = 0 – 0.32). Homogeneous and heterogeneous network architectures were achieved by varying electrostatic interactions between matrix and filler. The effect of emulsion droplet physical state (solid fat or liquid oil) and gelator concentration (2, 4, 6, or 8% gelatin) were also evaluated. The reported elastic modulus, and fracture properties were obtained from large deformation uniaxial compression tests. Power law scaling behavior was identified for the elastic modulus as a function of both ϕ_f and gelator concentration, which are also reported. This data is relevant to the evaluation of network properties on the applicability of small deformation particle reinforcement theories and models describing the fracture mechanics of filled composites such as fat-filled food systems.

Deep Learning in Classification of Covid-19 Coronavirus, Pneumonia and Healthy Lungs on CXR and CT Images

Purpose

In this paper, the transfer learning method has been implemented to chest X-ray (CXR) and computed tomography (CT) bio-images of diverse kinds of lungs maladies, including CORONAVIRUS 2019 (COVID-19). COVID-19 identification is a difficult assignment that constantly demands a careful analysis of a patient’s clinical images, as COVID-19 is found to be very alike to pneumonic viral lung infection. In this paper, a transfer learning model to accelerate prediction processes and to assist medical professionals is proposed. Finally, the main purpose is to do an accurate classification between Covid-19, pneumonia and, healthy lungs using CXR and CT images.

Methods

Learning transfer gives the possibility to find out about this new illness COVID-19, using the knowledge we have about the pneumonia virus. This demonstrates the apprehensiveness achieved from a new architecture trained to detect virus-related pneumonia that must be transferred for COVID-19 detection. Transfer learning presents a considerable dissimilarity in results when compared to the result of traditional groupings. It is not necessary to create a separate model for the classification of COVID-19. This simplifies complicated issues by adopting the available model for COVID-19 determination. Automated diagnosis of COVID-19 using Haralick texture features is focused on segmented lung images and problematic lung patches. Lung patches are necessary for the augmentation of COVID-19 image data.

Results

The obtained outcomes are quite reliable for all distinctive processes as the proposed architecture can distinguish healthy lungs, pneumonia, COVID-19.

Conclusions

The results suggest that the implemented model is improved considering other existing models because the obtained classification accuracy is over the recently obtained results. It is a belief that the new architecture that is implemented in this study, delivers a petite step in building refined Coronavirus 2019 diagnosis architecture using CXR and CT bio-images.

[Technical aspects of quality assurance for intravitreal injections (IVI)]

BACKGROUND

Successful quality assurance in intravitreal injection (IVI) of medications requires a complex information technology infrastructure. The main challenges are data availability independent of location, standardization of clinical data, integration of extensive and currently non-standardized image documentation from coherence tomography and compliance with data protection regulations.

OBJECTIVE

In this article the technical implementation and data protection principles are reviewed.

MATERIAL AND METHODS

Essential aspects in the implementation of quality assurance in the field of IVI are discussed in a systematic approach.

RESULTS

In the field of network architectures web-based applications supplemented by local virtual private networks (VPN) and/or other software instances have recently replaced the previously commonly used physical data medium exchange. The standardization of the data, e.g. by converting the visual acuity into logMAR, plays an important role in the collection of treatment data. Multiple non-standardized data formats in optical coherence tomography complicate the general quality assurance structure and comparability of data.

CONCLUSION

International standards will probably facilitate this in the near future. Until then individual solutions have to be found on site.

Injectable synthetic hydrogel for bone regeneration: Physicochemical characterisation of a high and a low pH gelling system

Hybrid poly(ethylene glycol)-co-peptide hydrogels are a versatile platform for bone regeneration. For the use as injectable scaffolds, a good understanding of reaction kinetics and physical properties is vital. However, these factors have not yet been comprehensively illuminated. We show that gelation time can be effectively controlled by pH without affecting the elasticity of the formed hydrogels. Maleimide functionalised PEG gels at lower pH and produces more densely cross-linked networks than vinylsulfone functionalised PEG. Both form non-ideal networks. The elastic moduli on the order of a few kPa are in good agreement with the structural characterisation. Primary human osteoblasts cultured in proximity to bulk gels were not adversely affected in vitro. The results demonstrate that hybrid PEG-peptide hydrogels can be tailored to the requirements of in situ gelation. Attributed to their increased structural properties and a higher tolerance towards low pH, maleimide functionalised hydrogels might provide a better alternative for injectable applications.