Novel Methods for Microglia Segmentation, Feature Extraction, and Classification

Progressive Feature Fusion Attention Dense Network for Speckle Noise Removal in OCT Images

Although deep learning for Big Data analytics has achieved promising results in the field of optical coherence tomography (OCT) image denoising, the low recognition rate caused by complex noise distribution and a large number of redundant features is still a challenge faced by deep learning-based denoising methods. Moreover, the network with large depth will bring high computational complexity. To this end, we propose a progressive feature fusion attention dense network (PFFADN) for speckle noise removal in OCT images. We arrange densely connected dense blocks in the deep convolution network, and sequentially connect the shallow convolution feature map with the deep one extracted from each dense block to form a residual block. We add attention mechanism to the network to extract the key features and suppress the irrelevant ones. We fuse the output feature maps from all dense blocks and input them to the reconstruction output layer. We compare PFFADN with the state-of-the-art denoising algorithms on retinal OCT images. Experiments show that our method has better improvement in denoising performance.

Morphology and Fractal-Based Classifications of Neurons and Microglia in Two and Three Dimensions

Microglia and neurons live physically intertwined, intimately related structurally and functionally in a dynamic relationship in which microglia change continuously over a much shorter timescale than do neurons. Although microglia may unwind and depart from the neurons they attend under certain circumstances, in general, together both contribute to the fractal topology of the brain that defines its computational capabilities. Both neuronal and microglial morphologies are well-described using fractal analysis complementary to more traditional measures. For neurons, the fractal dimension has proved valuable for classifying dendritic branching and other neuronal features relevant to pathology and development. For microglia, fractal geometry has substantially contributed to classifying functional categories, where, in general, the more pathological the biological status, the lower the fractal dimension for individual cells, with some exceptions, including hyper-ramification. This chapter provides a review of the intimate relationships between neurons and microglia, by introducing 2D and 3D fractal analysis methodology and its applications in neuron-microglia function in health and disease.

Thymosin beta 4 prevents systemic lipopolysaccharide-induced plaque load in middle-age APP/PS1 mice

Lipopolysaccharide (LPS) produced by the gut during systemic infections and inflammation is thought to contribute to Alzheimer's disease (AD) progression. Since thymosin beta 4 (Tβ4) effectively reduces LPS-induced inflammation in sepsis, we tested its potential to alleviate the impact of LPS in the brain of the APPswePS1dE9 mouse model of AD (APP/PS1) and wildtype (WT) mice. 12.5-month-old male APP/PS1 mice (n = 30) and their WT littermates (n = 29) were tested for baseline food burrowing performance, spatial working memory and exploratory drive in the spontaneous alternation and open-field tests, prior to being challenged with LPS (100ug/kg, i.v.) or its vehicle phosphate buffered saline (PBS). Tβ4 (5 mg/kg, i.v.) or PBS, was administered immediately following and at 2 and 4 h after the PBS or LPS challenge, and then once daily for 6 days (n = 7-8). LPS-induced sickness was assessed though monitoring of changes in body weight and behaviour over a 7-day period. Brains were collected for the determination of amyloid plaque load and reactive gliosis in the hippocampus and cortex. Treatment with Tβ4 alleviated sickness symptoms to a greater extent in APP/PS1 than in WT mice by limiting LPS-induced weight loss and inhibition of food burrowing behaviour. It prevented LPS-induced amyloid burden in APP/PS1 mice but increased astrocytic and microglial proliferation in the hippocampus of LPS-treated WT mice. These data show that Tβ4 can alleviate the adverse effects of systemic LPS in the brain by preventing exacerbation of amyloid deposition in AD mice and by inducing reactive microgliosis in aging WT mice.

Reduced and stable feature sets selection with random forest for neurons segmentation in histological images of macaque brain

In preclinical research, histology images are produced using powerful optical microscopes to digitize entire sections at cell scale. Quantification of stained tissue relies on machine learning driven segmentation. However, such methods require multiple additional information, or features, which are increasing the quantity of data to process. As a result, the quantity of features to deal with represents a drawback to process large series or massive histological images rapidly in a robust manner. Existing feature selection methods can reduce the amount of required information but the selected subsets lack reproducibility. We propose a novel methodology operating on high performance computing (HPC) infrastructures and aiming at finding small and stable sets of features for fast and robust segmentation of high-resolution histological images. This selection has two steps: (1) selection at features families scale (an intermediate pool of features, between spaces and individual features) and (2) feature selection performed on pre-selected features families. We show that the selected sets of features are stables for two different neuron staining. In order to test different configurations, one of these dataset is a mono-subject dataset and the other is a multi-subjects dataset to test different configurations. Furthermore, the feature selection results in a significant reduction of computation time and memory cost. This methodology will allow exhaustive histological studies at a high-resolution scale on HPC infrastructures for both preclinical and clinical research.

Brain Immunoinformatics: A Symmetrical Link between Informatics, Wet Lab and the Clinic

Breakthrough advances in informatics over the last decade have thoroughly influenced the field of immunology. The intermingling of machine learning with wet lab applications and clinical results has hatched the newly defined immunoinformatics society. Immunoinformatics of the central neural system, referred to as neuroimmunoinformatics (NII), investigates symmetrical and asymmetrical interactions of the brain-immune interface. This interdisciplinary overview on NII is addressed to bioscientists and computer scientists. We delineate the dominating trajectories and field-shaping achievements and elaborate on future directions using bridging language and terminology. Computation, varying from linear modeling to complex deep learning approaches, fuels neuroimmunology through three core directions. Firstly, by providing big-data analysis software for high-throughput methods such as next-generation sequencing and genome-wide association studies. Secondly, by designing models for the prediction of protein morphology, functions, and symmetrical and asymmetrical protein–protein interactions. Finally, NII boosts the output of quantitative pathology by enabling the automatization of tedious processes such as cell counting, tracing, and arbor analysis. The new classification of microglia, the brain’s innate immune cells, was an NII achievement. Deep sequencing classifies microglia in “sensotypes” to accurately describe the versatility of immune responses to physiological and pathological challenges, as well as to experimental conditions such as xenografting and organoids. NII approaches complex tasks in the brain-immune interface, recognizes patterns and allows for hypothesis-free predictions with ultimate targeted individualized treatment strategies, and personalizes disease prognosis and treatment response.

Classification of Microglial Morphological Phenotypes Using Machine Learning

Microglia are the brain's immunocompetent macrophages with a unique feature that allows surveillance of the surrounding microenvironment and subsequent reactions to tissue damage, infection, or homeostatic perturbations. Thereby, microglia's striking morphological plasticity is one of their prominent characteristics and the categorization of microglial cell function based on morphology is well established. Frequently, automated classification of microglial morphological phenotypes is performed by using quantitative parameters. As this process is typically limited to a few and especially manually chosen criteria, a relevant selection bias may compromise the resulting classifications. In our study, we describe a novel microglial classification method by morphological evaluation using a convolutional neuronal network on the basis of manually selected cells in addition to classical morphological parameters. We focused on four microglial morphologies, ramified, rod-like, activated and amoeboid microglia within the murine hippocampus and cortex. The developed method for the classification was confirmed in a mouse model of ischemic stroke which is already known to result in microglial activation within affected brain regions. In conclusion, our classification of microglial morphological phenotypes using machine learning can serve as a time-saving and objective method for post-mortem characterization of microglial changes in healthy and disease mouse models, and might also represent a useful tool for human brain autopsy samples.

Automated segmentation and analysis of retinal microglia within ImageJ

Microglia are immune cells of the central nervous system capable of distinct phenotypic changes and migration in response to injury. These changes most notably include the retraction of fine dendritic structures and adoption of a globular, phagocytic morphology. Due to their characteristic responses, microglia frequently act as histological indicators of injury progression. While algorithms seeking to automate microglia counts and morphological analysis are becoming increasingly popular, few exist that are adequate for use within the retina and manual analysis remains prevalent. To address this, we propose a novel segmentation routine, implemented within FIJI-ImageJ, to perform automated segmentation and cell counting of retinal microglia. We show that our routine could perform cell counts with accuracy similar to manual observers using the I307N Rho model. Tracking cell position relative to retinal vasculature, we observed population migration towards the photoreceptor layer beginning 12 h post light damage. Using feature selection with Chi² and principal component analysis, we resolved cells along a morphological gradient, demonstrating that extracted features were sufficiently descriptive to capture subtle morphological changes within cell populations in I307N Rho and Balb/c TLR2^-/- retinal degeneration models. Taken together, we introduce a novel automated routine capable of efficient image processing and segmentation. Using data retrieved following segmentation, we perform morphological analysis simultaneously on whole populations of cells, rather than individually. Our algorithm was built entirely with open-source software, for use on retinal microglia.

From Local to Global: A Graph Framework for Retinal Artery/Vein Classification

Fundus photography has been widely used for inspecting eye disorders by ophthalmologists or computer algorithms. Biomarkers related to retinal vessels plays an essential role to detect early diabetes. To quantify vascular biomarkers or the corresponding changes, an accurate artery and vein classification is necessary. In this work, we propose a new framework to boost local vessel classification with a global vascular network model using graph convolution. We compare our proposed method with two traditional state-of-the-art methods on a testing dataset of 750 images from the Maastricht Study. After incorporating global information, our model achieves the best accuracy of 86.45% compared to 85.5% from convolutional neural networks (CNN) and 82.9% from handcrafted pixel feature classification (HPFC). Our model also obtains the best area under receiver operating characteristic curve (AUC) of 0.95, compared to 0.93 from CNN and 0.90 from HPFC. The new classification framework has the advantage of easy deployment on top of local classification features. It corrects the local classification error by minimizing global classification error and it brings free additional classification performance.

An End-to-end System for Automatic Characterization of Iba1 Immunopositive Microglia in Whole Slide Imaging

Traumatic brain injury (TBI) is one of the leading causes of death and disability worldwide. Detailed studies of the microglial response after TBI require high throughput quantification of changes in microglial count and morphology in histological sections throughout the brain. In this paper, we present a fully automated end-to-end system that is capable of assessing microglial activation in white matter regions on whole slide images of Iba1 stained sections. Our approach involves the division of the full brain slides into smaller image patches that are subsequently automatically classified into white and grey matter sections. On the patches classified as white matter, we jointly apply functional minimization methods and deep learning classification to identify Iba1-immunopositive microglia. Detected cells are then automatically traced to preserve their complex branching structure after which fractal analysis is applied to determine the activation states of the cells. The resulting system detects white matter regions with 84% accuracy, detects microglia with a performance level of 0.70 (F1 score, the harmonic mean of precision and sensitivity) and performs binary microglia morphology classification with a 70% accuracy. This automated pipeline performs these analyses at a 20-fold increase in speed when compared to a human pathologist. Moreover, we have demonstrated robustness to variations in stain intensity common for Iba1 immunostaining. A preliminary analysis was conducted that indicated that this pipeline can identify differences in microglia response due to TBI. An automated solution to microglia cell analysis can greatly increase standardized analysis of brain slides, allowing pathologists and neuroscientists to focus on characterizing the associated underlying diseases and injuries.

Sex-specific hippocampal metabolic signatures at the onset of systemic inflammation with lipopolysaccharide in the APPswe/PS1dE9 mouse model of Alzheimer's disease

Systemic inflammation enhances the risk and progression of Alzheimer's disease (AD). Lipopolysaccharide (LPS), a potent pro-inflammatory endotoxin produced by the gut, is found in excess levels in AD where it associates with neurological hallmarks of pathology. Sex differences in susceptibility to inflammation and AD progression have been reported, but how this impacts on LPS responses remains under investigated. We previously reported in an APP/PS1 model of AD that systemic LPS administration rapidly altered hippocampal metabolism in males. Here, we used untargeted metabolomics to comprehensively identify hippocampal metabolic processes occurring at onset of systemic inflammation with LPS (100 µg/kg, i.v.) in APP/PS1 mice, at an early pathological stage, and investigated the sexual dimorphism in this response. Four hours after LPS administration, pathways regulating energy metabolism, immune and oxidative stress responses were simultaneously recruited in the hippocampi of 4.5-month-old mice with a more protective response in females despite their pro-inflammatory and pro-oxidant metabolic signature in the absence of immune stimulation. LPS induced comparable behavioural sickness responses in male and female wild-type and APP/PS1 mice and comparable activation of both the serotonin and nicotinamide pathways of tryptophan metabolism in their hippocampi. Elevations in N-methyl-2-pyridone-5-carboxamide, a major toxic metabolite of nicotinamide, correlated with behavioural sickness regardless of sex, as well as with the LPS-induced hypothermia seen in males. Males also exhibited a pro-inflammatory-like downregulation of pyruvate metabolism, exacerbated in APP/PS1 males, and methionine metabolism whereas females showed a greater cytokine response and anti-inflammatory-like downregulation of hippocampal methylglyoxal and methionine metabolism. Metabolic changes were not associated with morphological markers of immune cell activation suggesting that they constitute an early event in the development of LPS-induced neuroinflammation and AD exacerbation. These data suggest that the female hippocampus is more tolerant to acute systemic inflammation.

Segmentation, Tracing, and Quantification of Microglial Cells from 3D Image Stacks

Microglia play a central role in modulating synaptic structure and physiology, learning and memory processes. They exhibit morphological changes to perform these roles, therefore the morphological study of microglia can help to understand their functionality. Many promising methods are proposed to automatically segment the blood vessels or reconstruct the neuronal morphology. However, they often fail to accurately capture microglia organizations due to the striking structural differences. This requires a more sophisticated approach of reconstruction taking into account the varying nature of branch structures and soma sizes. To this end, we propose an automated method to reconstruct microglia, and quantify their features from 2D/3D image datasets. We first employ multilevel thresholding to segment soma volumes(3D)/areas(2D) and recognize foreground voxels/pixels. Seed points sampled from the foreground, are connected to form the skeleton of the branches via the tracing process. The reconstructed data is quantified and written in SWC standard file format. We have applied our method to 3D image datasets of microglia, then evaluated the results using ground truth data, and compared them to those achieved via the state-of-the-art methods. Our method outperforms the others both in accuracy and computational time.

T4SS Effector Protein Prediction with Deep Learning

Extensive research has been carried out on bacterial secretion systems, as they can pass effector proteins directly into the cytoplasm of host cells. The correct prediction of type IV protein effectors secreted by T4SS is important, since they are known to play a noteworthy role in various human pathogens. Studies on predicting T4SS effectors involve traditional machine learning algorithms. In this work we included a deep learning architecture, i.e., a Convolutional Neural Network (CNN), to predict IVA and IVB effectors. Three feature extraction methods were utilized to represent each protein as an image and these images fed the CNN as inputs in our proposed framework. Pseudo proteins were generated using ADASYN algorithm to overcome the imbalanced dataset problem. We demonstrated that our framework predicted all IVA effectors correctly. In addition, the sensitivity performance of 94.2% for IVB effector prediction exhibited our framework’s ability to discern the effectors in unidentified proteins.

Myoinositol CEST signal in animals with increased Iba-1 levels in response to an inflammatory challenge-Preliminary findings

Neuroinflammation plays an important role in the pathogenesis of a range of brain disorders. Non-invasive imaging of neuroinflammation is critical to help improve our understanding of the underlying disease mechanisms, monitor therapies and guide drug development. Generally, MRI lacks specificity to molecular imaging biomarkers, but molecular MR imaging based on chemical exchange saturation transfer (CEST) can potentially detect changes of myoinositol, a putative glial marker that may index neuroinflammation. In this pilot study we aimed to investigate, through validation with immunohistochemistry and in vivo magnetic resonance spectroscopy (MRS), whether CEST imaging can reflect the microglial response to a mild inflammatory challenge with lipopolysaccharide (LPS), in the APPSwe/ PS1 mouse model of Alzheimer’s disease and wild type controls. The response to the immune challenge was variable and did not align with genotype. Animals with a strong response to LPS (Iba1+, n = 6) showed an increase in CEST contrast compared with those who did not (Iba1-, n = 6). Changes of myoinositol levels after LPS were not significant. We discuss the difficulties of this mild inflammatory model, the role of myoinositol as a glial biomarker, and the technical challenges of CEST imaging at 0.6ppm.

pyHIVE, a health-related image visualization and engineering system using Python

Background

Imaging is one of the major biomedical technologies to investigate the status of a living object. But the biomedical image based data mining problem requires extensive knowledge across multiple disciplinaries, e.g. biology, mathematics and computer science, etc.

Results

pyHIVE (a Health-related Image Visualization and Engineering system using Python) was implemented as an image processing system, providing five widely used image feature engineering algorithms. A standard binary classification pipeline was also provided to help researchers build data models immediately after the data is collected. pyHIVE may calculate five widely-used image feature engineering algorithms efficiently using multiple computing cores, and also featured the modules of Principal Component Analysis (PCA) based preprocessing and normalization.

Conclusions

The demonstrative example shows that the image features generated by pyHIVE achieved very good classification performances based on the gastrointestinal endoscopic images. This system pyHIVE and the demonstrative example are freely available and maintained at http://www.healthinformaticslab.org/supp/resources.php.

Automated Morphological Analysis of Microglia After Stroke

Microglia are the resident immune cells of the brain and react quickly to changes in their environment with transcriptional regulation and morphological changes. Brain tissue injury such as ischemic stroke induces a local inflammatory response encompassing microglial activation. The change in activation status of a microglia is reflected in its gradual morphological transformation from a highly ramified into a less ramified or amoeboid cell shape. For this reason, the morphological changes of microglia are widely utilized to quantify microglial activation and studying their involvement in virtually all brain diseases. However, the currently available methods, which are mainly based on manual rating of immunofluorescent microscopic images, are often inaccurate, rater biased, and highly time consuming. To address these issues, we created a fully automated image analysis tool, which enables the analysis of microglia morphology from a confocal Z-stack and providing up to 59 morphological features. We developed the algorithm on an exploratory dataset of microglial cells from a stroke mouse model and validated the findings on an independent data set. In both datasets, we could demonstrate the ability of the algorithm to sensitively discriminate between the microglia morphology in the peri-infarct and the contralateral, unaffected cortex. Dimensionality reduction by principal component analysis allowed to generate a highly sensitive compound score for microglial shape analysis. Finally, we tested for concordance of results between the novel automated analysis tool and the conventional manual analysis and found a high degree of correlation. In conclusion, our novel method for the fully automatized analysis of microglia morphology shows excellent accuracy and time efficacy compared to traditional analysis methods. This tool, which we make openly available, could find application to study microglia morphology using fluorescence imaging in a wide range of brain disease models.

Dynamic metabolic patterns tracking neurodegeneration and gliosis following 26S proteasome dysfunction in mouse forebrain neurons

Metabolite profiling is an important tool that may better capture the multiple features of neurodegeneration. With the considerable parallels between mouse and human metabolism, the use of metabolomics in mouse models with neurodegenerative pathology provides mechanistic insight and ready translation into aspects of human disease. Using 400 MHz nuclear magnetic resonance spectroscopy we have carried out a temporal region-specific investigation of the metabolome of neuron-specific 26S proteasome knockout mice characterised by progressive neurodegeneration and Lewy-like inclusion formation in the forebrain. An early significant decrease in N-acetyl aspartate revealed evidence of neuronal dysfunction before cell death that may be associated with changes in brain neuroenergetics, underpinning the use of this metabolite to track neuronal health. Importantly, we show early and extensive activation of astrocytes and microglia in response to targeted neuronal dysfunction in this context, but only late changes in myo-inositol; the best established glial cell marker in magnetic resonance spectroscopy studies, supporting recent evidence that additional early neuroinflammatory markers are needed. Our results extend the limited understanding of metabolite changes associated with gliosis and provide evidence that changes in glutamate homeostasis and lactate may correlate with astrocyte activation and have biomarker potential for tracking neuroinflammation.

A portable image-based cytometer for rapid malaria detection and quantification

Increasing resistance by malaria parasites to currently used antimalarials across the developing world warrants timely detection and classification so that appropriate drug combinations can be administered before clinical complications arise. However, this is often challenged by low levels of infection (referred to as parasitemia) and presence of predominantly young parasitic forms in the patients' peripheral blood. Herein, we developed a simple, inexpensive and portable image-based cytometer that detects and numerically counts Plasmodium falciparum infected red blood cells (iRBCs) from Giemsa-stained smears derived from infected blood. Our cytometer is able to classify all parasitic subpopulations by quantifying the area occupied by the parasites within iRBCs, with high specificity, sensitivity and negligible false positives (~ 0.0025%). Moreover, we demonstrate the application of our image-based cytometer in testing anti-malarial efficacy against a commercial flow cytometer and demonstrate comparable results between the two methods. Collectively, these results highlight the possibility to use our image-based cytometer as a cheap, rapid and accurate alternative for antimalarial testing without compromising on efficiency and minimal processing time. With appropriate filters applied into the algorithm, to rule out leukocytes and reticulocytes, our cytometer may also be used for field diagnosis of malaria.