Automatic identification and morphometry of optic nerve fibers in electron microscopy images

Ultrastructural Analysis of the Large Neuronal Perikarya in an Injured Dentate Nucleus Using an Experimental Model of Hyperthermia-Induced Convulsions: The First Qualitative and Quantitative Study

Background: Febrile seizures are a common form of convulsions in childhood, with poorly known cellular mechanisms. The objective of this pioneering study was to provide qualitative and quantitative ultrastructural research on the large neuronal perikarya in the cerebellar dentate nucleus (DN), using an experimental model of hyperthermia-induced seizures (HSs), comparable to febrile seizures in children. Methods: The study used young male Wistar rats, divided into experimental and control groups. The HSs were evoked by a hyperthermic water bath at 45 °C for 4 min for four consecutive days. Specimens (1 mm3) collected from the DN were routinely processed for transmission electron microscopy studies. Results: The ultrastructure of the large neurons in the DN affected by hyperthermic stress showed variously pronounced lesions in the perikarya, including total cell disintegration. The most pronounced neuronal lesions exhibited specific morphological signs of aponecrosis, i.e., dark cell degeneration ('dark neurons'). In close vicinity to the 'dark neurons', the aponecrotic bodies were found. The findings of this qualitative ultrastructural study correspond with the results of the morphometric analysis of the neuronal perikarya. Conclusions: Our results may constitute interesting comparative material for similar submicroscopic observations on large DN neurons in HS morphogenesis and, in the future, may help to find potential treatment targets to prevent febrile seizures or reduce recurrent seizures in children.

AimSeg: A machine-learning-aided tool for axon, inner tongue and myelin segmentation

Electron microscopy (EM) images of axons and their ensheathing myelin from both the central and peripheral nervous system are used for assessing myelin formation, degeneration (demyelination) and regeneration (remyelination). The g-ratio is the gold standard measure of assessing myelin thickness and quality, and traditionally is determined from measurements made manually from EM images-a time-consuming endeavour with limited reproducibility. These measurements have also historically neglected the innermost uncompacted myelin sheath, known as the inner tongue. Nonetheless, the inner tongue has been shown to be important for myelin growth and some studies have reported that certain conditions can elicit its enlargement. Ignoring this fact may bias the standard g-ratio analysis, whereas quantifying the uncompacted myelin has the potential to provide novel insights in the myelin field. In this regard, we have developed AimSeg, a bioimage analysis tool for axon, inner tongue and myelin segmentation. Aided by machine learning classifiers trained on transmission EM (TEM) images of tissue undergoing remyelination, AimSeg can be used either as an automated workflow or as a user-assisted segmentation tool. Validation results on TEM data from both healthy and remyelinating samples show good performance in segmenting all three fibre components, with the assisted segmentation showing the potential for further improvement with minimal user intervention. This results in a considerable reduction in time for analysis compared with manual annotation. AimSeg could also be used to build larger, high quality ground truth datasets to train novel deep learning models. Implemented in Fiji, AimSeg can use machine learning classifiers trained in ilastik. This, combined with a user-friendly interface and the ability to quantify uncompacted myelin, makes AimSeg a unique tool to assess myelin growth.

Enhanced Automatic Morphometry of Nerve Histological Sections Using Ensemble Learning

There is a need for an automated morphometry algorithm to facilitate the otherwise labor-intensive task of the quantitative histological analysis of neural microscopic images. A benchmark morphometry algorithm is the convolutional neural network Axondeepseg (ADS), which yields a high segmentation accuracy for scanning and transmission electron microscopy images. Nevertheless, it shows decreased accuracy when applied to optical microscopy images, and it has been observed to yield sizable false positives when identifying small-sized neurons within the slides. In this study, ensemble learning is used to enhance the performance of ADS by combining it with the paired image-to-image translation algorithm PairedImageTranslation (PIT). Here, 120 optical microscopy images of peripheral nerves were used to train and test the ensemble learning model and the two base models individually for comparison. The results showed weighted pixel-wise accuracy for the ensemble model of 95.5%, whereas the ADS and PIT yielded accuracies of 93.4% and 90%, respectively. The automated measurements of the axon diameters and myelin thicknesses from the manually marked ground truth images were not statistically different (p = 0.05) from the measurements taken from the same images when segmented using the developed ensemble model, while they were different when they were measured from the segmented images by the two base models individually. The automated measurement of the G ratios indicated a higher similarity to the ground truth testing images for the ensemble model in comparison with the individual base models. The proposed model yielded automated segmentation of the nerve slides, which were sufficiently equivalent to the manual annotations and could be employed for axon diameters and myelin thickness measurements for fully automated histological analysis of the neural images.

High-throughput segmentation of unmyelinated axons by deep learning

Axonal characterizations of connectomes in healthy and disease phenotypes are surprisingly incomplete and biased because unmyelinated axons, the most prevalent type of fibers in the nervous system, have largely been ignored as their quantitative assessment quickly becomes unmanageable as the number of axons increases. Herein, we introduce the first prototype of a high-throughput processing pipeline for automated segmentation of unmyelinated fibers. Our team has used transmission electron microscopy images of vagus and pelvic nerves in rats. All unmyelinated axons in these images are individually annotated and used as labeled data to train and validate a deep instance segmentation network. We investigate the effect of different training strategies on the overall segmentation accuracy of the network. We extensively validate the segmentation algorithm as a stand-alone segmentation tool as well as in an expert-in-the-loop hybrid segmentation setting with preliminary, albeit remarkably encouraging results. Our algorithm achieves an instance-level [Formula: see text] score of between 0.7 and 0.9 on various test images in the stand-alone mode and reduces expert annotation labor by 80% in the hybrid setting. We hope that this new high-throughput segmentation pipeline will enable quick and accurate characterization of unmyelinated fibers at scale and become instrumental in significantly advancing our understanding of connectomes in both the peripheral and the central nervous systems.

AxonDeepSeg: automatic axon and myelin segmentation from microscopy data using convolutional neural networks

Segmentation of axon and myelin from microscopy images of the nervous system provides useful quantitative information about the tissue microstructure, such as axon density and myelin thickness. This could be used for instance to document cell morphometry across species, or to validate novel non-invasive quantitative magnetic resonance imaging techniques. Most currently-available segmentation algorithms are based on standard image processing and usually require multiple processing steps and/or parameter tuning by the user to adapt to different modalities. Moreover, only a few methods are publicly available. We introduce AxonDeepSeg, an open-source software that performs axon and myelin segmentation of microscopic images using deep learning. AxonDeepSeg features: (i) a convolutional neural network architecture; (ii) an easy training procedure to generate new models based on manually-labelled data and (iii) two ready-to-use models trained from scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Results show high pixel-wise accuracy across various species: 85% on rat SEM, 81% on human SEM, 95% on mice TEM and 84% on macaque TEM. Segmentation of a full rat spinal cord slice is computed and morphological metrics are extracted and compared against the literature. AxonDeepSeg is freely available at https://github.com/neuropoly/axondeepseg .

A New Method for Automated Identification and Morphometry of Myelinated Fibers Through Light Microscopy Image Analysis

Nerve morphometry is known to produce relevant information for the evaluation of several phenomena, such as nerve repair, regeneration, implant, transplant, aging, and different human neuropathies. Manual morphometry is laborious, tedious, time consuming, and subject to many sources of error. Therefore, in this paper, we propose a new method for the automated morphometry of myelinated fibers in cross-section light microscopy images. Images from the recurrent laryngeal nerve of adult rats and the vestibulocochlear nerve of adult guinea pigs were used herein. The proposed pipeline for fiber segmentation is based on the techniques of competitive clustering and concavity analysis. The evaluation of the proposed method for segmentation of images was done by comparing the automatic segmentation with the manual segmentation. To further evaluate the proposed method considering morphometric features extracted from the segmented images, the distributions of these features were tested for statistical significant difference. The method achieved a high overall sensitivity and very low false-positive rates per image. We detect no statistical difference between the distribution of the features extracted from the manual and the pipeline segmentations. The method presented a good overall performance, showing widespread potential in experimental and clinical settings allowing large-scale image analysis and, thus, leading to more reliable results.

AxonSeg: Open Source Software for Axon and Myelin Segmentation and Morphometric Analysis

Segmenting axon and myelin from microscopic images is relevant for studying the peripheral and central nervous system and for validating new MRI techniques that aim at quantifying tissue microstructure. While several software packages have been proposed, their interface is sometimes limited and/or they are designed to work with a specific modality (e.g., scanning electron microscopy (SEM) only). Here we introduce AxonSeg, which allows to perform automatic axon and myelin segmentation on histology images, and to extract relevant morphometric information, such as axon diameter distribution, axon density and the myelin g-ratio. AxonSeg includes a simple and intuitive MATLAB-based graphical user interface (GUI) and can easily be adapted to a variety of imaging modalities. The main steps of AxonSeg consist of: (i) image pre-processing; (ii) pre-segmentation of axons over a cropped image and discriminant analysis (DA) to select the best parameters based on axon shape and intensity information; (iii) automatic axon and myelin segmentation over the full image; and (iv) atlas-based statistics to extract morphometric information. Segmentation results from standard optical microscopy (OM), SEM and coherent anti-Stokes Raman scattering (CARS) microscopy are presented, along with validation against manual segmentations. Being fully-automatic after a quick manual intervention on a cropped image, we believe AxonSeg will be useful to researchers interested in large throughput histology. AxonSeg is open source and freely available at: https://github.com/neuropoly/axonseg.

Automated method for the segmentation and morphometry of nerve fibers in large-scale CARS images of spinal cord tissue

A fully automated method for large-scale segmentation of nerve fibers from coherent anti-Stokes Raman scattering (CARS) microscopy images is presented. The method is specifically designed for CARS images of transverse cross sections of nervous tissue but is also suitable for use with standard light microscopy images. After a detailed description of the two-part segmentation algorithm, its accuracy is quantified by comparing the resulting binary images to manually segmented images. We then demonstrate the ability of our method to retrieve morphological data from CARS images of nerve tissue. Finally, we present the segmentation of a large mosaic of CARS images covering more than half the area of a mouse spinal cord cross section and show evidence of clusters of neurons with similar g-ratios throughout the spinal cord.

AOTF based molecular hyperspectral imaging system and its applications on nerve morphometry

The neuroanatomical morphology of nerve fibers is an important description for understanding the pathological aspects of nerves. Different from the traditional automatic nerve morphometry methods, a molecular hyperspectral imaging system based on an acousto-optic tunable filter (AOTF) was developed and used to identify unstained nerve histological sections. The hardware, software, and system performance of the imaging system are presented and discussed. The gray correction coefficient was used to calibrate the system's spectral response and to remove the effects of noises and artifacts. A spatial-spectral kernel-based approach through the support vector machine formulation was proposed to identify nerve fibers. This algorithm can jointly use both the spatial and spectral information of molecular hyperspectral images for segmentation. Then, the morphological parameters such as fiber diameter, axon diameter, myelin sheath thickness, fiber area, and g-ratio were calculated and evaluated. Experimental results show that the hyperspectral-based method has the potential to recognize and measure the nerve fiber more accurately than traditional methods.

Automatic identification and quantitative morphometry of unstained spinal nerve using molecular hyperspectral imaging technology

Quantitative observation of nerve fiber sections is often complemented by morphological analysis in both research and clinical condition. However, existing manual or semi-automated methods are tedious and labour intensive, fully automated morphometry methods are complicated as the information of color or gray images captured by traditional microscopy is limited. Moreover, most of the methods are time-consuming as the nerve sections need to be stained with some reagents before observation. To overcome these shortcomings, a molecular hyperspectral imaging system is developed and used to observe the spinal nerve sections. The molecular hyperspectral images contain both the structural and biochemical information of spinal nerve sections which is very useful for automatic identification and quantitative morphological analysis of nerve fibers. This characteristic makes it possible for researchers to observe the unstained spinal nerve and live cells in their native environment. To evaluate the performance of the new method, the molecular hyperspectral images were captured and the improved spectral angle mapper algorithm was proposed and used to segment the myelin contours. Then the morphological parameters such as myelin thickness and myelin area were calculated and evaluated. With these morphological parameters, the three dimension surface view images were drawn to help the investigators observe spinal nerve at different angles. The experiment results show that the hyperspectral based method has the potential to identify the spinal nerve more accurate than the traditional method as the new method contains both the spectral and spatial information of nerve sections.

Visual system scaling in teleost fish

Teleost fish grow continuously throughout their lifespan, and this growth includes visual system components: eyes, optic nerves, and brain. As fish grow, the optic nerve lengthens and neural signals must travel increasing distances from the eye to the optic tectum along thousands of retinal ganglion cell (RGC) axons. Larger fish have better vision that enhances their ability to capture prey, but they are faced with the potential computational problem of changes in the relative timing of visual information arriving at the brain. Optic nerve conduction delays depend on RGC axon conduction velocities, and velocity is primarily determined by axon diameters. If axon diameters do not increase in proportion to body length, then absolute and relative conduction delays will vary with fish size. We have measured optic nerve lengths and axon diameter distributions in different sized zebrafish (Danio rerio) and goldfish (Carassius auratus) and find that, as both species of fish grow, axon diameters increase to reduce average conduction delays by about half and to keep relative delays constant. This invariance of relative conduction delays simplifies computational problems faced by the optic tectum.

A semi-automated method for identifying and measuring myelinated nerve fibers in scanning electron microscope images

Diagnosing illnesses, developing and comparing treatment methods, and conducting research on the organization of the peripheral nervous system often require the analysis of peripheral nerve images to quantify the number, myelination, and size of axons in a nerve. Current methods that require manually labeling each axon can be extremely time-consuming as a single nerve can contain thousands of axons. To improve efficiency, we developed a computer-assisted axon identification and analysis method that is capable of analyzing and measuring sub-images covering the nerve cross-section, acquired using a scanning electron microscope. This algorithm performs three main procedures - it first uses cross-correlation to combine the acquired sub-images into a large image showing the entire nerve cross-section, then identifies and individually labels axons using a series of image intensity and shape criteria, and finally identifies and labels the myelin sheath of each axon using a region growing algorithm with the geometric centers of axons as seeds. To ensure accurate analysis of the image, we incorporated manual supervision to remove mislabeled axons and add missed axons. The typical user-assisted processing time for a two-megapixel image containing over 2000 axons was less than 1h. This speed was almost eight times faster than the time required to manually process the same image. Our method has proven to be well suited for identifying axons and their characteristics, and represents a significant time savings over traditional manual methods.