Mu-Net a Light Architecture for Small Dataset Segmentation of Brain Organoid Bright-Field Images.

To characterize the growth of brain organoids (BOs), cultures that replicate some early physiological or pathological developments of the human brain are usually manually extracted. Due to their novelty, only small datasets of these images are available, but segmenting the organoid shape automatically with deep learning (DL) tools requires a larger number of images. Light U-Net segmentation architectures, which reduce the training time while increasing the sensitivity under small input datasets, have recently emerged. We further reduce the U-Net architecture and compare the proposed architecture (MU-Net) with U-Net and UNet-Mini on bright-field images of BOs using several data augmentation strategies. In each case, we perform leave-one-out cross-validation on 40 original and 40 synthesized images with an optimized adversarial autoencoder (AAE) or on 40 transformed images. The best results are achieved with U-Net segmentation trained on optimized augmentation. However, our novel method, MU-Net, is more robust: it achieves nearly as accurate segmentation results regardless of the dataset used for training (various AAEs or a transformation augmentation). In this study, we confirm that small datasets of BOs can be segmented with a light U-Net method almost as accurately as with the original method.

Brain organoid data synthesis and evaluation.

Introduction: Datasets containing only few images are common in the biomedical field. This poses a global challenge for the development of robust deep-learning analysis tools, which require a large number of images. Generative Adversarial Networks (GANs) are an increasingly used solution to expand small datasets, specifically in the biomedical domain. However, the validation of synthetic images by metrics is still controversial and psychovisual evaluations are time consuming. Methods: We augment a small brain organoid bright-field database of 40 images using several GAN optimizations. We compare these synthetic images to the original dataset using similitude metrcis and we perform an psychovisual evaluation of the 240 images generated. Eight biological experts labeled the full dataset (280 images) as syntetic or natural using a custom-built software. We calculate the error rate per loss optimization as well as the hesitation time. We then compare these results to those provided by the similarity metrics. We test the psychovalidated images in a training step of a segmentation task. Results and discussion: Generated images are considered as natural as the original dataset, with no increase of the hesitation time by experts. Experts are particularly misled by perceptual and Wasserstein loss optimization. These optimizations render the most qualitative and similar images according to metrics to the original dataset. We do not observe a strong correlation but links between some metrics and psychovisual decision according to the kind of generation. Particular Blur metric combinations could maybe replace the psychovisual evaluation. Segmentation task which use the most psychovalidated images are the most accurate.

giRAff: an automated atlas segmentation tool adapted to single histological slices.

Conventional histology of the brain remains the gold standard in the analysis of animal models. In most biological studies, standard protocols usually involve producing a limited number of histological slices to be analyzed. These slices are often selected into a specific anatomical region of interest or around a specific pathological lesion. Due to the lack of automated solutions to analyze such single slices, neurobiologists perform the segmentation of anatomical regions manually most of the time. Because the task is long, tedious, and operator-dependent, we propose an automated atlas segmentation method called giRAff, which combines rigid and affine registrations and is suitable for conventional histological protocols involving any number of single slices from a given mouse brain. In particular, the method has been tested on several routine experimental protocols involving different anatomical regions of different sizes and for several brains. For a given set of single slices, the method can automatically identify the corresponding slices in the mouse Allen atlas template with good accuracy and segmentations comparable to those of an expert. This versatile and generic method allows the segmentation of any single slice without additional anatomical context in about 1 min. Basically, our proposed giRAff method is an easy-to-use, rapid, and automated atlas segmentation tool compliant with a wide variety of standard histological protocols.

A general deep learning framework for neuron instance segmentation based on Efficient UNet and morphological post-processing.

Recent studies have demonstrated the superiority of deep learning in medical image analysis, especially in cell instance segmentation, a fundamental step for many biological studies. However, the excellent performance of the neural networks requires training on large, unbiased dataset and annotations, which is labor-intensive and expertise-demanding. This paper presents an end-to-end framework to automatically detect and segment NeuN stained neuronal cells on histological images using only point annotations. Unlike traditional nuclei segmentation with point annotation, we propose using point annotation and binary segmentation to synthesize pixel-level annotations. The synthetic masks are used as the ground truth to train the neural network, a U-Net-like architecture with a state-of-the-art network, EfficientNet, as the encoder. Validation results show the superiority of our model compared to other recent methods. In addition, we investigated multiple post-processing schemes and proposed an original strategy to convert the probability map into segmented instances using ultimate erosion and dynamic reconstruction. This approach is easy to configure and outperforms other classical post-processing techniques. This work aims to develop a robust and efficient framework for analyzing neurons using optical microscopic data, which can be used in preclinical biological studies and, more specifically, in the context of neurodegenerative diseases. Code is available at: https://github.com/MIRCen/NeuronInstanceSeg.

Automated Atlas-based Segmentation of Single Coronal Mouse Brain Slices using Linear 2D-2D Registration.

A significant challenge for brain histological data analysis is to precisely identify anatomical regions in order to perform accurate local quantifications and evaluate therapeutic solutions. Usually, this task is performed manually, becoming therefore tedious and subjective. Another option is to use automatic or semi-automatic methods, among which segmentation using digital atlases co-registration. However, most available atlases are 3D, whereas digitized histological data are 2D. Methods to perform such 2D-3D segmentation from an atlas are required. This paper proposes a strategy to automatically and accurately segment single 2D coronal slices within a 3D volume of atlas, using linear registration. We validated its robustness and performance using an exploratory approach at whole-brain scale.

Evaluation of Deep Learning Topcoders Method for Neuron Individualization in Histological Macaque Brain Section.

Cell individualization has a vital role in digital pathology image analysis. Deep Learning is considered as an efficient tool for instance segmentation tasks, including cell individualization. However, the precision of the Deep Learning model relies on massive unbiased dataset and manual pixel-level annotations, which is labor intensive. Moreover, most applications of Deep Learning have been developed for processing oncological data. To overcome these challenges, i) we established a pipeline to synthesize pixel-level labels with only point annotations provided; ii) we tested an ensemble Deep Learning algorithm to perform cell individualization on neurological data. Results suggest that the proposed method successfully segments neuronal cells in both object-level and pixel-level, with an average detection accuracy of 0.93.

Recent Trends and Perspectives in Cerebral Organoids Imaging and Analysis.

Purpose: Since their first generation in 2013, the use of cerebral organoids has spread exponentially. Today, the amount of generated data is becoming challenging to analyze manually. This review aims to overview the current image acquisition methods and to subsequently identify the needs in image analysis tools for cerebral organoids. Methods: To address this question, we went through all recent articles published on the subject and annotated the protocols, acquisition methods, and algorithms used. Results: Over the investigated period of time, confocal microscopy and bright-field microscopy were the most used acquisition techniques. Cell counting, the most common task, is performed in 20% of the articles and area; around 12% of articles calculate morphological parameters. Image analysis on cerebral organoids is performed in majority using ImageJ software (around 52%) and Matlab language (4%). Treatments remain mostly semi-automatic. We highlight the limitations encountered in image analysis in the cerebral organoid field and suggest possible solutions and implementations to develop. Conclusions: In addition to providing an overview of cerebral organoids cultures and imaging, this work highlights the need to improve the existing image analysis methods for such images and the need for specific analysis tools. These solutions could specifically help to monitor the growth of future standardized cerebral organoids.

Construction of a deformable spatiotemporal MRI atlas of the fetal brain: evaluation of similarity metrics and deformation models.

The development and identification of best methods in fetal brain MRI analysis is crucial as we expect an outburst of studies on groupwise and longitudinal analysis of early brain development in the upcoming years. To address this critical need, in this paper, we have developed a mathematical framework for the construction of an unbiased deformable spatiotemporal atlas of the fetal brain MRI and compared it to alternative configurations in terms of similarity metrics and deformation models. Our contributions are twofold: first we suggest a novel approach to fetal brain spatiotemporal atlas construction that shows high capability in capturing anatomic variation between subjects; and second, within our atlas construction framework we evaluate and compare a set of plausible configurations for inter-subject fetal brain MRI registration and identify the most accurate approach that can potentially lead to most accurate results in population atlas construction, atlas-based segmentation, and group analysis. Our evaluation results indicate that symmetric diffeomorphic deformable registration with cross correlation similarity metric outperforms other configurations in this application and results in sharp unbiased atlases that can be used in fetal brain MRI analysis.

Cortical thickness correlates of socioemotional difficulties in adults with Turner syndrome.

Turner syndrome (TS) is a non-inherited genetic disorder associated with a specific cognitive phenotype and socioemotional impairments. The present study aimed at characterizing the neuroanatomical basis of socioemotional dysfunctions in TS using the Emotional Quotient inventory (EQ-I) and cortical morphology analysis in 17 individuals with TS (45,X) and 17-age and verbal IQ matched healthy females. Individuals with TS reported significantly greater socioemotional impairment than controls. Cortical thickness analysis showed that participants with TS had an overall thicker cortex than controls, with extensive alterations in the temporal, frontal, parietal and insular regions bilaterally. Using the total EQ-I score as regressor in the cortical thickness analysis revealed a number of brain regions where the relationship between cortical thickness and EQ-I score differed between groups; these areas included brain regions critically involved in socioemotional processes, such as bilateral insula, the anterior cingulate and the orbitofrontal cortex. These results show that socioemotional dysfunctions seen in women with TS are associated with significant alterations in brain morphology.

Abnormal motor cortex excitability is associated with reduced cortical thickness in X monosomy.

Turner syndrome (TS) is a noninherited genetic disorder caused by the absence of one or part of one X chromosome. It is characterized by physical and cognitive phenotypes that include motor deficits that may be related to neuroanatomical abnormalities of sensorimotor pathways. Here, we used transcranial magnetic stimulation (TMS) and cortical thickness analysis to assess motor cortex excitability and cortical morphology in 17 individuals with TS (45, X) and 17 healthy controls. Exploratory analysis was performed to detect the effect of parental origin of the X chromosome (X(mat), X(pat)) on both measures. Results showed that long-interval intracortical inhibition was reduced and motor threshold (MT) was increased in TS relative to controls. Areas of reduced thickness were observed in the precentral gyrus of individuals with TS that correlated with MT. A significant difference between X(mat) (n = 11) and X(pat) (n = 6) individuals was found on the measure of long-interval intracortical inhibition. These findings demonstrate the presence of converging anatomical and neurophysiological abnormalities of the motor system in X monosomy.

Novel applications of quantitative MRI for the fetal brain.

The advent of ultrafast MRI acquisitions is offering vital insights into the critical maturational events that occur throughout pregnancy. Concurrent with the ongoing enhancement of ultrafast imaging has been the development of innovative image-processing techniques that are enabling us to capture and quantify the exuberant growth, and organizational and remodeling processes that occur during fetal brain development. This paper provides an overview of the role of advanced neuroimaging techniques to study in vivo brain maturation and explores the application of a range of new quantitative imaging biomarkers that can be used clinically to monitor high-risk pregnancies.

Quantitative in vivo MRI measurement of cortical development in the fetus.

Normal brain development is associated with expansion and folding of the cerebral cortex following a highly orchestrated sequence of gyral-sulcal formation. Although several studies have described the evolution of cerebral cortical development ex vivo or ex utero, to date, very few studies have characterized and quantified the gyrification process for the in vivo fetal brain. Recent advances in fetal magnetic resonance imaging and post-processing computational methods are providing new insights into fetal brain maturation in vivo. In this study, we investigate the in vivo fetal cortical folding pattern in healthy fetuses between 25 and 35 weeks gestational age using 3-D reconstructed fetal cortical surfaces. We describe the in vivo fetal gyrification process using a robust feature extraction algorithm applied directly on the cortical surface, providing an explicit delineation of the sulcal pattern during fetal brain development. We also delineate cortical surface measures, including surface area and gyrification index. Our data support an exuberant third trimester gyrification process and suggest a non-linear evolution of sulcal development. The availability of normative indices of cerebral cortical developing in the living fetus may provide critical insights on the timing and progression of impaired cerebral development in the high-risk fetus.

Normative fetal brain growth by quantitative in vivo magnetic resonance imaging.

OBJECTIVE: The objective of the study was to characterize total and regional volumetric brain growth in healthy fetuses during the second and third trimesters of pregnancy, using an automated method. STUDY DESIGN: We developed and validated an automated method to quantify global and regional in vivo brain volumes using fetal magnetic resonance imaging. We then computed the percentage of growth for each brain structure in a cohort of 64 healthy fetuses (25.4-36.6 weeks' gestational age). RESULTS: The cerebellum demonstrated the greatest maturation rate, with a 4-fold increase (384%) in volume between 25.4 and 36.6 weeks, and a relative growth rate of 12.87% per week. Both total brain and cerebral volumes increased by 230% and brain stem volume by 134% over the same gestational age period. Conversely, lateral ventricular volume decreased by 4.18% per week. CONCLUSION: The availability and ongoing validation of normative fetal brain growth trajectories will provide important tools for early detection of impaired fetal brain growth upon which to manage high-risk pregnancies.

Advancing fetal brain MRI: targets for the future.

Fetal MRI is becoming an increasingly powerful imaging tool for studying brain development in vivo. Until recently, the application of advanced magnetic resonance imaging techniques was limited by motion in the nonsedated fetus. Extensive research efforts currently underway are focusing on the development of dedicated magnetic resonance imaging sequences and sophisticated postprocessing techniques that are revolutionizing our ability to study the healthy and compromised fetus. The ongoing refinement of these magnetic resonance imaging techniques will undoubtedly lead to the development of cornerstone biomarkers that will provide healthcare caregivers with vital, and currently lacking, information upon which to counsel parents effectively, and base rational decisions regarding the timing and type of novel medical and surgical interventions currently on the horizon.

Surface-based structural group analysis of fMRI data.

As structural and surface-based analyses gain interest for activation detection, morphometry and intersubject matching purposes, this paper proposes a method to perform structural group analyses directly on the cortical surface. Scale-space blobs are extracted from surface-based functional maps and matched across subjects. The process aims at identifying activations within a population despite the various effects due to variability. Results of the method are presented with simulated activations and with data from a somatotopy protocol.

A new cortical surface parcellation model and its automatic implementation.

In this paper, we present an original method that aims at parcellating the cortical surface in regions functionally meaningful, from individual anatomy. The parcellation is obtained using an anatomically constrained surface-based coordinate system from which we define a complete partition of the surface. The aim of our method is to exhibit a new way to describe the cortical surface organization, in both anatomical and functional terms. The method is described together with results applied to a functional somatotopy experiments.