Skeleton optimization of neuronal morphology based on three-dimensional shape restrictions.

BACKGROUND: Neurons are the basic structural unit of the brain, and their morphology is a key determinant of their classification. The morphology of a neuronal circuit is a fundamental component in neuron modeling. Recently, single-neuron morphologies of the whole brain have been used in many studies. The correctness and completeness of semimanually traced neuronal morphology are credible. However, there are some inaccuracies in semimanual tracing results. The distance between consecutive nodes marked by humans is very long, spanning multiple voxels. On the other hand, the nodes are marked around the centerline of the neuronal fiber, not on the centerline. Although these inaccuracies do not seriously affect the projection patterns that these studies focus on, they reduce the accuracy of the traced neuronal skeletons. These small inaccuracies will introduce deviations into subsequent studies that are based on neuronal morphology files. RESULTS: We propose a neuronal digital skeleton optimization method to evaluate and make fine adjustments to a digital skeleton after neuron tracing. Provided that the neuronal fiber shape is smooth and continuous, we describe its physical properties according to two shape restrictions. One restriction is designed based on the grayscale image, and the other is designed based on geometry. These two restrictions are designed to finely adjust the digital skeleton points to the neuronal fiber centerline. With this method, we design the three-dimensional shape restriction workflow of neuronal skeleton adjustment computation. The performance of the proposed method has been quantitatively evaluated using synthetic and real neuronal image data. The results show that our method can reduce the difference between the traced neuronal skeleton and the centerline of the neuronal fiber. Furthermore, morphology metrics such as the neuronal fiber length and radius become more precise. CONCLUSIONS: This method can improve the accuracy of a neuronal digital skeleton based on traced results. The greater the accuracy of the digital skeletons that are acquired, the more precise the neuronal morphologies that are analyzed will be.

Direct 3D Analyses Reveal Barrel-Specific Vascular Distribution and Cross-Barrel Branching in the Mouse Barrel Cortex.

Whether vascular distribution is spatially specific among cortical columns is a fundamental yet controversial question. Here, we have obtained 1-µm resolution 3D datasets that cover the whole mouse barrel cortex by combining Nissl staining with micro-optical sectioning tomography to simultaneously visualize individual cells and blood vessels, including capillaries. Pinpointing layer IV of the posteromedial barrel subfield, direct 3D reconstruction and quantitative analysis showed that (1) penetrating vessels preferentially locate in the interbarrel septa/barrel wall (75.1%) rather than the barrel hollows, (2) the branches of 70% penetrating vessels only reach the neighboring but not always all the neighboring barrels and the other 30% extend beyond the neighboring barrels and may provide cross-barrel blood supply or drainage, (3) the branches of 59.6% penetrating vessels reach all the neighboring barrels, while the rest only reach part of them, and (4) the length density of microvessels in the interbarrel septa/barrel wall is lower than that in the barrel hollows with a ratio of 0.92. These results reveal that the penetrating vessels and microvessels exhibit a barrel-specific organization, whereas the branches of penetrating vessels do not, which suggests a much more complex vascular distribution pattern among cortical columns than previously thought.

Prolonged deficits in parvalbumin neuron stimulation-evoked network activity despite recovery of dendritic structure and excitability in the somatosensory cortex following global ischemia in mice.

Relatively few studies have examined plasticity of inhibitory neuronal networks following stroke in vivo, primarily due to the inability to selectively monitor inhibition. We assessed the structure of parvalbumin (PV) interneurons during a 5 min period of global ischemia and reperfusion in mice, which mimicked cerebral ischemia during cardiac arrest or forms of transient ischemic attack. The dendritic structure of PV-neurons in cortical superficial layers was rapidly swollen and beaded during global ischemia, but recovered within 5-10 min following reperfusion. Using optogenetics and a multichannel optrode, we investigated the function of PV-neurons in mouse forelimb somatosensory cortex. We demonstrated pharmacologically that PV-channelrhodopsin-2 (ChR2) stimulation evoked activation in layer IV/V, which resulted in rapid current sinks mediated by photocurrent and action potentials (a measure of PV-neuron excitability), which was then followed by current sources mediated by network GABAergic synaptic activity. During ischemic depolarization, the PV-ChR2-evoked current sinks (excitability) were suppressed, but recovered rapidly following reperfusion concurrent with repolarization of the DC-EEG. In contrast, the current sources reflecting GABAergic synaptic network activity recovered slowly and incompletely, and was coincident with the partial recovery of the forepaw stimulation-evoked current sinks in layer IV/V 30 min post reperfusion. Our in vivo data suggest that the excitability of PV inhibitory neurons was suppressed during global ischemia and rapidly recovered during reperfusion. In contrast, PV-ChR2 stimulation-evoked GABAergic synaptic network activity exhibited a prolonged suppression even ∼1 h after reperfusion, which could contribute to the dysfunction of sensation and cognition following transient global ischemia.

3D BrainCV: simultaneous visualization and analysis of cells and capillaries in a whole mouse brain with one-micron voxel resolution.

Systematic cellular and vascular configurations are essential for understanding fundamental brain anatomy and metabolism. We demonstrated a 3D brainwide cellular and vascular (called 3D BrainCV) visualization and quantitative protocol for a whole mouse brain. We developed a modified Nissl staining method that quickly labeled the cells and blood vessels simultaneously in an entire mouse brain. Terabytes 3D datasets of the whole mouse brains, with unprecedented details of both individual cells and blood vessels, including capillaries, were simultaneously imaged at 1-µm voxel resolution using micro-optical sectioning tomography (MOST). For quantitative analysis, we proposed an automatic image-processing pipeline to perform brainwide vectorization and analysis of cells and blood vessels. Six representative brain regions from the cortex to the deep, including FrA, M1, PMBSF, V1, striatum, and amygdala, and six parameters, including cell number density, vascular length density, fractional vascular volume, distance from the cells to the nearest microvessel, microvascular length density, and fractional microvascular volume, had been quantitatively analyzed. The results showed that the proximity of cells to blood vessels was linearly correlated with vascular length density, rather than the cell number density. The 3D BrainCV made overall snapshots of the detailed picture of the whole brain architecture, which could be beneficial for the state comparison of the developing and diseased brain.

Altered functional connectivity after pilocarpine-induced seizures revealed by intrinsic optical signals imaging in awake mice.

Significance: Comorbidities such as mood and cognitive disorders are often found in individuals with epilepsy after seizures. Cortex processes sensory, motor, and cognitive information. Brain circuit changes can be studied by observing functional network changes in epileptic mice's cortex. Aim: The cortex is easily accessible for non-invasive brain imaging and electroencephalogram recording (EEG). However, the impact of seizures on cortical activity and functional connectivity has been rarely studied in vivo. Approach: Intrinsic optical signal and EEG were used to monitor cortical activity in awake mice within 4 h after pilocarpine induction. It was divided into three periods according to the behavior and EEG of the mice: baseline, onset of seizures (onset, including seizures and resting in between seizure events), and after seizures (post, without seizures). Changes in cortical activity were compared between the baseline and after seizures. Results: Hemoglobin levels increased significantly, particularly in the parietal association cortex (PT), retrosplenial cortex (RS), primary visual cortex (V1), and secondary visual cortex (V2). The network-wide functional connectivity changed post seizures, e.g., hypoconnectivity between PT and visual-associated cortex (e.g., V1 and V2). In contrast, connectivity between the motor-associated cortex and most other regions increased. In addition, the default mode network (DMN) also changed after seizures, with decreased connectivity between primary somatosensory region (SSp) and visual region (VIS), but increased connectivity involving anterior cingulate cortex (AC) and RS. Conclusions: Our results provide references for understanding the mechanisms behind changes in brain circuits, which may explain the profound effects of seizures on comorbid health conditions.

Optogenetic analysis of neuronal excitability during global ischemia reveals selective deficits in sensory processing following reperfusion in mouse cortex.

We have developed an approach to directly probe neuronal excitability during the period beginning with induction of global ischemia and extending after reperfusion using transgenic mice expressing channelrhodopsin-2 (ChR2) to activate deep layer cortical neurons independent of synaptic or sensory stimulation. Spontaneous, ChR2, or forepaw stimulation-evoked electroencephalogram (EEG) or local field potential (LFP) records were collected from the somatosensory cortex. Within 20 s of ischemia, a >90% depression of spontaneous 0.3-3 Hz EEG and LFP power was detected. Ischemic depolarization followed EEG depression with a ∼2 min delay. Surprisingly, neuronal excitability, as assessed by the ChR2-mediated EEG response, was intact during the period of strong spontaneous EEG suppression and actually increased before ischemic depolarization. In contrast, a decrease in the somatosensory-evoked potential (forepaw-evoked potential, reflecting cortical synaptic transmission) was coincident with the EEG suppression. After 5 min of ischemia, the animal was reperfused, and the ChR2-mediated response mostly recovered within 30 min (>80% of preischemia value). However, the recovery of the somatosensory-evoked potential was significantly delayed compared with the ChR2-mediated response (<40% of preischemia value at 60 min). By assessing intrinsic optical signals in combination with EEG, we found that neuronal excitability approached minimal values when the spreading ischemic depolarization wave propagated to the ChR2-stimulated cortex. Our results indicate that the ChR2-mediated EEG/LFP response recovers much faster than sensory-evoked EEG/LFP activity in vivo following ischemia and reperfusion, defining a period where excitable but synaptically silent neurons are present.

[Study of coupling mechanism between hydrocarbon generation and structure evolution in low rank coal].

The structure evolution in the process of low rank coal hydrocarbon generation was studied using Fourier transform infrared spectroscopy and gas chromatography. Gaseous hydrocarbon yield and change law of functional groups were obtained. The results show that: the coal pyrolysis products are mainly gaseous hydrocarbon C1-5. Methane generation instantaneous yield curve contains four peak of hydrocarbon pyrolysis. Oxygen-containing functional group and alkyl side chain of low rank coal chemical structure reduced while aromatization degree increased along with coal rank. The characteristic absorption peak of coal structure of aliphatic hydrocarbons, aromatic hydrocarbon, methyl C=O base C=C of alkanes and aromatic structure of methyl and methylene were characterized by 2 950, 2 920, 2 860, 1 730, 1 705, 1 600 and 1 380-1 460 cm(-1) selected in FTIR spectra. Temperature 420 degrees C is the turning point, before the absorption peak intensity gradually decreases, and then increases slightly. Three major structural evolution stages of coalification mechanism were revealed. Finally, the low rank coal hydrocarbon structure evolution pattern was put forward.

Probing microdomain Ca2+ activity and synaptic transmission with a node-based tripartite synapse model.

Astrocytic fine processes are the most minor structures of astrocytes but host much of the Ca2+ activity. These localized Ca2+ signals spatially restricted to microdomains are crucial for information processing and synaptic transmission. However, the mechanistic link between astrocytic nanoscale processes and microdomain Ca2+ activity remains hazily understood because of the technical difficulties in accessing this structurally unresolved region. In this study, we used computational models to disentangle the intricate relations of morphology and local Ca2+ dynamics involved in astrocytic fine processes. We aimed to answer: 1) how nano-morphology affects local Ca2+ activity and synaptic transmission, 2) and how fine processes affect Ca2+ activity of large process they connect. To address these issues, we undertook the following two computational modeling: 1) we integrated the in vivo astrocyte morphological data from a recent study performed with super-resolution microscopy that discriminates sub-compartments of various shapes, referred to as nodes and shafts to a classic IP3R-mediated Ca2+ signaling framework describing the intracellular Ca2+ dynamics, 2) we proposed a node-based tripartite synapse model linking with astrocytic morphology to predict the effect of structural deficits of astrocytes on synaptic transmission. Extensive simulations provided us with several biological insights: 1) the width of nodes and shafts could strongly influence the spatiotemporal variability of Ca2+ signals properties but what indeed determined the Ca2+ activity was the width ratio between nodes and shafts, 2) the connectivity of nodes to larger processes markedly shaped the Ca2+ signal of the parent process rather than nodes morphology itself, 3) the morphological changes of astrocytic part might potentially induce the abnormality of synaptic transmission by affecting the level of glutamate at tripartite synapses. Taken together, this comprehensive model which integrated theoretical computation and in vivo morphological data highlights the role of the nanomorphology of astrocytes in signal transmission and its possible mechanisms related to pathological conditions.

An efficient reduced-lattice model of IP3R for probing Ca2+ dynamics.

Numerous cellular processes are regulated by Ca2+ signals, and the endoplasmic reticulum (ER) membrane's inositol triphosphate receptor (IP3R) is critical for modulating intracellular Ca2+ dynamics. The IP3Rs are seen to be clustered in a variety of cell types. The combination of IP3Rs clustering and IP3Rs-mediated Ca2+-induced Ca2+ release results in the hierarchical organization of the Ca2+ signals, which challenges the numerical simulation given the multiple spatial and temporal scales that must be covered. The previous methods rather ignore the spatial feature of IP3Rs or fail to coordinate the conflicts between the real biological relevance and the computational cost. In this work, a general and efficient reduced-lattice model is presented for the simulation of IP3Rs-mediated multiscale Ca2+ dynamics. The model highlights biological details that make up the majority of the calcium events, including IP3Rs clustering and calcium domains, and it reduces the complexity by approximating the minor details. The model's extensibility provides fresh insights into the function of IP3Rs in producing global Ca2+ events and supports the research under more physiological circumstances. Our work contributes to a novel toolkit for modeling multiscale Ca2+ dynamics and advances knowledge of Ca2+ signals.

Using head-mounted eye trackers to explore children's color preferences and perceptions of toys with different color gradients.

This study investigated how color gradients affect the attraction and visual comfort of children aged 4 to 7 years. We analyzed 108 eye-tracking datasets, including the color attraction index (COI), visual comfort index (PUI), and saccade rate (SR). The findings revealed that children are more attracted to colors as saturation decreases and brightness increases within a specific range. Beyond this range, reduced saturation diminishes color appeal. Moderate brightness and contrast enhance visual comfort during play, while extremely low contrast hinders concentration. Warm colors (red, orange, and yellow) slightly dominate preferences; however, the roles of hue, saturation, and brightness in children's color preferences remain inconclusive. These insights have practical implications for age-appropriate toy design and marketing. Future research should explore age-specific color preferences for more targeted design strategies.

DDeep3M+: adaptive enhancement powered weakly supervised learning for neuron segmentation.

Significance: Robust segmentations of neurons greatly improve neuronal population reconstruction, which could support further study of neuron morphology for brain research. Aim: Precise segmentation of 3D neuron structures from optical microscopy (OM) images is crucial to probe neural circuits and brain functions. However, the high noise and low contrast of images make neuron segmentation challenging. Convolutional neural networks (CNNs) can provide feasible solutions for the task but they require large manual labels for training. Labor-intensive labeling is highly expensive and heavily limits the algorithm generalization. Approach: We devise a weakly supervised learning framework Docker-based deep network plus (DDeep3M+) for neuron segmentation without any manual labeling. A Hessian analysis based adaptive enhancement filter is employed to generate pseudo-labels for segmenting neuron images. The automated segmentation labels are input for training a DDeep3M to extract neuronal features. We mine more undetected weak neurites from the probability map based on neuronal structures, thereby modifying the pseudo-labels. We iteratively refine the pseudo-labels and retrain the DDeep3M model with the pseudo-labels to obtain a final segmentation result. Results: The proposed method achieves promising results with the F1 score of 0.973, which is close to that of the CNN model with manual labels and superior to several segmentation algorithms. Conclusions: We propose an accurate weakly supervised neuron segmentation method. The high precision results achieved on 3D OM datasets demonstrate the superior generalization of our DDeep3M+.

Astrocytes exhibit diverse Ca2+ changes at subcellular domains during brain aging.

Astrocytic Ca2+ transients are essential for astrocyte integration into neural circuits. These Ca2+ transients are primarily sequestered in subcellular domains, including primary branches, branchlets and leaflets, and endfeet. In previous studies, it suggests that aging causes functional defects in astrocytes. Until now, it was unclear whether and how aging affects astrocytic Ca2+ transients at subcellular domains. In this study, we combined a genetically encoded Ca2+ sensor (GCaMP6f) and in vivo two-photon Ca2+ imaging to determine changes in Ca2+ transients within astrocytic subcellular domains during brain aging. We showed that aging increased Ca2+ transients in astrocytic primary branches, higher-order branchlets, and terminal leaflets. However, Ca2+ transients decreased within astrocytic endfeet during brain aging, which could be caused by the decreased expressions of Aquaporin-4 (AQP4). In addition, aging-induced changes of Ca2+ transient types were heterogeneous within astrocytic subcellular domains. These results demonstrate that the astrocytic Ca2+ transients within subcellular domains are affected by aging differently. This finding contributes to a better understanding of the physiological role of astrocytes in aging-induced neural circuit degeneration.

Simultaneous monitoring of intracellular pH changes and hemodynamic response during cortical spreading depression by fluorescence-corrected multimodal optical imaging.

Cortical spreading depression (CSD) plays an important role in trauma, migraine and ischemia. CSD could induce pronounced hemodynamic changes and the disturbance of pH homeostasis which has been postulated to contribute to cell death following ischemia. In this study, we described a fluorescence-corrected multimodal optical imaging system to simultaneously monitor CSD associated intracellular pH (pH(i)) changes and hemodynamic response including hemoglobin concentrations and cerebral blood flow (CBF). CSD was elicited by application of KCl on rat cortex and direct current (DC) potential was recorded as a typical characteristic of CSD. The pH(i) shift was mapped by neutral red (NR) fluorescence which was excited at 516-556 nm and emitted at 625 nm. The changes in hemoglobin concentrations were determined by dual-wavelength optical intrinsic signal imaging (OISI) at 550 nm and 625 nm. Integration of fluorescence imaging and dual-wavelength OISI was achieved by a time-sharing camera equipped with a liquid crystal tunable filter (LCTF). CBF was visualized by laser speckle contrast imaging (LSCI) through a separate camera. Besides, based on the dual-wavelength optical intrinsic signals (OISs) obtained from our system, NR fluorescence was corrected according to our method of fluorescence correction. We found that a transient intracellular acidification followed by a small alkalization occurred during CSD. After CSD, there was a prolonged intracellular acidification and the recovery of pH(i) from CSD took much longer time than those of hemodynamic response. Our results suggested that the new multimodal optical imaging system had the potential to advance our knowledge of CSD and might work as a useful tool to exploit neurovascular coupling under physiological and pathological conditions.

Simultaneous automatic arteries-veins separation and cerebral blood flow imaging with single-wavelength laser speckle imaging.

Automatic separation of arteries and veins in optical cerebral cortex images is important in clinical practice and preclinical study. In this paper, a simple but effective automatic artery-vein separation method which utilizes single-wavelength coherent illumination is presented. This method is based on the relative temporal minimum reflectance analysis of laser speckle images. The validation is demonstrated with both theoretic simulations and experimental results applied to the rat cortex. Moreover, this method can be combined with laser speckle contrast analysis so that the artery-vein separation and blood flow imaging can be simultaneously obtained using the same raw laser speckle images data to enable more accurate analysis of changes of cerebral blood flow within different tissue compartments during functional activation, disease dynamic, and neurosurgery, which may broaden the applications of laser speckle imaging in biology and medicine.

Adaptive region of interest method for analytical micro-CT reconstruction.

The real-time imaging is important in automatic successive inspection with micro-computerized tomography (micro-CT). Generally, the size of the detector is chosen according to the most probable size of the measured object to acquire all the projection data. Given enough imaging area and imaging resolution of X-ray detector, the detector is larger than specimen projection area, which results in redundant data in the Sinogram. The process of real-time micro-CT is computation-intensive because of the large amounts of source and destination data. The speed of the reconstruction algorithm can't always meet the requirements of real-time applications. A preprocessing method called adaptive region of interest (AROI), which detects the object's boundaries automatically to focus the active Sinogram regions, is introduced into the analytical reconstruction algorithm in this paper. The AROI method reduces the volume of the reconstructing data and thus directly accelerates the reconstruction process. It has been further shown that image quality is not compromised when applying AROI, while the reconstruction speed is increased as the square of the ratio of the sizes of the detector and the specimen slice. In practice, the conch reconstruction experiment indicated that the process is accelerated by 5.2 times with AROI and the imaging quality is not degraded. Therefore, the AROI method improves the speed of analytical micro-CT reconstruction significantly.

Unraveling Aß-Mediated Multi-Pathway Calcium Dynamics in Astrocytes: Implications for Alzheimer's Disease Treatment From Simulations.

The accumulation of amyloid ß peptide (Aß) in the brain is hypothesized to be the major factor driving Alzheimer's disease (AD) pathogenesis. Mounting evidence suggests that astrocytes are the primary target of Aß neurotoxicity. Aß is known to interfere with multiple calcium fluxes, thus disrupting the calcium homeostasis regulation of astrocytes, which are likely to produce calcium oscillations. Ca2+ dyshomeostasis has been observed to precede the appearance of clinical symptoms of AD; however, it is experimentally very difficult to investigate the interactions of many mechanisms. Given that Ca2+ disruption is ubiquitously involved in AD progression, it is likely that focusing on Ca2+ dysregulation may serve as a potential therapeutic approach to preventing or treating AD, while current hypotheses concerning AD have so far failed to yield curable therapies. For this purpose, we derive and investigate a concise mathematical model for Aß-mediated multi-pathway astrocytic intracellular Ca2+ dynamics. This model accounts for how Aß affects various fluxes contributions through voltage-gated calcium channels, Aß-formed channels and ryanodine receptors. Bifurcation analysis of Aß level, which reflected the corresponding progression of the disease, revealed that Aß significantly induced the increasing [Ca2+] i and frequency of calcium oscillations. The influence of inositol 1,4,5-trisphosphate production (IP3) is also investigated in the presence of Aß as well as the impact of changes in resting membrane potential. In turn, the Ca2+ flux can be considerably changed by exerting specific interventions, such as ion channel blockers or receptor antagonists. By doing so, a "combination therapy" targeting multiple pathways simultaneously has finally been demonstrated to be more effective. This study helps to better understand the effect of Aß, and our findings provide new insight into the treatment of AD.

Src Family Kinases in the Central Nervous System: Their Emerging Role in Pathophysiology of Migraine and Neuropathic Pain.

Src family kinases (SFK) are a group of non-receptor tyrosine kinases which play a pivotal role in cellular responses and oncogenesis. Accumulating evidence suggest that SFK also act as a key component in signalling pathways of the central nervous system (CNS) in both physiological and pathological conditions. Despite the crucial role of SFK in signal transduction of the CNS, the relationship between SFK and molecules implicated in pain has been relatively unexplored. This article briefly reviews the recent advances uncovering the interplay of SFK with diverse membrane proteins and intracellular proteins in the CNS and the importance of SFK in the pathophysiology of migraine and neuropathic pain. Mechanisms underlying the role of SFK in these conditions and potential clinical applications of SFK inhibitors in neurological diseases are also summarised. We propose that SFK are the convergent point of signalling pathways in migraine and neuropathic pain and may constitute a promising therapeutic target for these diseases.

Boosting Multilabel Semantic Segmentation for Somata and Vessels in Mouse Brain.

Deep convolutional neural networks (DCNNs) are widely utilized for the semantic segmentation of dense nerve tissues from light and electron microscopy (EM) image data; the goal of this technique is to achieve efficient and accurate three-dimensional reconstruction of the vasculature and neural networks in the brain. The success of these tasks heavily depends on the amount, and especially the quality, of the human-annotated labels fed into DCNNs. However, it is often difficult to acquire the gold standard of human-annotated labels for dense nerve tissues; human annotations inevitably contain discrepancies or even errors, which substantially impact the performance of DCNNs. Thus, a novel boosting framework consisting of a DCNN for multilabel semantic segmentation with a customized Dice-logarithmic loss function, a fusion module combining the annotated labels and the corresponding predictions from the DCNN, and a boosting algorithm to sequentially update the sample weights during network training iterations was proposed to systematically improve the quality of the annotated labels; this framework eventually resulted in improved segmentation task performance. The microoptical sectioning tomography (MOST) dataset was then employed to assess the effectiveness of the proposed framework. The result indicated that the framework, even trained with a dataset including some poor-quality human-annotated labels, achieved state-of-the-art performance in the segmentation of somata and vessels in the mouse brain. Thus, the proposed technique of artificial intelligence could advance neuroscience research.

Accurate Neuronal Soma Segmentation Using 3D Multi-Task Learning U-Shaped Fully Convolutional Neural Networks.

Neuronal soma segmentation is a crucial step for the quantitative analysis of neuronal morphology. Automated neuronal soma segmentation methods have opened up the opportunity to improve the time-consuming manual labeling required during the neuronal soma morphology reconstruction for large-scale images. However, the presence of touching neuronal somata and variable soma shapes in images brings challenges for automated algorithms. This study proposes a neuronal soma segmentation method combining 3D U-shaped fully convolutional neural networks with multi-task learning. Compared to existing methods, this technique applies multi-task learning to predict the soma boundary to split touching somata, and adopts U-shaped architecture convolutional neural network which is effective for a limited dataset. The contour-aware multi-task learning framework is applied to the proposed method to predict the masks of neuronal somata and boundaries simultaneously. In addition, a spatial attention module is embedded into the multi-task model to improve neuronal soma segmentation results. The Nissl-stained dataset captured by the micro-optical sectioning tomography system is used to validate the proposed method. Following comparison to four existing segmentation models, the proposed method outperforms the others notably in both localization and segmentation. The novel method has potential for high-throughput neuronal soma segmentation in large-scale optical imaging data for neuron morphology quantitative analysis.

DDeep3M: Docker-powered deep learning for biomedical image segmentation.

BACKGROUND: Deep learning models are turning out to be increasingly popular in biomedical image processing. The fruitful utilization of these models, in most cases, is substantially restricted by the complicated configuration of computational environments, resulting in the noteworthy increment of the time and endeavors to reproduce the outcomes of the models. NEW METHOD: We thus present a Docker-based method for better use of deep learning models and quicker reproduction of model performance for multiple data sources, permitting progressively more biomedical scientists to attempt the new technology conveniently in their domain. Here, we introduce a Docker-powered deep learning model, named as DDeep3M and validated it with the electron microscopy data volumes (microscale). RESULTS: DDeep3M is utilized to the 3D optical microscopy image stack in mouse brain for the image segmentation (mesoscale). It achieves high accuracy on both vessels and somata structures with all the recall/precision scores and Dice indexes over 0.96. DDeep3M also reports the state-of-the-art performance in the MRI data (macroscale) for brain tumor segmentation. COMPARISON WITH EXISTING METHODS: We compare the performance and efficiency of DDeep3M with three existing models on image datasets varying from micro- to macro-scales. CONCLUSION: DDeep3M is a friendly, convenient and efficient tool for image segmentations in biomedical research. DDeep3M is open sourced with the codes and pretrained model weights available at https://github.com/cakuba/DDeep3m.