Morphological Reconstruction Research Articles

Deep Learning-Based Reconstruction of 3D Morphology of Geomaterial Particles from Single-View 2D Images.

The morphology of particles formed in different environments contains critical information. Thus, the rapid and effective reconstruction of their three-dimensional (3D) morphology is crucial. This study reconstructs the 3D morphology from two-dimensional (2D) images of particles using artificial intelligence (AI). More than 100,000 particles were sampled from three sources: naturally formed particles (desert sand), manufactured particles (lunar soil simulant), and numerically generated digital particles. A deep learning approach based on a voxel representation of the morphology and multi-dimensional convolutional neural networks was proposed to rapidly upscale and reconstruct particle morphology. The trained model was tested using the three particle types and evaluated using different multi-scale morphological descriptors. The results demonstrated that the statistical properties of the morphological descriptors were consistent for the real 3D particles and those derived from the 2D images and the model. This finding confirms the model's validity and generalizability in upscaling and reconstructing diverse particle samples. This study provides a method for generating 3D numerical representations of geological particles, facilitating in-depth analysis of properties, such as mechanical behavior and transport characteristics, from 2D images.

Multi-scale spiking network model of human cerebral cortex.

Although the structure of cortical networks provides the necessary substrate for their neuronal activity, the structure alone does not suffice to understand the activity. Leveraging the increasing availability of human data, we developed a multi-scale, spiking network model of human cortex to investigate the relationship between structure and dynamics. In this model, each area in one hemisphere of the Desikan-Killiany parcellation is represented by a $1\,\mathrm{mm^{2}}$ column with a layered structure. The model aggregates data across multiple modalities, including electron microscopy, electrophysiology, morphological reconstructions, and diffusion tensor imaging, into a coherent framework. It predicts activity on all scales from the single-neuron spiking activity to the area-level functional connectivity. We compared the model activity with human electrophysiological data and human resting-state functional magnetic resonance imaging (fMRI) data. This comparison reveals that the model can reproduce aspects of both spiking statistics and fMRI correlations if the inter-areal connections are sufficiently strong. Furthermore, we study the propagation of a single-spike perturbation and macroscopic fluctuations through the network. The open-source model serves as an integrative platform for further refinements and future in silico studies of human cortical structure, dynamics, and function.

A Novel Mechanism to Accelerate Stress Relaxation of Toughened Blends: Stress-Induced Core-Shell Morphological Reconstruction and Its Application in Thermoforming and Dimensional Stabilization

Improving thermoforming efficiency and dimensional stability in thermoplastic products is a common challenge. This study investigates toughened polyamide 612 (PA612) blends made by adding maleic anhydride-functionalized SEBS (mSEBS) elastomers, along with high-density polyethylene (HDPE), polyphenylene Oxide (PPO), and polyphenylene Sulfide (PPS). Analyses showed that mSEBS forms a core-shell structure with HDPE or PPO and a sea-island structure with PPS in the PA612 matrix. The study uniquely examines how these structures affect stress relaxation. The results, modeled by a steady-state creep model, revealed that the core-shell structures reduced the characteristic relaxation time (λ) by over three orders of magnitude compared to PA612/mSEBS blends. Additionally, PA612/mSEBS/HDPE blends were more temperature-sensitive, reducing λ by six orders of magnitude compared to PA612/mSEBS/PPO blends. Further analysis showed that stress-induced core-shell morphological reconstruction (SCMR) significantly improved stress relaxation by promoting extensive plastic deformation and energy dissipation. These toughened PA612 blends exhibited excellent thermoforming efficiency and dimensional stability. A 3D finite element model confirmed SCMR as an effective strategy for stress relaxation, providing valuable insights for designing toughened blends with superior processing efficiency and stability.

The Regulatory Role of miRNAs in Zebrafish Fin Regeneration

Since Teleostei fins have a strong regenerative capacity, further research was conducted on the regulation of gene expression during fin regeneration. This research focuses on miRNA, which is a key post-transcriptional regulatory molecule. In this study, a miRNA library for the fin regeneration of zebrafish was constructed to reveal the differential expression of miRNA during fin regeneration and to explore the regulatory pathway for fin regeneration. Following the injection of miRNA agomir into zebrafish, the proliferation of blastema cells and the overall fin regeneration area were significantly reduced. It was observed that the miRNAs impaired blastocyte formation by affecting fin regeneration through the inhibition of the expressions of genes and proteins associated with blastocyte formation (including yap1 and Smad1/5/9), which is an effect associated with the Hippo pathway. Furthermore, it has been demonstrated that miRNAs can impair the patterns and mineralization of newly formed fin rays. The miRNAs influenced fin regeneration by inhibiting the expression of a range of bone-related genes and proteins in osteoblast lineages, including sp7, runx2a, and runx2b. This study provides a valuable reference for the further exploration of morphological bone reconstruction in aquatic vertebrates.

Collaborative augmented reconstruction of 3D neuron morphology in mouse and human brains

Digital reconstruction of the intricate 3D morphology of individual neurons from microscopic images is a crucial challenge in both individual laboratories and large-scale projects focusing on cell types and brain anatomy. This task often fails in both conventional manual reconstruction and state-of-the-art artificial intelligence (AI)-based automatic reconstruction algorithms. It is also challenging to organize multiple neuroanatomists to generate and cross-validate biologically relevant and mutually agreed upon reconstructions in large-scale data production. Based on collaborative group intelligence augmented by AI, we developed a collaborative augmented reconstruction (CAR) platform for neuron reconstruction at scale. This platform allows for immersive interaction and efficient collaborative editing of neuron anatomy using a variety of devices, such as desktop workstations, virtual reality headsets and mobile phones, enabling users to contribute anytime and anywhere and to take advantage of several AI-based automation tools. We tested CAR’s applicability for challenging mouse and human neurons toward scaled and faithful data production.

Transcriptomic cell-type specificity of local cortical circuits.

Complex neocortical functions rely on networks of diverse excitatory and inhibitory neurons. While local connectivity rules between major neuronal subclasses have been established, the specificity of connections at the level of transcriptomic subtypes remains unclear. We introduce single transcriptome assisted rabies tracing (START), a method combining monosynaptic rabies tracing and single-nuclei RNA sequencing to identify transcriptomic cell types, providing inputs to defined neuron populations. We employ START to transcriptomically characterize inhibitory neurons providing monosynaptic input to 5 different layer-specific excitatory cortical neuron populations in mouse primary visual cortex (V1). At the subclass level, we observe results consistent with findings from prior studies that resolve neuronal subclasses using antibody staining, transgenic mouse lines, and morphological reconstruction. With improved neuronal subtype granularity achieved with START, we demonstrate transcriptomic subtype specificity of inhibitory inputs to various excitatory neuron subclasses. These results establish local connectivity rules at the resolution of transcriptomic inhibitory cell types.

Learning meaningful representation of single-neuron morphology via large-scale pre-training.

Single-neuron morphology, the study of the structure, form, and shape of a group of specialized cells in the nervous system, is of vital importance to define the type of neurons, assess changes in neuronal development and aging and determine the effects of brain disorders and treatments. Despite the recent surge in the amount of available neuron morphology reconstructions due to advancements in microscopy imaging, existing computational and deep learning methods for modeling neuron morphology have been limited in both scale and accuracy. In this paper, we propose MorphRep, a model for learning meaningful representation of neuron morphology pre-trained with over 250 000 existing neuron morphology data. By encoding the neuron morphology into graph-structured data, using graph transformers for feature encoding and enforcing the consistency between multiple augmented views of neuron morphology, MorphRep achieves the state of the art performance on widely used benchmarking datasets. Meanwhile, MorphRep can accurately characterize the neuron morphology space across neuron morphometrics, fine-grained cell types, brain regions and ages. Furthermore, MorphRep can be applied to distinguish neurons under a wide range of conditions, including genetic perturbation, drug injection, environment change and disease. In summary, MorphRep provides an effective strategy to embed and represent neuron morphology and can be a valuable tool in integrating cell morphology into single-cell multiomics analysis. The codebase has been deposited in https://github.com/YaxuanLi-cn/MorphRep.

Exploring de-anonymization risks in PET imaging: Insights from a comprehensive analysis of 853 patient scans

Due to their high resolution, anonymized CT scans can be reidentified using face recognition tools. However, little is known regarding PET deanonymization because of its lower resolution. In this study, we analysed PET/CT scans of 853 patients from a TCIA-restricted dataset (AutoPET). First, we built denoised 2D morphological reconstructions of both PET and CT scans, and then we determined how frequently a PET reconstruction could be matched to the correct CT reconstruction with no other metadata. Using the CT morphological reconstructions as ground truth allows us to frame the problem as a face recognition problem and to quantify our performance using traditional metrics (top k accuracies) without any use of patient pictures. Using our denoised PET 2D reconstructions, we achieved 72% top 10 accuracy after the realignment of all CTs in the same reference frame, and 71% top 10 accuracy after realignment and mixing within a larger face dataset of 10, 168 pictures. This highlights the need to consider face identification issues when dealing with PET imaging data.

StrainNet-LD: Large Displacement digital image correlation based on deep learning and displacement-field decomposition

First, a displacement-field decomposition strategy is proposed that effectively enhances the accuracy of digital image correlation (DIC) algorithms under large displacement or deformation conditions. The implementation methodologies of displacement-field decomposition under both Lagrangian and Eulerian descriptions are derived and presented. Second, displacement-field decomposition is applied to speckle image correlation algorithms based on deep learning, resolving the contradiction between registration capability of large displacement and the accuracy. A CUDA-accelerated and residual learning-based strain fitting algorithm is proposed for high-precision and real-time strain computation. This method is named “StrainNet for Large Displacement” (StrainNet-LD). StrainNet-LD achieves a calculation speed of 768 × 768 × 6 fps (768 × 768 × 3 fps with strain calculated simultaneously) on GPU with displacement MAE < 0.05 pixels and strain MAE < 0.002ε. Finally, rubber tensile tests, 3D motion measurement of composite shells, and morphology reconstruction experiments of speckle-free composite repair structures are conducted to validate the algorithm's robustness and efficiency. Some vital discussions of displacement-field decomposition are given in the Appendixes. The code and dataset are available at https://github.com/GW-Wang-thu/StrainNet-LD.

Correcting angular distortions in Bragg coherent X-ray diffraction imaging.

Bragg coherent X-ray diffraction imaging (BCDI) has emerged as a powerful technique for strain imaging and morphology reconstruction of nanometre-scale crystals. However, BCDI often suffers from angular distortions that appear during data acquisition, caused by radiation pressure, heating or imperfect scanning stages. This limits the applicability of BCDI, in particular for small crystals and high-flux X-ray beams. Here, we present a pre-processing algorithm that recovers the 3D datasets from the BCDI dataset measured under the impact of large angular distortions. We systematically investigate the performance of this method for different levels of distortion and find that the algorithm recovers the correct angles for distortions up to 16.4× (1640%) the angular step size dθ = 0.004°. We also show that the angles in a continuous scan can be recovered with high accuracy. As expected, the correction provides marked improvements in the subsequent phase retrieval.

Novel mathematical approach to accurately quantify 3D endothelial cell morphology and vessel geometry based on fluorescently marked endothelial cell contours: Application to the dorsal aorta of wild-type and Endoglin-deficient zebrafish embryos.

Endothelial cells, which line the lumen of blood vessels, locally sense and respond to blood flow. In response to altered blood flow dynamics during early embryonic development, these cells undergo shape changes that directly affect vessel geometry: In the dorsal aorta of zebrafish embryos, elongation of endothelial cells in the direction of flow between 48 and 72 hours post fertilization (hpf) reduces the vessel's diameter. This remodeling process requires Endoglin; excessive endothelial cell growth in the protein's absence results in vessel diameter increases. To understand how these changes in vessel geometry emerge from morphological changes of individual endothelial cells, we developed a novel mathematical approach that allows 3D reconstruction and quantification of both dorsal aorta geometry and endothelial cell surface morphology. Based on fluorescently marked endothelial cell contours, we inferred cross-sections of the dorsal aorta that accounted for dorsal flattening of the vessel. By projection of endothelial cell contours onto the estimated cross-sections and subsequent triangulation, we finally reconstructed 3D surfaces of the individual cells. By simultaneously reconstructing vessel cross-sections and cell surfaces, we found in an exploratory analysis that morphology varied between endothelial cells located in different sectors of the dorsal aorta in both wild-type and Endoglin-deficient zebrafish embryos: In wild-types, ventral endothelial cells were smaller and more elongated in flow direction than dorsal endothelial cells at both 48 hpf and 72 hpf. Although dorsal and ventral endothelial cells in Endoglin-deficient embryos had similar sizes at 48 hpf, dorsal endothelial cells were much larger at 72 hpf. In Endoglin-deficient embryos, elongation in flow direction increased between 48 hpf and 72 hpf in ventral endothelial cells but hardly changed in dorsal endothelial cells. Hereby, we provide evidence that dorsal endothelial cells contribute most to the disparate changes in dorsal aorta diameter in wild-type and Endoglin-deficient embryos between 48 hpf and 72 hpf.

Infrared Bilateral Polarity Ship Detection in Complex Maritime Scenarios.

In complex maritime scenarios where the grayscale polarity of ships is unknown, existing infrared ship detection methods may struggle to accurately detect ships among significant interference. To address this issue, this paper first proposes an infrared image smoothing method composed of Grayscale Morphological Reconstruction (GMR) and a Relative Total Variation (RTV). Additionally, a detection method considering the grayscale uniformity of ships and integrating shape and spatiotemporal features is established for detecting bright and dark ships in complex maritime scenarios. Initially, the input infrared images undergo opening (closing)-based GMR to preserve dark (bright) blobs with the opposite suppressed, followed by smoothing the image with the relative total variation model to reduce clutter and enhance the contrast of the ship. Subsequently, Maximally Stable Extremal Regions (MSER) are extracted from the smoothed image as candidate targets, and the results from the bright and dark channels are merged. Shape features are then utilized to eliminate clutter interference, yielding single-frame detection results. Finally, leveraging the stability of ships and the fluctuation of clutter, true targets are preserved through a multi-frame matching strategy. Experimental results demonstrate that the proposed method outperforms ITDBE, MRMF, and TFMSER in seven image sequences, achieving accurate and effective detection of both bright and dark polarity ship targets.

Host circuit engagement of human cortical organoids transplanted in rodents.

Human neural organoids represent promising models for studying neural function; however, organoids grown in vitro lack certain microenvironments and sensory inputs that are thought to be essential for maturation. The transplantation of patient-derived neural organoids into animal hosts helps overcome some of these limitations and offers an approach for neural organoid maturation and circuit integration. Here, we describe a method for transplanting human stem cell-derived cortical organoids (hCOs) into the somatosensory cortex of newborn rats. The differentiation of human induced pluripotent stem cells into hCOs occurs over 30-60 days, and the transplantation procedure itself requires ~0.5-1 hours per animal. The use of neonatal hosts provides a developmentally appropriate stage for circuit integration and allows the generation and experimental manipulation of a unit of human neural tissue within the cortex of a living animal host. After transplantation, animals can be maintained for hundreds of days, and transplanted hCO growth can be monitored by using brain magnetic resonance imaging. We describe the assessment of human neural circuit function in vivo by monitoring genetically encoded calcium responses and extracellular activity. To demonstrate human neuron-host functional integration, we also describe a procedure for engaging host neural circuits and for modulating animal behavior by using an optogenetic behavioral training paradigm. The transplanted human neurons can then undergo ex vivo characterization across modalities including dendritic morphology reconstruction, single-nucleus transcriptomics, optogenetic manipulation and electrophysiology. This approach may enable the discovery of cellular phenotypes from patient-derived cells and uncover mechanisms that contribute to human brain evolution from previously inaccessible developmental stages.

High-throughput analysis of dendrite and axonal arbors reveals transcriptomic correlates of neuroanatomy

Neuronal anatomy is central to the organization and function of brain cell types. However, anatomical variability within apparently homogeneous populations of cells can obscure such insights. Here, we report large-scale automation of neuronal morphology reconstruction and analysis on a dataset of 813 inhibitory neurons characterized using the Patch-seq method, which enables measurement of multiple properties from individual neurons, including local morphology and transcriptional signature. We demonstrate that these automated reconstructions can be used in the same manner as manual reconstructions to understand the relationship between some, but not all, cellular properties used to define cell types. We uncover gene expression correlates of laminar innervation on multiple transcriptomically defined neuronal subclasses and types. In particular, our results reveal correlates of the variability in Layer 1 (L1) axonal innervation in a transcriptomically defined subpopulation of Martinotti cells in the adult mouse neocortex.

Functional architecture of intracellular oscillations in hippocampal dendrites

Fast electrical signaling in dendrites is central to neural computations that support adaptive behaviors. Conventional techniques lack temporal and spatial resolution and the ability to track underlying membrane potential dynamics present across the complex three-dimensional dendritic arbor in vivo. Here, we perform fast two-photon imaging of dendritic and somatic membrane potential dynamics in single pyramidal cells in the CA1 region of the mouse hippocampus during awake behavior. We study the dynamics of subthreshold membrane potential and suprathreshold dendritic events throughout the dendritic arbor in vivo by combining voltage imaging with simultaneous local field potential recording, post hoc morphological reconstruction, and a spatial navigation task. We systematically quantify the modulation of local event rates by locomotion in distinct dendritic regions, report an advancing gradient of dendritic theta phase along the basal-tuft axis, and describe a predominant hyperpolarization of the dendritic arbor during sharp-wave ripples. Finally, we find that spatial tuning of dendritic representations dynamically reorganizes following place field formation. Our data reveal how the organization of electrical signaling in dendrites maps onto the anatomy of the dendritic tree across behavior, oscillatory network, and functional cell states.

Brain Image Segmentation for Ultrascale Neuron Reconstruction via an Adaptive Dual-Task Learning Network.

Accurate morphological reconstruction of neurons in whole brain images is critical for brain science research. However, due to the wide range of whole brain imaging, uneven staining, and optical system fluctuations, there are significant differences in image properties between different regions of the ultrascale brain image, such as dramatically varying voxel intensities and inhomogeneous distribution of background noise, posing an enormous challenge to neuron reconstruction from whole brain images. In this paper, we propose an adaptive dual-task learning network (ADTL-Net) to quickly and accurately extract neuronal structures from ultrascale brain images. Specifically, this framework includes an External Features Classifier (EFC) and a Parameter Adaptive Segmentation Decoder (PASD), which share the same Multi-Scale Feature Encoder (MSFE). MSFE introduces an attention module named Channel Space Fusion Module (CSFM) to extract structure and intensity distribution features of neurons at different scales for addressing the problem of anisotropy in 3D space. Then, EFC is designed to classify these feature maps based on external features, such as foreground intensity distributions and image smoothness, and select specific PASD parameters to decode them of different classes to obtain accurate segmentation results. PASD contains multiple sets of parameters trained by different representative complex signal-to-noise distribution image blocks to handle various images more robustly. Experimental results prove that compared with other advanced segmentation methods for neuron reconstruction, the proposed method achieves state-of-the-art results in the task of neuron reconstruction from ultrascale brain images, with an improvement of about 49% in speed and 12% in F1 score.

Emodin attenuates hypoxic-ischemic brain damage by inhibiting neuronal apoptosis in neonatal mice

Neonatal hypoxic-ischemic brain damage (HIBD) can lead to mortality and severe neurological dysfunction. Emodin is a natural anthraquinone derivative that is easy to obtain and has good neuroprotective effects. This study aimed to investigate the neuroprotective effect of emodin on neonatal mouse HIBD. The modified Rice–Vannucci method was used to induce HIBD in mouse pups. Eighty postnatal 7-day (P7) C57BL/6 neonatal mice were randomly divided into the sham group (sham), vehicle group (vehicle), and emodin group (emodin). TTC staining and whole-brain morphology were used to evaluate the infarct volume and morphology of the brain tissue. The condition of the neurons was observed through Nissl staining, HE staining, FJC staining, immunofluorescence and Western blot for NeuN, IBA-1, and GFAP. The physiological status of the mice was evaluated using weight measurements. The neural function of the mice was assessed using the negative geotaxis test, righting reflex test, and grip test. TUNEL staining was used to detect apoptosis in brain cells. Finally, Western blot and immunofluorescence were used to detect the expression levels of apoptosis-related proteins, such as P53, cleaved caspase-3, Bax and Bcl-2, in the brain. Experiments have shown that emodin can reduce the cerebral infarct volume, brain oedema, neuronal apoptosis, and degeneration and improve the reconstruction of brain tissue morphology, neuronal morphology, physiological conditions, and neural function. Additionally, emodin inhibited the expression of proapoptotic proteins such as P53, Bax and cleaved caspase-3 and promoted the expression of the antiapoptotic protein Bcl-2. Emodin attenuates HIBD by inhibiting neuronal apoptosis in neonatal mice.

RESPAN: an accurate, unbiased and automated pipeline for analysis of dendritic morphology and dendritic spine mapping.

Accurate and unbiased reconstructions of neuronal morphology, including quantification of dendritic spine morphology and distribution, are widely used in neuroscience but remain a major roadblock for large-scale analysis. Traditionally, spine analysis has required labor-intensive manual annotation, which is prone to human error and impractical for large 3D datasets. Previous automated tools for reconstructing neuronal morphology and quantitative dendritic spine analysis face challenges in generating accurate results and, following close inspection, often require extensive manual correction. While recent tools leveraging deep learning approaches have substantially increased accuracy, they lack functionality and useful outputs, necessitating additional tools to perform a complete analysis and limiting their utility. In this paper, we describe Restoration Enhanced SPine And Neuron (RESPAN) analysis, a new comprehensive pipeline developed as an open-source, easily deployable solution that harnesses recent advances in deep learning and GPU processing. Our approach demonstrates high accuracy and robustness, validated extensively across a range of imaging modalities for automated dendrite and spine mapping. It also offers extensive visual and tabulated data outputs, including detailed morphological and spatial metrics, dendritic spine classification, and 3D renderings. Additionally, RESPAN includes tools for validating results, ensuring scientific rigor and reproducibility.

A Morphological Reconstruction of T-Stems in Proto-Semitic

A Morphological Reconstruction of T-Stems in Proto-Semitic

Electrophysiological Properties and Morphology of Cardiac and Pulmonary Motoneurons within the Dorsal Motor Nucleus of the Vagus of Rats

The dorsal motor nucleus of the vagus (DMV) contains parasympathetic motoneurons that project to the heart and lungs. These motoneurons control ventricular excitability/contractility and airways secretions/blood flow, respectively. However, their electrophysiological properties, morphology and synaptic input activity remain unknown. One important ionic current described in DMV motoneurons controlling their electrophysiological behaviour is the A-type mediated by voltage-dependent K+ (Kv) channels. Thus, we compared the electrophysiological properties, synaptic activity, morphology, A-type current density, and single cell expression of Kv subunits, that contribute to macroscopic A-type currents, between DMV motoneurons projecting to either the heart or lungs of adult male rats. Using retrograde labelling, we visualized distinct DMV motoneurons projecting to the heart or lungs in acutely prepared medullary slices. Subsequently, whole cell recordings, morphological reconstruction and single motoneuron qRT-PCR studies were performed. DMV pulmonary motoneurons were more depolarized, electrically excitable, presented higher membrane resistance, broader action potentials and received greater excitatory synaptic inputs compared to cardiac DMV motoneurons. These differences were in part due to highly branched dendritic complexity and lower magnitude of A-type K+ currents. By evaluating expression of channels that mediate A-type currents from single motoneurons, we demonstrated a lower level of Kv4.2 in pulmonary versus cardiac motoneurons, whereas Kv4.3 and Kv1.4 levels were similar. Thus, with the distinct electrical, morphological, and molecular properties of DMV cardiac and pulmonary motoneurons, we surmise that these cells offer a new vista of opportunities for genetic manipulation providing improvement of parasympathetic function in cardiorespiratory diseases such heart failure and asthma.