The mouse cortico-basal ganglia-thalamic network.

The cortico-basal ganglia-thalamo-cortical loop is one of the fundamental network motifs in the brain. Revealing its structural and functional organization is critical to understanding cognition, sensorimotor behaviour, and the natural history of many neurological and neuropsychiatric disorders. Classically, this network is conceptualized to contain three information channels: motor, limbic and associative1-4. Yet this three-channel view cannot explain the myriad functions of the basal ganglia. We previously subdivided the dorsal striatum into 29 functional domains on the basis of the topography of inputs from the entire cortex5. Here we map the multi-synaptic output pathways of these striatal domains through the globus pallidus external part (GPe), substantia nigra reticular part (SNr), thalamic nuclei and cortex. Accordingly, we identify 14 SNr and 36 GPe domains and a direct cortico-SNr projection. The striatonigral direct pathway displays a greater convergence of striatal inputs than the more parallel striatopallidal indirect pathway, although direct and indirect pathways originating from the same striatal domain ultimately converge onto the same postsynaptic SNr neurons. Following the SNr outputs, we delineate six domains in the parafascicular and ventromedial thalamic nuclei. Subsequently, we identify six parallel cortico-basal ganglia-thalamic subnetworks that sequentially transduce specific subsets of cortical information through every elemental node of the cortico-basal ganglia-thalamic loop. Thalamic domains relay this output back to the originating corticostriatal neurons of each subnetwork in a bona fide closed loop.

Sharing massive biomedical data at magnitudes lower bandwidth using implicit neural function.

Efficient storage and sharing of massive biomedical data would open up their wide accessibility to different institutions and disciplines. However, compressors tailored for natural photos/videos are rapidly limited for biomedical data, while emerging deep learning-based methods demand huge training data and are difficult to generalize. Here, we propose to conduct Biomedical data compRession with Implicit nEural Function (BRIEF) by representing the target data with compact neural networks, which are data specific and thus have no generalization issues. Benefiting from the strong representation capability of implicit neural function, BRIEF achieves 2[Formula: see text]3 orders of magnitude compression on diverse biomedical data at significantly higher fidelity than existing techniques. Besides, BRIEF is of consistent performance across the whole data volume, and supports customized spatially varying fidelity. BRIEF's multifold advantageous features also serve reliable downstream tasks at low bandwidth. Our approach will facilitate low-bandwidth data sharing and promote collaboration and progress in the biomedical field.

Whole-brain mapping of histaminergic projections in mouse brain.

Histamine is a conserved neuromodulator in mammalian brains and critically involved in many physiological functions. Understanding the precise structure of the histaminergic network is the cornerstone in elucidating its function. Herein, using histidine decarboxylase (HDC)-CreERT2 mice and genetic labeling strategies, we reconstructed a whole-brain three dimensional (3D) structure of histaminergic neurons and their outputs at 0.32 × 0.32 × 2 µm3 pixel resolution with a cutting-edge fluorescence microoptical sectioning tomography system. We quantified the fluorescence density of all brain areas and found that histaminergic fiber density varied significantly among brain regions. The density of histaminergic fiber was positively correlated with the amount of histamine release induced by optogenetic stimulation or physiological aversive stimulation. Lastly, we reconstructed a fine morphological structure of 60 histaminergic neurons via sparse labeling and uncovered the largely heterogeneous projection pattern of individual histaminergic neurons. Collectively, this study reveals an unprecedented whole-brain quantitative analysis of histaminergic projections at the mesoscopic level, providing a foundation for future functional histaminergic study.

Cross-modal coherent registration of whole mouse brains.

Recent whole-brain mapping projects are collecting large-scale three-dimensional images using modalities such as serial two-photon tomography, fluorescence micro-optical sectioning tomography, light-sheet fluorescence microscopy, volumetric imaging with synchronous on-the-fly scan and readout or magnetic resonance imaging. Registration of these multi-dimensional whole-brain images onto a standard atlas is essential for characterizing neuron types and constructing brain wiring diagrams. However, cross-modal image registration is challenging due to intrinsic variations of brain anatomy and artifacts resulting from different sample preparation methods and imaging modalities. We introduce a cross-modal registration method, mBrainAligner, which uses coherent landmark mapping and deep neural networks to align whole mouse brain images to the standard Allen Common Coordinate Framework atlas. We build a brain atlas for the fluorescence micro-optical sectioning tomography modality to facilitate single-cell mapping, and used our method to generate a whole-brain map of three-dimensional single-neuron morphology and neuron cell types.

The cytoarchitectonic landscape revealed by deep learning method facilitated precise positioning in mouse neocortex.

Neocortex is a complex structure with different cortical sublayers and regions. However, the precise positioning of cortical regions can be challenging due to the absence of distinct landmarks without special preparation. To address this challenge, we developed a cytoarchitectonic landmark identification pipeline. The fluorescence micro-optical sectioning tomography method was employed to image the whole mouse brain stained by general fluorescent nucleotide dye. A fast 3D convolution network was subsequently utilized to segment neuronal somas in entire neocortex. By approach, the cortical cytoarchitectonic profile and the neuronal morphology were analyzed in 3D, eliminating the influence of section angle. And the distribution maps were generated that visualized the number of neurons across diverse morphological types, revealing the cytoarchitectonic landscape which characterizes the landmarks of cortical regions, especially the typical signal pattern of barrel cortex. Furthermore, the cortical regions of various ages were aligned using the generated cytoarchitectonic landmarks suggesting the structural changes of barrel cortex during the aging process. Moreover, we observed the spatiotemporally gradient distributions of spindly neurons, concentrated in the deep layer of primary visual area, with their proportion decreased over time. These findings could improve structural understanding of neocortex, paving the way for further exploration with this method.

Dissection of the long-range projections of specific neurons at the synaptic level in the whole mouse brain.

Through synaptic connections, long-range circuits transmit information among neurons and connect different brain regions to form functional motifs and execute specific functions. Tracing the synaptic distribution of specific neurons requires submicron-level resolution information. However, it is a great challenge to map the synaptic terminals completely because these fine structures span multiple regions, even in the whole brain. Here, we develop a pipeline including viral tracing, sample embedding, fluorescent micro-optical sectional tomography, and big data processing. We mapped the whole-brain distribution and architecture of long projections of the parvalbumin neurons in the basal forebrain at the synaptic level. These neurons send massive projections to multiple downstream regions with subregional preference. With three-dimensional reconstruction in the targeted areas, we found that synaptic degeneration was inconsistent with the accumulation of amyloid-ß plaques but was preferred in memory-related circuits, such as hippocampal formation and thalamus, but not in most hypothalamic nuclei in 8-month-old mice with five familial Alzheimer's disease mutations. Our pipeline provides a platform for generating a whole-brain atlas of cell-type-specific synaptic terminals in the physiological and pathological brain, which can provide an important resource for the study of the organizational logic of specific neural circuits and the circuitry changes in pathological conditions.

High-definition imaging using line-illumination modulation microscopy.

The microscopic visualization of large-scale three-dimensional (3D) samples by optical microscopy requires overcoming challenges in imaging quality and speed and in big data acquisition and management. We report a line-illumination modulation (LiMo) technique for imaging thick tissues with high throughput and low background. Combining LiMo with thin tissue sectioning, we further develop a high-definition fluorescent micro-optical sectioning tomography (HD-fMOST) method that features an average signal-to-noise ratio of 110, leading to substantial improvement in neuronal morphology reconstruction. We achieve a >30-fold lossless data compression at a voxel resolution of 0.32 × 0.32 × 1.00 µm3, enabling online data storage to a USB drive or in the cloud, and high-precision (95% accuracy) brain-wide 3D cell counting in real time. These results highlight the potential of HD-fMOST to facilitate large-scale acquisition and analysis of whole-brain high-resolution datasets.

Extension of Endocardium-Derived Vessels Generate Coronary Arteries in Neonates.

BACKGROUND: Unraveling how new coronary arteries develop may provide critical information for establishing novel therapeutic approaches to treating ischemic cardiac diseases. There are 2 distinct coronary vascular populations derived from different origins in the developing heart. Understanding the formation of coronary arteries may provide insights into new ways of promoting coronary artery formation after myocardial infarction. METHODS: To understand how intramyocardial coronary arteries are generated to connect these 2 coronary vascular populations, we combined genetic lineage tracing, light sheet microscopy, fluorescence micro-optical sectioning tomography, and tissue-specific gene knockout approaches to understand their cellular and molecular mechanisms. RESULTS: We show that a subset of intramyocardial coronary arteries form by angiogenic extension of endocardium-derived vascular tunnels in the neonatal heart. Three-dimensional whole-mount fluorescence imaging showed that these endocardium-derived vascular tunnels or tubes adopt an arterial fate in neonates. Mechanistically, we implicate Mettl3 (methyltransferase-like protein 3) and Notch signaling in regulating endocardium-derived intramyocardial coronary artery formation. Functionally, these intramyocardial arteries persist into adulthood and play a protective role after myocardial infarction. CONCLUSIONS: A subset of intramyocardial coronary arteries form by extension of endocardium-derived vascular tunnels in the neonatal heart.

Aß42 and ROS dual-targeted multifunctional nanocomposite for combination therapy of Alzheimer's disease.

Amyloid-ß (Aß) readily misfolds into neurotoxic aggregates, generating high levels of reactive oxygen species (ROS), leading to progressive oxidative damage and ultimately cell death. Therefore, simultaneous inhibition of Aß aggregation and scavenging of ROS may be a promising therapeutic strategy to alleviate Alzheimer's disease pathology. Based on the previously developed antibody 1F12 that targets all forms of Aß42, we developed an Aß42 and ROS dual-targeting nanocomposite using biodegradable mesoporous silica nanoparticles as carriers to load ultra-small cerium oxide nanocrystals (bMSNs@Ce-1F12). By modifying the brain-targeted rabies virus glycoprotein 29 (RVG29-bMSNs@Ce-1F12), this intelligent nanocomposite can efficiently target brain Aß-rich regions. Combined with peripheral and central nervous system treatments, RVG29-bMSNs@Ce-1F12 can significantly alleviate AD symptoms by inhibiting Aß42 misfolding, accelerating Aß42 clearance, and scavenging ROS. Furthermore, this synergistic effect of ROS scavenging and Aß clearance exhibited by this Aß42 and ROS dual-targeted strategy also reduced the burden of hyperphosphorylated tau, alleviated glial cell activation, and ultimately improved cognitive function in APP/PS1 mice. Our findings indicate that RVG29-bMSNs@Ce-1F12 is a promising nanodrug that can facilitate multi-target treatment of AD.

Rapid and High-Sensitive Phosphoproteomics Elucidated the Spatial Dynamics of the Mouse Brain.

Recent developments in phosphoproteomics have enabled signaling studies where over 10,000 phosphosites can be routinely identified and quantified. Yet, current analyses are limited in sample size, reproducibility, and robustness, hampering experiments that involve low-input samples such as rare cells and fine-needle aspiration biopsies. To address these challenges, we introduced a simple and rapid phosphorylation enrichment method (miniPhos) that uses a minimal amount of the sample to get enough information to decipher biological significance. The miniPhos approach completed the sample pretreatment within 4 h and high effectively collected the phosphopeptides in a single-enrichment format with an optimized enrichment process and miniaturized system. This resulted in an average of 22,000 phosphorylation peptides quantified from 100 µg of proteins and even confidently localized over 4500 phosphosites from as little as 10 µg of peptides. Further application was carried out on different layers of mouse brain micro-sections; our miniPhos method provided quantitative information on protein abundance and phosphosite regulation for the most relevant neurodegenerative diseases, cancers, and signaling pathways in the mouse brain. Surprisingly, the phosphoproteome exhibited more spatial variations than the proteome in the mouse brain. Overall, spatial dynamics of phosphosites are integrated with proteins to gain insights into crosstalk of cellular regulation at different layers, thereby facilitating a more comprehensive understanding of mouse brain development and activity.

RecV recombinase system for in vivo targeted optogenomic modifications of single cells or cell populations.

Brain circuits comprise vast numbers of interconnected neurons with diverse molecular, anatomical and physiological properties. To allow targeting of individual neurons for structural and functional studies, we created light-inducible site-specific DNA recombinases based on Cre, Dre and Flp (RecVs). RecVs can induce genomic modifications by one-photon or two-photon light induction in vivo. They can produce targeted, sparse and strong labeling of individual neurons by modifying multiple loci within mouse and zebrafish genomes. In combination with other genetic strategies, they allow intersectional targeting of different neuronal classes. In the mouse cortex they enable sparse labeling and whole-brain morphological reconstructions of individual neurons. Furthermore, these enzymes allow single-cell two-photon targeted genetic modifications and can be used in combination with functional optical indicators with minimal interference. In summary, RecVs enable spatiotemporally precise optogenomic modifications that can facilitate detailed single-cell analysis of neural circuits by linking genetic identity, morphology, connectivity and function.

Differential synthetic illumination based on multi-line detection for resolution and contrast enhancement of line confocal microscopy.

Line confocal (LC) microscopy is a fast 3D imaging technique, but its asymmetric detection slit limits resolution and optical sectioning. To address this, we propose the differential synthetic illumination (DSI) method based on multi-line detection to enhance the spatial resolution and optical sectioning capability of the LC system. The DSI method allows the imaging process to simultaneously accomplish on a single camera, which ensures the rapidity and stability of the imaging process. DSI-LC improves X- and Z-axis resolution by 1.28 and 1.26 times, respectively, and optical sectioning by 2.6 times compared to LC. Furthermore, the spatially resolved power and contrast are also demonstrated by imaging pollen, microtubule, and the fiber of the GFP fluorescence-labeled mouse brain. Finally, Video-rate imaging of zebrafish larval heart beating in a 665.6 × 332.8 µm2 field-of-view is achieved. DSI-LC provides a promising approach for 3D large-scale and functional imaging in vivo with improved resolution, contrast, and robustness.

Resolution and uniformity improvement of parallel confocal microscopy based on microlens arrays and a spatial light modulator.

In traditional fluorescence microscopy, it is hard to achieve a large uniform imaging field with high resolution. In this manuscript, we developed a confocal fluorescence microscope combining the microlens array with spatial light modulator to address this issue. In our system, a multi-spot array generated by a spatial light modulator passes through the microlens array to form an optical probe array. Then multi-spot adaptive pixel-reassignment method for image scanning microscopy (MAPR-ISM) will be introduced in this parallelized imaging to improve spatial resolution. To generate a uniform image, we employ an optimized double weighted Gerchberg-Saxton algorithm (ODWGS) using signal feedback from the camera. We have built a prototype system with a FOV of 3.5 mm × 3.5 mm illuminated by 2500 confocal points. The system provides a lateral resolution of ∼0.82 µm with ∼1.6 times resolution enhancement after ISM processing. And the nonuniformity across the whole imaging field is 3%. Experimental results of fluorescent beads, mouse brain slices and melanoma slices are presented to validate the applicability and effectiveness of our system.

Partially interpretable image deconvolution framework based on the Richardson-Lucy model.

Fluorescence microscopy typically suffers from aberration induced by system and sample, which could be circumvented by image deconvolution. We proposed a novel, to the best of our knowledge, Richardson-Lucy (RL) model-driven deconvolution framework to improve reconstruction performance and speed. Two kinds of neural networks within this framework were devised, which are partially interpretable compared with previous deep learning methods. We first introduce RL into deep feature space, which has superior generalizability to the convolutional neural networks (CNN). We further accelerate it with an unmatched backprojector, providing a five times faster reconstruction speed than classic RL. Our deconvolution approaches outperform both CNN and traditional methods regarding image quality for blurred images caused by out-of-focus or imaging system aberration.

Deep background-mismodeling-learned reconstruction for high-accuracy fluorescence diffuse optical tomography.

We present a deep background-mismodeling-learned reconstruction framework for high-accuracy fluorescence diffuse optical tomography (FDOT). A learnable regularizer incorporating background mismodeling is formulated in the form of certain mathematical constraints. The regularizer is then learned to obtain the background mismodeling automatically using a physics-informed deep network implicitly. Here, a deep-unrolled FIST-Net for optimizing L1-FDOT is specially designed to obtain fewer learning parameters. Experiments show that the accuracy of FDOT is significantly improved via implicitly learning the background mismodeling, which proves the validity of the deep background-mismodeling-learned reconstruction. The proposed framework can also be used as a general method to improve a class of image modalities based on linear inverse problems with unknown background modeling errors.

High-fidelity mesoscopic fluorescence molecular tomography based on SSB-Net.

The imaging fidelity of mesoscopic fluorescence molecular tomography (MFMT) in reflective geometry suffers from spatial nonuniformity of measurement sensitivity and ill-posed reconstruction. In this study, we present a spatially adaptive split Bregman network (SSB-Net) to simultaneously overcome the spatial nonuniformity of measurement sensitivity and promote reconstruction sparsity. The SSB-Net is derived by unfolding the split Bregman algorithm. In each layer of the SSB-Net, residual block and 3D convolution neural networks (3D-CNNs) can adaptively learn spatially nonuniform error compensation, the spatially dependent proximal operator, and sparsity transformation. Simulations and experiments show that the proposed SSB-Net enables high-fidelity MFMT reconstruction of multifluorophores at different positions within a depth of a few millimeters. Our method paves the way for a practical reflection-mode diffuse optical imaging technique.

Characterization of N-glycome profile in mouse brain tissue regions by MALDI-TOF/MS.

Glycosylation is one of the most common types of post-translational modifications in mammals. It is well known that N-glycans play a key role in cell adhesion, differentiation, synapsis, and myelination during the development of the mammalian central nervous system (CNS). Neuropathological symptoms (such as epilepsy and Alzheimer's disease) are usually accompanied by N-glycosylation changes. In this study, we extracted N-glycan chains from eight regions of the mouse brain, and combined high-throughput, high-resolution matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF/MS) with the Fmoc N-hydroxysuccinimide ester (Fmoc-OSU) derivatization method to improve the sensitivity of glycan detection to characterize the total N-glycans in the mouse brain. A total of 96 N-glycan moieties were detected. An exhaustive examination of the relative abundance of N-glycans, coupled with a comparative analysis of differences, has uncovered discernible variations of statistical significance, including high mannose, fucosylated, sialylated, and galactosylated N-glycans. According to our investigations, a thorough and regionally specific cartography of glycans within the brain can facilitate the investigation of glycan-mediated mechanisms related to both the developmental trajectory and functional output of the brain. Additionally, this approach may serve as a basis for identifying potential biomarkers that are relevant to various brain-associated pathologies.

Cerebral Organoid Arrays for Batch Phenotypic Analysis in Sections and Three Dimensions.

Organoids can recapitulate human-specific phenotypes and functions in vivo and have great potential for research in development, disease modeling, and drug screening. Due to the inherent variability among organoids, experiments often require a large sample size. Embedding, staining, and imaging each organoid individually require a lot of reagents and time. Hence, there is an urgent need for fast and efficient methods for analyzing the phenotypic changes in organoids in batches. Here, we provide a comprehensive strategy for array embedding, staining, and imaging of cerebral organoids in both agarose sections and in 3D to analyze the spatial distribution of biomarkers in organoids in situ. We constructed several disease models, particularly an aging model, as examples to demonstrate our strategy for the investigation of the phenotypic analysis of organoids. We fabricated an array mold to produce agarose support with microwells, which hold organoids in place for live/dead imaging. We performed staining and imaging of sectioned organoids embedded in agarose and 3D imaging to examine phenotypic changes in organoids using fluorescence micro-optical sectioning tomography (fMOST) and whole-mount immunostaining. Parallel studies of organoids in arrays using the same staining and imaging parameters enabled easy and reliable comparison among different groups. We were able to track all the data points obtained from every organoid in an embedded array. This strategy could help us study the phenotypic changes in organoids in disease models and drug screening.

Large depth-of-field fluorescence microscopy based on deep learning supported by Fresnel incoherent correlation holography.

Fluorescence microscopy plays an irreplaceable role in biomedicine. However, limited depth of field (DoF) of fluorescence microscopy is always an obstacle of image quality, especially when the sample is with an uneven surface or distributed in different depths. In this manuscript, we combine deep learning with Fresnel incoherent correlation holography to describe a method to obtain significant large DoF fluorescence microscopy. Firstly, the hologram is restored by the Auto-ASP method from out-of-focus to in-focus in double-spherical wave Fresnel incoherent correlation holography. Then, we use a generative adversarial network to eliminate the artifacts introduced by Auto-ASP and output the high-quality image as a result. We use fluorescent beads, USAF target and mouse brain as samples to demonstrate the large DoF of more than 400µm, which is 13 times better than that of traditional wide-field microscopy. Moreover, our method is with a simple structure, which can be easily combined with many existing fluorescence microscopic imaging technology.

Interpretable model-driven projected gradient descent network for high-quality fDOT reconstruction.

In fluorescence diffuse optical tomography (fDOT), the quality of reconstruction is severely limited by mismodeling and ill-posedness of inverse problems. Although data-driven deep learning methods improve the quality of image reconstruction, the network architecture lacks interpretability and requires a lot of data for training. We propose an interpretable model-driven projected gradient descent network (MPGD-Net) to improve the quality of fDOT reconstruction using only a few training samples. MPGD-Net unfolds projected gradient descent into a novel deep network architecture that is naturally interpretable. Simulation and in vivo experiments show that MPGD-Net greatly improves the fDOT reconstruction quality with superior generalization ability.