Cell peruseed . The world's first systematic acquisition of a 3D map of the mouse brain, greatly promoting the study of brain science (worth collecting)

With the development of modern whole brain imaging technology, it is very important to keep brain atlas synchronized with mass data collection.recently, many research institutions are studying the cell types and neural connections of mouse brain, and collecting a large amount of data across modal, spatial scale and multi brain regions.however, successful integration of the above data requires a standard 3D reference map.to this end, Allen Institute of Brain Sciences released the third edition of ccfv3, which can be used to analyze, visualize and integrate multi-modal and multi-scale 3D data sets.current standard 2D mouse brain maps, such as Allen reference map (ARA) and mouse brain stereotactic coordinates (MBSC).the annotation was drawn on a single reference brain slice image in a single plane, and then stained and imaged by light microscope with cell resolution.traditional segmentation of brain regions is based on the staining of cell structure or myelin sheath structure. Different gene expression, connection mode and functional properties of brain regions can also be used for brain division.because each mode may reveal unique features of certain brain regions, when used together, they are expected to greatly improve the structural description.however, there is no systematic integration of multimodal mammalian brain maps.in addition, although the first and second editions of 3D maps have been published by Allen Institute, there are still two main problems: 1) image distortion caused by converting 2D structure of coronal annotation into 3D volume in noncoronary plane; 2) high cell resolution mapping is required for large-scale data generation.the ccfv3 published in this study just solved the above limitations.team Julie A. Harris and Lydia ng from the Allen Brain Science Institute in Seattle, USA, published the title "the Allen mouse brain common coordinate framework: a3d reference" in cell In this paper, we analyzed in detail how ccfv3 divided the mouse brain of each hemisphere into 43 neocortex regions, 329 subcortical gray matter structures, 81 fiber tracts and 8 ventricular structures.due to the subtle differences in size and structure of the brains of each mouse.ccfv3 is a global mean Allen mouse brain connectivity map constructed by serial two-photon tomography (STPT) in 1675 adult mouse brain, which follows the method of human MRI based population mean template.the author uses the iterative process to register the deformation of each sample in the template, takes the average value of all samples, calculates the average deformation field of all samples, and then reverses it and applies it to the created average image.the algorithm continues until the difference of the average amplitude of the average deformation field between iterations drops below a certain threshold and tends to be stable.specifically, figure 1A shows a clear average image converging to a clear anatomical detail. in order to improve the computational efficiency, the method is first applied at 50 μ m resolution until convergence, then it is processed at 25 μ m resolution, and finally 3D atlas model with anatomical structure is presented at 10 μ M. at 10 μ m voxel resolution, the average template contains 506 million voxels (its 3D is 13.2 mm ′ 11.4 mm ′ 8 mm). the final average template contains many distinguishing, detailed and anatomical features. Compared with the fuzzy structure of a single sample STPT, the barrel cortex is very clearly visible in the average template. this shows that the average template provides cell scale resolution, and appropriate 3D reference brain segmentation and annotation. the annotation of the average template brain is completed through a coordinated and iterative workflow. by reviewing previous published maps and literature, the authors visually analyzed the average template and five multimodal reference data sets (STPT imaging of transgenic mice data; axon projection data from Allen linkage map; immunohistochemistry and cell structure staining; in situ hybridization data of Allen mouse brain atlas). ② according to the size and shape of structures, neuroanatomists outline each structure in coronal, sagittal and horizontal planes at certain intervals. ③ the braided structure was filled, refined and smoothed by illustrators, and then verified by neuroanatomists. ④ once a certain amount of 3D reconstruction is achieved, individual and local structures are merged. after merging, small overlaps and gaps are fixed, so all voxels are assigned to a single brain structure. ⑤ voxel labeling was completed and flipped in one hemisphere to generate symmetrical ccfv3, which was finally evaluated by neuroanatomists. draw the boundary of the whole neocortex. the authors used calb1 CRE (CALB expressed in the superficial cortex) and Fezf2 creer (fezf expressed in the deep cortex) mice to distinguish the deep and the superficial cortex; in the coronal section, the dorsal boundary between the cortex and piriform area; in the sagittal plane, the posterior boundary with the posterior inferior support and the anterior area; on the horizontal plane, the posterior boundary with the olfactory bulb. the author draws boundaries directly on these 2D views, as described below to integrate data, and then converts the surface map into 3D volume blocks by extrapolating from the surface along the streamline. firstly, the neocortex was divided into primary visual area, auditory area, somatosensory area and posterior cortex area. then, the four regions were subdivided by transgenic mouse (for example, the visual area was divided into nine small regions). the authors reconstructed 31 neocortical regions from the top view and 33 from the lateral view. all the subcortical areas were divided into three parts. in addition, the neocortex was divided into L1, L2 / 3, L4, L5, l6a and l6b levels by histological data and using five transgenic mice (all with genes specifically expressed in each layer of neocortex). and the 6 layers of these neocortex intersected with 43 cortical regions, forming 242 structural blocks. next, the author described 329 subcortical gray matter structures, including 11 major brain regions. the author still draws the boundary from five multimodal reference data sets. taking 329 structures of IPN as an example, this paper first reviews the existing maps and Literature (IPN is subdivided into 7 / 8 subdomains). The author divides IPN into 8 subdomains by Nissl or antibody staining, transgene expression, axon projection IPN labeling and average template itself. the size, shape and location of the subzones were also described in coronal, sagittal and horizontal planes. most of the voxels in the mean template volume were labeled after subdividing 329 subcortical gray matter structures, and there were some gaps in the gray matter (not labeled voxels). after that, the authors described the white matter tracts by anterograde tracing and antibody staining. By comparing the mean template with the corresponding anterograde tracking data (directly marking the axons that make up these bundles), the contour and trajectory of these white matter tracts can be easily defined. the authors reconstructed 81 fiber bundles in mouse brain by three-dimensional reconstruction, and identified their location, shape, size and trajectory. at the same time, based on the inherent contrast of the average template, combined with the corresponding Nissl staining, the authors delineated the 3D structure of the ventricles, and reconstructed 8 ventricles and related structures. finally, compared with other maps, the author listed some unique structures of ccfv3: IPRL of IPN was described for the first time; the dorsal side of lateral geniculate body (LGD) was further subdivided into LGD sh, LG CO and LG IP; four fiber bundles were added, including visual and auditory radiation, corpus callosum (CCB) and supracallosal white matter (SCWM). What are the advantages of ccfv3? ① The atlas's image viewer provides the ability to view and navigate multiple datasets in a single window, comparing different expression or connection patterns from multiple experiments. ② support the link browsing and visualization of interested structures in ccfv3 ontology tree, or loading 3D models of different structure sets. ③ ccfv3 collected high-resolution images of the whole brain with high signal-to-noise ratio (the spatial resolution of ccfv3 is more than 1000 times that of the second version of Atlas), which makes it easier to distinguish anatomical details. ④ ccfv3 adds two new functions: reverse mapping and volume estimation in the original atlas. ⑤ ccfv3 has a spatial coordinate system and an ordered structure ontology, which can access the atlas interactively with multi-resolution. of course, you can also browse in 2D or 3D mode. Ccfv3 also provides anatomical background for large-scale electrophysiological experiments and wide field calcium imaging of cerebral cortex. annotation workflow for constructing three-dimensional structure (image from cell) in general, we have upgraded the three-dimensional brain atlas of mouse brain anatomy by high-resolution planar serial two-photon tomography of 1675 C57BL / 6J mice. provides a digital open access interactive atlas (full annotated anatomical reference space) for brain researchers. inspiration and questions: 1. the author has upgraded the three-dimensional brain atlas, which greatly facilitates the brain scientists. promoted the fine development of brain science. The authors point out that ccfv3 does not include the data of Ara in situ hybridization and the experiments of a few transgenic lines can only be obtained by downloading. The authors point out that the relationship between the average template and the bregma is unknown, so ccfv3 is not suitable for the determination of stereotactic coordinates. ccfv3 should be used for reference with the traditional two-dimensional atlas. Although mouse is one of the most commonly used experimental animal models, with the refinement of scientific research requirements and some functions of mice can not better simulate human physiological function, we often use rat model and non-human primate model. it is hoped that the Allen Institute of brain science will also devote itself to developing the atlas of the above two animal models in the future. Reference News: detailed explanation of the article album cell close reading | Dr. Yang peiguo found that g3bp1 phase transition mediates the formation of ribonucleoprotein particles. Nature Neuroscience | found that neurons regulating "eating" were also involved in controlling "fighting" behavior. Nature close reading | artifact "was born. 3D tumor cell culture is more suitable for cancer research| It is amazing to find that the immune function of offspring will be damaged under the external pressure during pregnancy. The breakthrough of nature close reading will lead to the breakthrough of cell death in nervous system. It is an important discovery in autophagy field that phospholipid synthesis pathway is the source of membrane phospholipid of autophagy