Processing And Analysis Of 3D Mouse Brain Image Obtained By Bessel Light-sheet Microscopy

Compared to the traditional epi-illumination fluorescence microscopy,the combination of light-sheet fluorescence microscopy(LSFM)and tissue clearing has the advantage of high throughput,high axial resolution,low phototoxicity,which bring great improvements in imaging speed and spatial resolution to the imaging of sizeable volumes of whole organ samples in the field of anatomy imaging.In addition to the advantages mentioned,compared with conventional Gaussian light sheet microscopy,Bessel light-sheet microscopy has the advantages of both large field of view and high axial resolution,which is benefit to the three dimensional imaging of the whole organ samples.Corresponding to advanced imaging methods,visualization and qualitative / quantitative data analysis based on three dimensional images are the key to connect the acquisition raw images with meaningful biomedical research,and therefore have important research value.Under this background,this paper carries out resolution enhanced processing,visualization and quantitative analysis of mouse brain 3D imaging data based on Bessel light-sheet microscopy,and is committed to providing accurate and legible 3D visualization results and large-scale,high-precision statistical information for brain science.The main contents of this study include:(1)Achieve the isotropic subcellular-resolution whole brain data acquisition with high throughput based on Bessel light-sheet microscopy(THY1-GFP-M transgenic mouse whole brain: imaging magnification: 6.4×;propidine iodide(PI)nuclear-labeled half mouse brain:imaging magnification 8×).And Large raw data are processed by three-dimension stitching,registration and multi-view fusion and reconstruction to obtain THY1-GFP-M sparsely labeled whole mouse brain and PI labeled half brain.Finally,data visualization modeling of whole brain and half brain are completed to observe the details from a global perspective.In addition,the study traces multiple long-distance neuronal projections through different regions in THY1-GFP-M labeled brain.PI nuclear-labeled half brain is also partitioned into several regions and the study makes statistical analysis of the cell number,density in different brain regions as well.(2)For the research needs of higher resolution,this study introduces a method that combines compressed sensing(CS)with Bessel light-sheet microscopy.CS is used to enhance4× resolution in a single dimension with 3D Bessel light-sheet acquisition low-resolution data.Compared with conventional high-magnification acquisition method obtaining image data with the same resolution,the throughput increases about 64 times.And CS is applied to THY1-GFP-M and PI labeled low-resolution data of different brain regions in the mouse brain.Compared with high-magnification acquisition data,the accuracy and effectiveness of the method could be verified.Due to the limitation of calculation speed and the density of the original image,this study only shows the combination of compressive sensing and Bessel light-sheet microscopy being used to smaller regions.In the future,it is expected that compressive sensing can complete more resolution improvement for large blocks of data at faster rate of processing speed.