A multimodal, multidimensional atlas of C57BL/6J mouse brain at development stage P0

The object of this study is to establish an anatomical framework for the neonatal mouse brain which can serve as a standard space for gene expression analysis. The results include a multi-modal 3D digital brain atlas for mouse of postnatal day 0 (P0) and an atlas interface implemented with a atlas-based registration toolset and a data management system. The atlas contains high-resolution anatomical models based on registered histological images, in which the 2D distortion introduced during sample preparation was corrected by registering the images to a MR reference image volume. A standard atlas space with stereotaxic coordinates for the P0 C57BL/6J mouse brain was defined using the average anatomical space of eight individual co-registered MR image volumes. Accuracy of registration and morphometric variations in structures between subjects was analyzed statistically. We also applied this atlas coordinate system to data acquired using different imaging protocols as well as to the reconstructed high-resolution histological atlas obtained from separate animals. Mapping accuracy in the atlas space was examined to determine the applicability of this atlas framework. The results show that the atlas space defined here provides a stable framework for image registration for P0 normal mouse brains. With an appropriate feature-based co-registration strategy, the probability atlas can also provide an accurate anatomical map for images acquired using invasive imaging methods. This high-resolution atlas serves as a framework for data-to-atlas mapping and also facilitates the integration the data from anatomical and genetic databases. A Java-coded atlas interface was implemented to enable atlas-based registration and mapping for the gene expression data images that were acquired without any spatial framework. This interface also provides a management system for representing the spatial relations of multiple datasets and modeling the gene expression pattern the data presented in the atlas space. With the atlas coordinates given to these 3D objects, the user can perform spatial queries over the atlas-associated database. Concurrently displaying the 3D co-expression patterns of multiple registered dataset provides better insight into the interaction of the genes and their roles in brain functions.