Waxholm Space atlas of the rat brain: A 3D atlas supporting data analysis and integration

doi:10.21203/rs.3.rs-2466303/v1

Download PDF

Resource

Waxholm Space atlas of the rat brain: A 3D atlas supporting data analysis and integration

https://doi.org/10.21203/rs.3.rs-2466303/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 02 Oct, 2023

Read the published version in Nature Methods →

Version 1

posted

You are reading this latest preprint version

Volumetric brain atlases are increasingly used to integrate and analyse diverse experimental neuroscience data acquired from animal models, but until recently a publicly available digital atlas with complete coverage of the rat brain has been missing. Here we present the new Waxholm Space rat brain atlas, a comprehensive open-access volumetric atlas resource. This full brain atlas features annotations of 222 structures, of which 112 are new and 57 revised compared to previous versions. It maps in detail the cerebral cortex, hippocampal region, striatopallidal areas, midbrain dopaminergic system, thalamic cell groups, the auditory system and main fibre tracts. We document the criteria underlying the new annotations and demonstrate how the atlas and related tools and workflows can be used by neuroscientists to support interpretation, integration, analysis, and dissemination of experimental rat brain data.

Biological sciences/Neuroscience/Computational neuroscience

Biological sciences/Neuroscience/Cellular neuroscience

Progress in neuroscience increasingly requires integrated analysis of large amounts of data acquired with a broad range of methods spanning different spatial and temporal scales. Structural brain atlases are key spatial reference tools in this endeavor. These atlases provide state-of-the-art anatomical ontologies and delineations defined in a representative graphical brain data set, accompanied with a coordinate system, suitable for indexing research findings. They allow accurate visualization and navigation of brain tissue structures, and provide a platform for the consistent comparison across datasets in the same model species. Three-dimensional (3D) digital atlases introduced over the past decade have provided several important advantages over the conventional serial section atlases. They cover the entire brain volume, and not only a selected set of 2D planes. Furthermore, in combination with a growing suite of software tools, they allow dynamic viewing and analysis, e.g., data visualization of atlas structures¹, image registration to atlas^2,3, and spatial analysis of features in images^4–6.

For analysis and interpretation of murine brain data, a range of atlas resources are available. For the mouse, the Common Coordinate Framework, CCFv3⁷ (https://portal.brain-map.org/#) has become a widely used resource for placing mouse brain image data in a common reference space^2,3,8,9, for analysis and computational modelling^10–13, as well as for the indexation of neuronal morphologies and gene expression data^14,15. For the rat, the most detailed open access 3D atlas is the Waxholm Space atlas of the Sprague Dawley rat brain (WHS rat brain atlas; RRID: SCR_017124)^16–18. This atlas has been adopted by research services such as AFNI¹⁹ and EBRAINS (https://ebrains.eu/service/rat-brain-atlas). Previous versions of the WHS rat brain atlas featured a high-level delineation of major brain regions¹⁶, but detailed, functionally relevant delineations were limited to hippocampal and parahippocampal regions¹⁷, and the brainstem auditory system¹⁸. Despite this limitation, the atlas has already been employed in a wide range of studies, including interpretation of functional neuroimaging data^20,21; reconstruction of fibre tracts and receptor densities²²; visualization of recording electrodes²³; and brain-wide analysis of microscopic features²⁴.

Here, we present a new full brain coverage version (Version 4) of the WHS rat brain atlas. New, detailed 3D brain structure annotations were created by combining interpretation of the MRI/DTI data sets of the atlas template with multimodal cyto-, chemo-, and myeloarchitecture, other reference atlases, and literature. Comprehensive subdivisions of the cerebral cortex, striatopallidal region, midbrain dopaminergic system, and thalamus are now included, in addition to a full revision to the previous anatomical delineations. We present explicit criteria for structure identification, with an emphasis on consistency with previous literature. We also provide a consistent hierarchical organization scheme, version-specific statistics, and detailed metadata. Lastly, we provide examples of how the atlas, can be incorporated in several tools and workflows for analysis, integration and interpretation of a broad range of rat brain data.

The Waxholm Space (WHS) rat brain atlas is the most comprehensive open access volumetric rat brain atlas available, covering all major systems of the brain with detailed delineations. Below, we discuss its key features regarding terminology, structural organization, and annotations of the atlas, before exemplifying how it can be utilized for integrating and analysing experimental rat brain data. More detailed descriptions of the procedures involved in the creation of the digital atlas, including choice of terminology and procedures and criteria used to identify and delineate anatomical regions are provided in the Methods section.

Key features of the atlas

The atlas is based on an isotropic, high-resolution, contrast enhanced sMRI/DTI data set acquired ex vivo from an adult male Sprague Dawley rat brain¹⁶. In this reference data set, a “Waxholm Space” coordinate system²⁷ was defined using internal brain landmarks, and within this spatial reference framework, several sets of annotations have been defined and released as incremental versions of the atlas^16–18,28 with increasing detail.

Terminology and structural hierarchy of the atlas

The terminology used in our new atlas (Supplementary File 1) is largely compatible with the one used by Paxinos and Watson^29,30, but has at several points been adapted to be compatible with terms more commonly used in the field. The structure terms were hierarchically organized to facilitate dynamic visualization and analyses of brain regions and system at different levels of granularity. At the highest level of the hierarchy, the brain is subdivided into white matter, grey matter, ventricular system, and spinal cord. Grey matter structures are hierarchically sorted according in five main domains according to the embryonic neural tube segments from which they derive, namely the telencephalon, diencephalon, mesencephalon, metencephalon, and myelencephalon.

Within each vesicular node, we sorted cortical areas, brain regions, and subregions into a consistent hierarchy of structures that are largely consistent with terminologies used in other atlases^7,30,31. The telencephalon domain is divided into the externally located pallium (further divided into laminated and non-laminated regions), and the deeply located subpallium. The laminated pallium is here divided into the cerebral cortex and the olfactory bulb, while the subpallium consists of the striatum, pallidum and basal forebrain. The diencephalon domain is divided into the pre-, epi- and dorsal thalamus, the hypothalamus, and the pretectum. The mesencephalon domain is the midbrain, while the rhombencephalon is divided into the metencephalon and myelencephalon.

Brain region annotations of the atlas

The WHS rat brain atlas v4 features 222 brain region annotations covering all major brain systems. All region annotations were manually delineated in 3D, based on interpretation of sMRI/DTI signal contrast. Interpretation of signals in the reference data was aided by morphological observations of cyto-, myelo-, or chemoarchitecture in histological section images spatially registered to the MRI reference data, combined with positional information derived from other reference atlases and published data. Fifty-seven annotations are revised and 112 are new in the present version^16–18. The new annotations include detailed subdivisions of the cerebral cortex, striatopallidal region, midbrain dopaminergic system, and thalamus (Fig. 1 and 2). A listing of all annotations is provided in Supplementary File 1, and information about the provenance of each structure through the different versions is available together with the atlas files on the Neuroimaging Tools and Resources Collaboratory (NITRC; https://www.nitrc.org/projects/whs-sd-atlas/).

Within its telencephalon domain, the atlas contains detailed delineation of the cerebral cortex, with subdivisions in the cingulate, frontal, parietal, occipital and temporal cortical regions (Fig. 1) based on information from literature describing structural and functionally defined maps^32–36, and extrapolation of borders from 2D atlases^29,31,37. The olfactory system is represented with the olfactory bulb and its glomerular layer delineated, in addition to the piriform cortex and nucleus of the lateral olfactory tract. Three structures are included in the non-laminated pallium, adjacent to the cerebral cortex: the claustrum, endopiriform nucleus, and the amygdaloid area. In the subpallium, the atlas contains detailed delineation of striatopallidal regions (Fig. 2) defined by combined use of the reference data and histological images. The atlas also contains detailed subdivisions of the hippocampal and parahippocampal regions¹⁷ based on observed sMRI/DTI contrast, corresponding to histologically defined borders³⁸.

In its diencephalon domain, the thalamus, hypothalamus and pretectum are subdivided into a total of 62 nuclei (Fig. 2), whose boundaries are defined by consistent differences in the high-resolution MRI signal of the grey matter as well as by the 3D geometry of several white matter tracts coursing through the thalamus. The delineation of thalamic subregions is largely compatible with the annotations of Paxinos and Watson^29,30, except in the posterior thalamus where our atlas features more detailed annotations.

In the midbrain domain, the atlas contains detailed delineations of the superior and inferior colliculi^16,18. The atlas also covers the ventral tegmental area and peripeduncular nucleus, and has detailed subdivisions of the substantia nigra, with the reticular, compact and lateral parts included. In the rhombencephalon domain, the atlas includes the cerebellum and the pontine nuclei, as well as detailed subdivisions of the nuclei of the ascending auditory system ¹⁸.

The white matter delineations include all major tracts of the brain¹⁶, with particular detail for the thalamus, ascending auditory system¹⁸, and hippocampal region¹⁷.

Using the atlas

The WHS rat brain atlas is provided as an open-access resource that can be downloaded from the NeuroImaging Tools & Resources Collaboratory (NITRC; https://www.nitrc.org/). Download options include individual image files (NiFTi format) suitable for use with a range of standalone neuroimaging tools, or ready-to-use file bundles with labels and configuration files for the software tools ITK-snap³⁹ (http://www.itksnap.org), the Mouse BIRN Atlasing Toolkit (MBAT; https://loni.usc.edu/research/software), and the Biomedical Imaging Quantification tool PMOD (https://www.pmod.com/). The open-source ITK-snap tool for delineation of medical images³⁹ (http://www.itksnap.org) was used to develop the atlas and is well suited for inspection of the atlas reference data and annotations, and for 3D rendering of atlas structures with custom colour and transparency level (Fig. 2 and 3). The atlas is also embedded in several neuroinformatics tools allowing users to explore and visualize the atlas, or to utilize it for integration and analysis of experimental rat brain data, as exemplified below.

Exploration and visualization of atlas structures

Our atlas files (see Data availability, below) include everything sufficient and necessary to implement the atlas in web-based or standalone tools, or for opening the atlas in existing tools (see above). In addition, various online tools exist to allow exploration and visualization of the atlas and its structures, bypassing the need to download individual files and thus further increasing the accessibility of the atlas for a broad range of users. For example, online exploration of all versions of the WHS rat brain atlas is offered through different viewers available through the EBRAINS research infrastructure. The Interactive Atlas Viewer (https://interactive-viewer.apps.hbp.eu/#/) displays the sMRI reference data with annotation overlay in the three standard planes, features arbitrary slicing of the atlas in any angle of view, and allows 3D surface rendering view of selected brain structures (Fig. 2 and 3). The viewer also allows users to navigate the atlas spatially using WHS or voxel coordinates, or semantically using the hierarchical atlas terminology. The Interactive Atlas Viewer is connected to query tools providing access to experimental data registered to the atlas and publicly shared via the EBRAINS Knowledge Graph. Such shared data sets provide another useful way to explore the WHS rat brain atlas: through links provided in the EBRAINS infrastructure data cards, users can launch the LocaliZoom viewer to show histologically and immunohistochemically stained data with atlas overlay images^40–42. The atlas overlay images can be displayed as contour lines or as filled contours at different levels of opacity. The viewer provides mouse-over information about structure names and information about mouse pointer position in WHS or stereotaxic coordinates. Lastly, the web application MeshView (https://www.nesys.uio.no/MeshView/, RRID:SCR_017222) allows 3D rendering of individual atlas structures, as well as display of point data coordinates. Together, this suite of online tools provides users with interactive options for exploring the atlas, either separately, in context of spatially registered data, or through data derived by using atlas-based tools.

Integration and analysis of data using the WHS rat brain atlas

The atlas provides a complete framework for integration of heterogeneous neuroscience data. Experimental rat brain data can be integrated by spatial registration of 2D or 3D images to the atlas based on image contrast or anatomical features. Such spatial registration can be achieved by use of a range of tools, typically specialized for the registration of either volumetric^43–46 or 2D histological^2,11,47 image data. Having experimental data registered to the atlas enables researchers to interpret their data in context of brain regions defined by the atlas annotations (Fig. 3c), and to extract spatial information about specific locations of interest in their data, as single or multiple points of interest, such as the location of labelled cells (Fig. 3e) or boundaries indicating regions of interest (Fig. 4a).

Semi-automatic analysis across the brain with the EBRAINS QUINT workflow

As a volumetric, open-source atlas with highly detailed annotations, the WHS rat brain atlas is well-suited for defining anatomical locations for brain-wide quantitative analyses. There are several combinations of tools that allows such analyses^12,48, one of which is the EBRAINS tools for spatial registration (QuickNII and VisuAlign²) and quantification (Nutil^4,49). These tools are included in the QUINT workflow for quantification of features of interest in microscopic images⁴. To demonstrate the practical value of using WHSv4 for brain-wide cell counting, we here re-analysed an existing data set showing parvalbumin neurons across the rat brain^24,50 using the QUINT workflow. The results of this analysis, shown in Figure 4, demonstrates that detailed quantitative analyses can now be performed across the brain, whereas previously only the hippocampal region was delineated to the level of detail required for such an analysis²⁴.

The Waxholm Space (WHS) rat brain atlas provides neuroscientists with a detailed and complete volumetric open-source reference atlas. The atlas has been embedded in tools and workflows offered by the EBRAINS research infrastructure, as well as by other providers of analytical tools and services. It supports research requiring data integration and analysis across a large number of investigations. Below, we first discuss some of the challenges with the creation of the atlas, before summarizing the opportunities it provides for open science.

While the new atlas covers most of the forebrain with high granularity, only the outer boundaries are included for the hypothalamus and amygdala. We chose not to subdivide these regions in the atlas because the reference MRI data lack the necessary structural resolution here. We recommend that investigators add subregions in such structures using their own criteria and literature definitions⁵¹, while referring to the spatial coordinates of the atlas to facilitate interpretation and opportunities for data reuse.

In conventional 2D atlases, uncharted regions are often left blank or annotated only with names, and not complete borders. This is not an option for volumetric atlases, since all voxels within the brain volume needs to be assigned a label. Particular challenges include: 1) voxels that cannot be unambiguously assigned to one of two subregions and 2) remaining voxels from a collective region where some parts have been subdivided. One possible solution is to assign such voxels with a label corresponding to a region higher up in the hierarchy (i.e. the “parent” region)⁷. For example, in the olfactory bulb of the WHS atlas, some subregions have been delineated (including the glomerular layers of the olfactory and accessory olfactory bulb, and the nucleus of the lateral olfactory tract). The remaining part could not be unambiguously assigned to layers at the same level of detail and could be labelled “olfactory bulb”. However, by naming these voxels by the name “olfactory bulb”, the user might get the false impression that they represent the entire structure, which is not the case due to the subregions that have been delineated separately. We have therefore chosen to give such remaining parts of larger areas the extension “unspecified” (in this case, “Olfactory bulb, unspecified”), clarifying that these voxels belong to a parent structure but do not represent it in its entirety. Another challenge with these regions is that they often span hierarchical levels and thus cannot be accurately sorted to a single node; however, they do need to be incorporated in the hierarchy for the atlas to be compatible with tools. In these cases, which include the “brainstem, unspecified”, “amygdaloid area, unspecified”, and “basal forebrain, unspecified”, we have simply chosen to place them at one of the alternative levels.

The anatomical structures in previously published versions of the WHS rat brain atlas have largely been delineated based on distinct sMRI/DTI signal intensity differences, using the white matter tracts separating and encapsulating many nuclei as robust and reproducible starting points for the delineations. In several of the gray matter regions delineated for version 4, the sMRI/DTI data provided insufficient signal intensity differences for identification of changes in the neuroarchitecture. In these regions, the location of borders were identified aided by information registered to the atlas from histological section images, e.g., showing the cytoarchitectonic organization, from other reference atlases, or from published maps containing stereotaxic coordinates. In this way, the delineations of the WHS rat brain atlas v4 builds on several sources of anatomical information. While contrast-based landmarks are generally reproducible and accepted as meaningful criteria for defining brain regions, other criteria, as exemplified above, are typically more arguable and often interpreted differently among domain experts. We have therefore provided explanations on how the borders in the atlas were established (see Methods). It is widely acknowledged that anatomical interpretations vary among experts. Consequently, information about spatial coordinates in addition region names improves the specificity of communicated anatomical results, particularly for smaller subregions of the atlas.

Open access, digital 3D brain atlases are used to plan and perform neuroscientific experiments and analyses^4,14,52–54, visualize and disseminate data^24,55, as well as integrate new data with already available data registered to the same atlas. The WHS atlas supports these functionalities through its compatibility with analytical workflows^4,56,57. More detailed studies employing these workflows have been described previously^24,58. The open access sharing of the atlas is also an important feature allowing anyone to implement it in tools and workflows developed independently of the atlas itself^1,23,59,60, which will facilitate and incentivize the integration of a broad range of rat brain data to the atlas. Such integration of data using the WHS rat brain atlas has practical value both for the individual researcher and the neuroscience community at large. Individual researchers can visualize their data and ensure continuity across their projects. When researcher share their spatially integrated data, the community will benefit from increased access to important heterogeneous data.

In conclusion, the WHS rat brain atlas is an open and accurate neuroanatomical resource as well as a tool supporting analysis and integration of broad range of rat brain data. Motivated by the success stories of our users, we remain committed to further development of the atlas, not at least to solve problems related to uncharted regions and alternative interpretations. We envision further integration with tools for data analysis relying on precise anatomical information.

Acknowledgements

We thank Jala Imad and Arvind E. Wennberg for valuable contributions to early versions of the anatomical delineations for the WHS rat brain atlas v4, Ulrike Schlegel and Ingrid Reiten for assistance with sharing data sets via the EBRAINS Data and Knowledge services, and Gergely Csucs and Eszter A. Papp for expert technical assistance. We also thank Maja A. Puchades, Sharon C. Yates, Nicolaas Groeneboom, and Xiaoyun Gui for implementation of the atlas in tools and viewers. This work was funded by the European Union’s Horizon 2020 Framework Programme for Research and Innovation under the Specific Grant Agreement No. 785907 (Human Brain Project SGA2) and Specific Grant Agreement No. 945539 (Human Brain Project SGA3), and The Research Council of Norway under Grant Agreement No. 269774 (INCF Norwegian Node). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data availability

All data generated or analysed in this study are included in this article or available through the EBRAINS Knowledge Graphe^{40–42,50,74–76} and the Neuroimaging Tools and Resources Collaboratory (NITRC; https://www.nitrc.org/projects/whs-sd-atlas/).

The new version of the WHS atlas of the Sprague Dawley rat brain (version 4) is shared on the atlas home page through NITRC and consist of three files:

WHS_SD_rat_atlas_v4.nii.gz; Volumetric atlas file of 222 anatomical structures.
WHS_SD_rat_atlas_v4.label; Label file specifying the ID, colour code, and name of each anatomical structure.
WHS_SD_rat_atlas_v4_PMOD.zip; MBAT-ready atlas with label (.ilf) and startup file (.atlas).

Creation of the WHS rat brain atlas

The Waxholm Space (WHS) rat brain atlas consist of four parts: a reference data set, a coordinate system, an annotation set, and a terminology (Fig. 5a). The reference data set is volumetric and has a coordinate system based on internal landmarks, 222 three-dimensional annotations representing brain regions manually drawn. The terminology is largely consistent with existing reference atlases but also includes additional expert neuroanatomist opinions. In the following, we present the methods for the generation of the different parts of the atlas.

Reference data and coordinate system

The reference data used as a basis for the WHS rat brain atlas comprise high-resolution, contrast-enhanced T₂*-weighted gradient recalled echo anatomic images with an isotropic spatial resolution of 39 μm (sMRI), and colour-coded principal diffusion direction maps (DTI) with a resolution of 78 μm (Fig. 5a1). These data were acquired ex vivo from the head of an adult male Sprague Dawley rat (age 80 days, weight 397.6 g, Charles River, Wilmington, MS, USA), perfusion-fixed using a mixture of formalin and a gadolinium-based MRI contrast agent (Johnson et al., 2002). Imaging was performed using a 7T small animal MRI system (Magnex Scientific, Yarnton, Oxford, UK) at the Duke Centre for In Vivo Microscopy (Durham, NC, USA). Technical details are provided in Papp et al. (2014, 2015).

The WHS is a continuous, three-dimensional Cartesian coordinate system with the origin defined at the decussation of the anterior commissure (Fig. 5a2). The origin of the WHS, as first described by Papp and colleagues (page 376)¹⁶: “at the intersection of: a) the mid-sagittal plane, b) a coronal plane passing midway (rostro-caudal) through the decussation of the anterior and posterior part of the anterior commissure, and c) a horizontal plane passing midway through the most dorsal and ventral aspect of the decussation of the anterior commissure”. The stereotaxic orientation of the WHS is defined in the flat skull position with the lambda and bregma landmarks at the same height. The WHS reference data deviates 4 degrees from the flat skull position.

Terminology and structural hierarchy

The terminology used in the first version of the WHS rat brain atlas¹⁶ is based on the standard rat brain atlases from Paxinos and Watson (version 6) and Swanson (version 3)^29,61. Subsequent revisions of the atlas incorporated domain specific nomenclatures for the annotations of the hippocampal formation and parahippocampal region⁶², and sought consensus among several sources for structures of the auditory system^29,63,64. The terminology used for new structures in cortical, thalamic, striatopallidal, and midbrain regions (Supplementary File 1) is largely compatible with the terminology of Paxinos and Watson^29,30, but have at several points been adapted to be more compatible with the terminology commonly used by domain experts. This particularly concerns the orbitofrontal, posterior parietal and insular cortex, which reflect the terminology used in seminal papers on the prefrontal cortex in several species^65–67. For striatopallidal and midbrain dopaminergic regions, the terminology used by Paxinos and Watson has changed considerably across atlas versions, and our terminology is mostly consistent with their most recent MRI-based atlas³⁷. For the subdivisions of the thalamus most of our terms are compatible with Paxinos and Watson^29,30, with addition of some terms used by domain experts for structures in the posterior thalamus⁶⁸, where the Paxinos and Watson atlas is less detailed. The hierarchical organization was based on the five secondary brain vesicles, with cortical areas, brain regions, and subregions for each vesicle sorted into a hierarchy of structures that largely followed the terminologies used in other atlases^7,30,31, in particular the the Allen Mouse brain Common Coordinate Framework⁷. The hierarchy is provided as a structured .ilf file.

Annotation set and delineation procedure

All anatomical delineation (Fig. 5a3) are manually drawn based on a combination of features in the reference data, histological material, and information available in the literature. Below, we describe the general procedure of the delineation process, before giving a brief summary of the annotations from previously published versions of the atlas. Lastly, we summarize the annotations created for version 4 of the atlas.

General delineation procedure. Annotations in the WHS rat brain atlas were made using the ITK-SNAP software (version 3.6.0; http://itksnap.org³⁹). Delineations were made in the principal orthogonalplanes (coronal, sagittal, and horizontal) using three approaches (Fig. 5b): 1) interpretation of sMRI/DTI contrast in the reference data; 2) inspection of histological reference material showing cyto-, chemo- or myeloarchitecture; and 3) consultation of literature, expert knowledge, and other brain atlases. These three approaches have been used for all previous versions of the atlas, as well as for the new version 4, and are briefly described below.

When interpreting sMRI/DTI contrast in the reference data, distinct contrast between coherent axonal fibre bundles and surrounding cell-rich areas forms a starting point for the delineation process by distinguishing between white and grey matter. The T₂*-weighted sMRI greyscale maps show high grey to white matter contrast and highlights several cytoarchitectonic features. In the DTI maps, the prevailing orientation and magnitude of water diffusion in each DTI voxel is represented by the red, green, and blue colours, each signifying diffusion in one of the three principal directions (red = mediolateral, green = rostrocaudal, blue = dorsoventral). Intermediate diffusion orientations are indicated by intermediate colours according to the RGB model. The brightness of the colours represents the magnitude of oriented diffusion in each voxel, as determined by fractional anisotropy (FA) values⁶⁹. Fluid-containing spaces appear white in T2*-weighted sMRI and dark in DTI, while air-filled regions are black in both modalities. Combined interpretation of sMRI and DTI maps viewed in coronal, sagittal, and horizontal planes allows identification of a large number of regional and subregional borders, aided by interactive inspection of surface rendering of structures to ensure smooth and 3D coherent surfaces.

At locations where the sMRI/DTI images is difficult to interpret, images of histological images showing myeloarchitecture and cytoarchitecture^40,41,70 aid more detailed interpretations of the sMRI/DTI images, and refinement of border delineation. These public images have been spatially registered to the WHS rat brain atlas, and are available for inspection in an online image viewer, together with reference atlas overlay images^40–42. Thus, observed features in the reference data can be interpreted in comparison with cyto- and myelin-stained histological sections, as previously described¹⁸.

Lastly, the delineation process is aided by consulting the literature, standard reference atlases^29,61, and neuroanatomy experts. Key literature sources include studies of Sprague Dawley rats describing cyto-, chemo, or myelo-architectural characteristics of brain regions, with specific delineation criteria and clear descriptions of borders between regions^71–73. For additional validation of the delineations, we used datasets containing 2D atlas plates from three reference atlases^29,31,37 spatially registered to our volumetric reference data set^74–76. This allowed us to directly compare spatial correspondences of anatomical landmarks and structural delineation across atlases in the standard planes provided by the 2D atlas plates.

Annotations from previously published versions. The annotation set in version 1 of the atlas included major anatomical structures that were identified from readily distinguishable differences in sMRI and DTI signals¹⁶. For version 2 of the atlas, detailed subdivisions in the hippocampus and parahippocampal region were identified based on observed sMRI/DTI contrast, corresponding to histologically defined borders¹⁷. For version 3 of the atlas, the complete ascending auditory system was delineated based on the interpretation of image features in the WHS rat brain sMRI/DTI data, and validated in relation to spatially corresponding images cell- and myelin-stained histological sections¹⁸.

New annotations in version 4. To create the new annotations, we used the following files from the atlas home page on NITRC (https://www.nitrc.org/projects/whs-sd-atlas/):

WHS_SD_rat_T2star_v1.01.nii.gz; Image template showing anatomical MRI, T₂*- weighted gradient echo image at 39 μm original resolution (1024x512x512 voxels).
WHS_SD_rat_DWI_v1.01.nii.gz; Image template showing diffusion weighted image map (DWI), resampled to 39 µm resolution (1024x512x512 voxels).
WHS_SD_rat_FA_color_v1.01.nii.gz; Image template of diffusion tensor (DTI) showing colour Fractional Anisotropy (FA) map resampled to 39 μm resolution (1024x512x512 voxels).
WHS_SD_rat_atlas_v3.nii.gz; Volumetric atlas file containing 118 anatomical structures.
WHS_SD_rat_atlas_v3.label; Label file specifying the ID, colour code, and name of each anatomical structure.

We delineated new regions across the brain, focusing on the cerebral cortex, striatopallidal region, thalamus, and midbrain dopaminergic regions, using the criteria outlined below.

Cerebral cortex.We added 36 new structures in the cerebral cortex. The cerebral cortex appears highly homogeneous in the sMRI/DTI data (Fig. 5), which has insufficient spatial resolution to resolve finer details of cortical layers and subtle cytoarchitectonic borders. We therefore used the cortical delineation of the 6th edition of the atlas by Paxinos and Watson²⁹, the 4th edition of Swanson’s atlas³¹, and the MRI based atlas by Paxinos and colleagues³⁷ as starting points. Diagrams from these atlases have been spatially registered to the WHS reference data (v1.01)^74–76, allowing co-visualization and comparison of the different cortical delineation from these atlases to anatomical features visible in the sMRI/DTI data. This allowed us to identify the approximate stereotaxic location of cortical areas in coronal, sagittal, and horizontal levels in the sMRI/DTI data, and create the initial delineation. For the somatomotor and somatosensory areas, the locations and shapes of borders were adjusted to stereotaxic positions transferred from published maps defined by electrophysiological measurements^{33–36,77,78}. For areas of the orbitofrontal, insular and posterior parietal cortex, the delineation were adapted in three planes on the basis of descriptions from previous anatomical studies^79–81. Visual areas were adjusted in accordance with additional literature^73,82, abutting the boundaries of the parahippocampal region¹⁷ (v2) and the auditory cortex¹⁸ (v3). In this way, the new delineations of cortical areas represent a composite of several sources, where most of the delineation are largely compatible with other reference atlases and published cortical maps. Adjacent to the cerebral cortex we defined three new structures, the claustrum, endopiriform nucleus, and a collective area called the amygdaloid area, unspecified, based on observed grey-white matter contrast.

Striatopallidal regions.We added 7 new and revised 3 existing structures in the striatopallidal regions. These regions were only coarsely delineated in previous versions. The annotation of the striatum from previous versions¹⁶ was subdivided into the caudate putamen as well as the core and shell of the nucleus accumbens. The dorsal boundary of the caudate-putamen complex is clearly demarcated by white matter, and core / shell region of the nucleus accumbens were defined by sMRI signal intensity contrast and their location relative to the anterior commissure. Furthermore, we delineated an area located directly caudal to the nucleus accumbens and ventral to the caudate putamen. This ventral striatal region can be identified by its sMRI signal intensity, and corresponds partly to regions referred to as the fundus of striatum and interstitial limb of the posterior part of the anterior commissure (IPAC) in other atlases^29,31. However, we did not find adequate information in the template or in our reference data to subdivide the fundus of striatum or IPAC separately, and thus termed it as an ‘unspecified’ ventral striatal region. This area was included in the striatum annotation in previous versions of the WHS atlas. The ventral boundary of the ventral striatal region and shell of the nucleus accumbens towards the basal forebrain region was defined by using principal diffusion orientation differences visible in the DTI data. In the pallidum we subdivided the delineation of the external globus pallidus (referred to as globus pallidus in previous versions) into a medial and a lateral part. Furthermore, we delineated the ventral pallidum, which extends ventrally and rostrally from the external globus pallidus. The ventral pallidum is highly clustered and was identified by highly oriented (anisotropic) diffusion signal related to the anterioposteriorly oriented fibres from the olfactory tract and neighbouring basal forebrain areas. Lastly, in the pallidum, we also revised the delineation of the entopeduncular nucleus.

Thalamus.We added 55 new subregions in the thalamus, which was delineated as a single structure in previously published versions. Within the thalamus, white matter bands such as the medial lemniscus, the external and internal medullary lamina, and the superior cerebellar peduncle, are distinctly visible in the sMRI/DTI data and provide a framework for delineating subregions of the pre-, epi- and dorsal thalamus. In the prethalamus, we delineated the reticular (pre)thalamic nucleus (only the auditory segment was delineated as part of the previous version¹⁸, then referred to as the auditory segment of the reticular thalamic nucleus). We furthermore delineated rostral, dorsal, ventral and caudal parts of the zona incerta, the A13 and A11 dopamine cell groups, and the Fields of Forel.In the epithalamus, we delineated the habenula and subdivided it into lateral and medial parts. In the dorsal thalamus, we delineated 43 areas belonging to nine regional groups: the anterior nuclei; the dorsal-caudal midline group; the ventral midline group; the mediodorsal nucleus; the ventral nuclei; the intralaminar nuclei; the posterior complex; the lateral posterior (pulvinar) complex; the laterodorsal thalamic nuclei; and the medial geniculate complex. The medial geniculate complex was subdivided in the previously published version 3 of the atlas¹⁸, but all subregions except the marginal zone were revised and have therefore been assigned new region IDs in version 4. In this version of the atlas, the medial geniculate complex thus includes the dorsal, ventral, and medial divisions, as well as the marginal zone and the suprageniculate nucleus. In general, the delineation of thalamic subregions largely follows the annotations of Paxinos and Watson^29,30. However, in the posterior thalamus, the atlases by Paxinos and Watson^29,30 are incomplete, with large areas remaining unspecified. The posterior complex of the thalamus in the WHS rat brain atlas includes the posterior thalamic nucleus and the posterior thalamic nuclear group, triangular part.

Midbrain dopaminergic regions. We added 4 new structures to the midbrain dopaminergic regions. We delineated the subregions of the substantia nigra, i.e. the reticular, compact, and lateral parts. The ventral tegmental area was also delineated based on its appearance in the sMRI/DTI data and its relative position to the substantia nigra.

Quantification of parvalbumin neurons using the QUINT workflow with WHSv4

To demonstrate the practical value of the complete rat brain atlas for analysis, interpretation, visualization and communication of data, we re-analysed a published data set with the new atlas using the QUINT workflow. The raw and derived data used for this analysis²⁴ are available through the EBRAINS Knowledge Graph, and were previously interpreted using WHS rat brain atlas v2. We here re-analysed these data for one of the subjects (rat 25205), using the derived data set⁵⁰ and version 4 of the atlas. To optimize the spatial registration to WHSv4, we adjusted the .json file (ext-d000008_PVRat_25205_nonlinear.json) provided with the data set⁴² using VisuAlign (v0.9; RRID:SCR_017978). We then exported the new customized atlas maps and combined these with the existing segmentation maps using Nutil Quantifier. We used the same Nutil parameters as previously²⁴ except that we created new custom regions tailored to WHS rat brain atlas v4. The custom regions largely correspond to the highest level of detail in the atlas, but all white matter tracts were merged, as well as very fine grey matter areas such as cortical layers. Upon running Nutil Quantifier, we post-processed the reports as described previously²⁴ to arrive at total number and density estimates for each custom brain region.

Bakker, R., Tiesinga, P. & Kötter, R. The Scalable Brain Atlas: Instant Web-Based Access to Public Brain Atlases and Related Content. Neuroinformatics 13, 353–366 (2015).
Puchades, M., Csucs, G., Ledergerber, D., Leergaard, T. & Bjaalie, J. Spatial registration of serial microscopic brain images to three-dimensional reference atlases with the QuickNII tool. PLoS One 14, e0216796 (2019).
Krepl, J. et al. Supervised Learning With Perceptual Similarity for Multimodal Gene Expression Registration of a Mouse Brain Atlas. Front. Neuroinform. 15, 37 (2021).
Yates, S. et al. QUINT: Workflow for Quantification and Spatial Analysis of Features in Histological Images From Rodent Brain. Front. Neuroinform. 13, 1–14 (2019).
Oh, S. et al. A mesoscale connectome of the mouse brain. Nature 508, 207–214 (2014).
Mano, T. et al. CUBIC-Cloud: An Integrative Computational Framework Towards Community-driven Whole-Mouse-Brain Mapping. bioRxiv 2020.08.28.271031 (2020) doi:10.1101/2020.08.28.271031.
Wang, Q. et al. The Allen Mouse Brain Common Coordinate Framework: A 3D reference atlas. Cell 181, 1–18 (2020).
Bjerke, I. et al. Data integration through brain atlasing: Human Brain Project tools and strategies. Eur. Psychiatry 50, 70–76 (2018).
Qu, L. et al. Cross-modal coherent registration of whole mouse brains. Nat. Methods 19, 111–118 (2022).
Kim, Y. et al. Brain-wide maps reveal stereotyped cell-type-based cortical architecture and subcortical sexual dimorphism. Cell 171, 456–469 (2017).
Fürth, D. et al. An interactive framework for whole-brain maps at cellular resolution. Nat. Neurosci. 21, 139–153 (2018).
Pallast, N., Wieters, F., Fink, G. & Aswendt, M. Atlas-based imaging data analysis tool for quantitative mouse brain histology (AIDAhisto). J. Neurosci. Methods 326, 108394 (2019).
Hjorth, J. J. J. et al. The microcircuits of striatum in silico. Proc. Natl. Acad. Sci. U. S. A. 117, 9554–9565 (2020).
Lein, E. et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature 445, 168–176 (2007).
Winnubst, J. et al. Reconstruction of 1,000 Projection Neurons Reveals New Cell Types and Organization of Long-Range Connectivity in the Mouse Brain. Cell 179, 268-281.e13 (2019).
Papp, E., Leergaard, T., Calabrese, E., Johnson, G. & Bjaalie, J. Waxholm Space atlas of the Sprague Dawley rat brain. Neuroimage 97, 374–386 (2014).
Kjonigsen, L., Lillehaug, S., Bjaalie, J., Witter, M. & Leergaard, T. Waxholm Space atlas of the rat brain hippocampal region: Three-dimensional delineations based on magnetic resonance and diffusion tensor imaging. Neuroimage 108, 441–449 (2015).
Osen, K., Imad, J., Wennberg, A., Papp, E. & Leergaard, T. Waxholm Space atlas of the rat brain auditory system: Three-dimensional delineations based on structural and diffusion tensor magnetic resonance imaging. Neuroimage 199, 38–56 (2019).
Cox, R. W. & Hyde, J. S. Software tools for analysis and visualization of fMRI data. NMR Biomed. 10, 171–178 (1997).
Rabut, C. et al. 4D functional ultrasound imaging of whole-brain activity in rodents. Nat. Methods 16, 994–997 (2019).
Gesnik, M. et al. 3D functional ultrasound imaging of the cerebral visual system in rodents. Neuroimage 149, 267–274 (2017).
Schubert, N. et al. 3D Reconstructed Cyto-, Muscarinic M2 Receptor, and Fiber Architecture of the Rat Brain Registered to the Waxholm Space Atlas. Front. Neuroanat. 10, 51 (2016).
Fuglstad, J., Saldanha, P., Paglia, J. & Whitlock, J. HERBS: Histological E-data Registration in rodent Brain Spaces. bioRxiv 2021.10.01.462770 (2021) doi:10.1101/2021.10.01.462770.
Bjerke, I. et al. Densities and numbers of calbindin and parvalbumin positive neurons across the rat and mouse brain. iScience 24, 101906 (2021).
Crittenden, J. & Graybiel, A. Basal ganglia disorders associated with imbalances in the striatal striosome and matrix compartments. Front. Neuroanat. 5, 1–25 (2011).
Graybiel, A. Neurotransmitters and neuromodulators in the basal ganglia. Trends Neurosci. 13, 244–254 (1990).
Hawrylycz, M. et al. The INCF Digital Atlasing Program: Report on Digital Atlasing Standards in the Rodent Brain. Nat. Preced. (2009) doi:10.1038/npre.2009.4000.
Papp, E., Leergaard, T., Calabrese, E., Johnson, G. & Bjaalie, J. Addendum to “Waxholm Space atlas of the Sprague Dawley rat brain” [NeuroImage 97 (2014) 374-386]. Neuroimage (2014) doi:10.1016/j.neuroimage.2014.10.017.
Paxinos, G. & Watson, C. The rat brain in stereotaxic coordinates. (Academic Press, 2007).
Paxinos, G. & Watson, C. The rat brain in stereotaxic coordinates. (Elsevier Inc, 2013).
Swanson, L. Brain maps 4.0—Structure of the rat brain: An open access atlas with global nervous system nomenclature ontology and flatmaps. J. Comp. Neurol. 526, 935–943 (2018).
Hall, R. & Lindholm, E. Organization of motor and somatosensory neocortex in the albino rat. Brain Res. 66, 23–38 (1974).
Chapin, J. K. & Lin, C. ‐S. Mapping the body representation in the SI cortex of anesthetized and awake rats. J. Comp. Neurol. 229, 199–213 (1984).
Fabri, M. & Burton, H. Ipsilateral cortical connections of primary somatic sensory cortex in rats. J. Comp. Neurol. 311, 405–424 (1991).
Leergaard, T. B. et al. Three-dimensional topography of corticopontine projections from rat sensorimotor cortex: Comparisons with corticostriatal projections reveal diverse integrative organization. J. Comp. Neurol. 478, 306–322 (2004).
Hoffer, Z. S., Hoover, J. E. & Alloway, K. D. Sensorimotor corticocortical projections from rat barrel cortex have an anisotropic organization that facilitates integration of inputs from whiskers in the same row. J. Comp. Neurol. 466, 525–544 (2003).
Paxinos, G., Watson, C., Calabrese, E., Badea, A. & Johnson, G. MRI/DTI Atlas of the Rat Brain. (Academic Press Inc, 2015).
Boccara, C. N. et al. A three-plane architectonic atlas of the rat hippocampal region. Hippocampus 25, 838–857 (2015).
Yushkevich, P. et al. User-guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability. Neuroimage 31, 1116–1128 (2006).
Boccara, C. et al. Coronal section images series showing neuronal nuclei, calbindin and parvalbumin in the rat hippocampal region. (Human Brain Project Neuroinformatics Platform, 2019). doi:10.25493/AMW1-Z16.
Leergaard, T., Lillehaug, S., Dale, A. & Bjaalie, J. Atlas of normal rat brain cyto- and myeloarchitecture. at https://doi.org/10.25493/C63A-FEY (2018).
Laja, A., Bjerke, I., Leergaard, T. & Witter, M. Distribution of parvalbumin-positive interneurons in the normal adult rat brain. (EBRAINS, 2020). doi:10.25493/8KCQ-3C7.
Klein, S., Staring, M., Murphy, K., Viergever, M. A. & Pluim, J. P. W. Elastix: A toolbox for intensity-based medical image registration. IEEE Trans. Med. Imaging 29, 196–205 (2010).
Tustison, N. J. et al. The ANTsX ecosystem for quantitative biological and medical imaging. Sci. Rep. 11, 9068 (2021).
Hill, B. & Baldock, R. A. Constrained distance transforms for spatial atlas registration. BMC Bioinformatics 16, 90 (2015).
Avants, B. B., Johnson, H. J. & Tustison, N. J. Neuroinformatics and the The Insight ToolKit. Front. Neuroinform. 9, 5 (2015).
Vandenberghe, M. et al. High-throughput 3D whole-brain quantitative histopathology in rodents. Sci. Rep. 6, 1–12 (2016).
Tappan, S. J. et al. Automatic navigation system for the mouse brain. J. Comp. Neurol. 527, 2200–2211 (2019).
Groeneboom, N., Yates, S., Puchades, M. & Bjaalie, J. Nutil: A Pre- and Post-processing Toolbox for Histological Rodent Brain Section Images. Front. Neuroinform. 14, 37 (2020).
Bjerke, I., Yates, S., Puchades, M., Bjaalie, J. & Leergaard, T. Brain-wide quantitative data on parvalbumin positive neurons in the rat. (EBRAINS, 2020). doi:10.25493/KR92-C33.
Zaborsky, L., Pang, K., Somogyi, J., Nadasdy, Z. & Kallo, I. The Basal Forebrain Corticopetal System Revisited. Ann. N. Y. Acad. Sci. 877, 339–367 (1999).
Newmaster, K. et al. Quantitative cellular-resolution map of the oxytocin receptor in postnatally developing mouse brains. Nat. Commun. 11, 1–12 (2020).
Kim, Y. et al. Mapping social behavior-induced brain activation at cellular resolution in the mouse. Cell Rep. 10, 292–305 (2015).
Zhang, C. et al. A platform for stereological quantitative analysis of the brain- wide distribution of type-specific neurons. Sci. Rep. 7, 1–12 (2017).
Tocco, C., Øvsthus, M., Bjaalie, J., Leergaard, T. & Studer, M. Topography of corticopontine projections is controlled by postmitotic expression of the area-mapping gene Nr2f1. bioRxiv 1–45 (2021).
Papp, E., Leergaard, T., Csucs, G. & Bjaalie, J. Brain-Wide Mapping of Axonal Connections: Workflow for Automated Detection and Spatial Analysis of Labeling in Microscopic Sections. Front. Neuroinform. 10, 1–11 (2016).
Schadt, F., Israel, I. & Samnick, S. Development and Validation of a Semi-Automated, Preclinical, MRI-Template Based PET Image Data Analysis Tool for Rodents. Front. Neuroinform. 15, 1–14 (2021).
Dempsey, B. et al. Mapping and analysis of the connectome of sympathetic premotor neurons in the rostral ventrolateral medulla of the rat using a volumetric brain atlas. Front. Neural Circuits 11, 1–16 (2017).
Barrière, D. et al. The SIGMA rat brain templates and atlases for multimodal MRI data analysis and visualization. Nat. Commun. 10, 1–13 (2019).
Osanlouy, M. et al. The SPARC DRC: Building a Resource for the Autonomic Nervous System Community. Front. Physiol. 12, 1–18 (2021).
Swanson, L. Brain Maps III: Structure of the rat brain. (Elsevier, 2004).
Cappaert, N. L. M., Van Strien, N. M. & Witter, M. P. Hippocampal formation. 511–573 (2015).
Terminology, F. C. on A. Terminologia anatomica: international anatomical terminology. (Thieme, 1998).
Malmierca, M. Auditory system. in The Rat Nervous System (Elsevier, Academic Press, 2015).
Groenewegen, H. J., Berendse, H. W., Wolters, J. G. & Lohman, A. H. M. The anatomical relationship of the prefrontal cortex with the striatopallidal system, the thalamus and the amygdala: Evidence for a parallel organization. Prog. Brain Res. 85, 95–118 (1991).
Groenewegen, H. J., Wright, C. I. & Uylings, H. B. M. The anatomical relationships of the prefrontal cortex with limbic structures and the basal ganglia. Journal of Psychopharmacology vol. 11 99–106 at https://doi.org/10.1177/026988119701100202 (1997).
Heidbreder, C. A. & Groenewegen, H. J. The medial prefrontal cortex in the rat: Evidence for a dorso-ventral distinction based upon functional and anatomical characteristics. Neuroscience and Biobehavioral Reviews vol. 27 555–579 at https://doi.org/10.1016/j.neubiorev.2003.09.003 (2003).
Márquez-Legorreta, E., Horta-Júnior, J. de A. C., Berrebi, A. S. & Saldaña, E. Organization of the Zone of Transition between the Pretectum and the Thalamus, with Emphasis on the Pretectothalamic Lamina. Front. Neuroanat. 10, 1–20 (2016).
Le Bihan, D. et al. Diffusion tensor imaging: Concepts and applications. J. Magn. Reson. Imaging 13, 534–546 (2001).
Boccara, C. et al. Sagittal and horizontal section images showing neuronal nuclei, calbindin and parvalbumin staining in the rat hippocampal region. (Human Brain Project Neuroinformatics Platform, 2019). doi:10.25493/JQ8F-TNF.
Olsen, G. et al. Organization of Posterior Parietal–Frontal Connections in the Rat. Front. Syst. Neurosci. 13, 38 (2019).
Halley, A., Baldwin, M., Cooke, D., Englund, M. & Krubitzer, L. Distributed motor control of limb movements in rat motor and somatosensory cortex: The sensorimotor amalgam revisited. Cereb. Cortex 30, 6296–6312 (2020).
Miller, M. & Vogt, B. Direct connections of rat visual cortex with sensory, motor, and association cortices. J. Comp. Neurol. 226, 184–202 (1984).
Bjerke, I., Schlegel, U., Puchades, M., Bjaalie, J. & Leergaard, T. Paxinos & Watson’s “The Rat Brain in Stereotaxic Coordinates” (6th edition) spatially registered to the Waxholm Space atlas of the rat brain. (Human Brain Project Neuroinformatics Platform, 2019). doi:10.25493/XQ8J-TNE.
Bjerke, I., Schlegel, U., Puchades, M., Bjaalie, J. & Leergaard, T. Swanson’s “Brain Maps: Structure of the Rat Brain” (4th edition) spatially registered to the Waxholm Space atlas of the rat brain. (Human Brain Project Neuroinformatics Platform, 2019). doi:10.25493/486N-966.
Bjerke, I., Schlegel, U., Puchades, M., Bjaalie, J. & Leergaard, T. Paxinos and colleagues’ “MRI/DTI Atlas of the Rat Brain” (1st edition) spatially registered to the Waxholm Space atlas of the rat brain. at (2019).
Welker, C. Microelectrode delineation of fine grain somatotopic organization of SmI cerebral neocortex in albino rat. Brain Res. 26, (1971).
Hall, R. & Lindholm, E. Organization of motor and somatosensory neocortex in the albino rat. Brain Res. 66, 23–38 (1974).
Kondo, H. & Witter, M. Topographic organization of orbitofrontal projections to the parahippocampal region in rats. J. Comp. Neurol. 522, 772–793 (2014).
Mathiasen, M., Hansen, L. & Witter, M. Insular projections to the parahippocampal region in the rat. J. Comp. Neurol. 523, 1379–1398 (2015).
Olsen, G. & Witter, M. Posterior parietal cortex of the rat: Architectural delineation and thalamic differentiation. J. Comp. Neurol. 524, 3774–3809 (2016).
Paxinos, G. The Rat Nervous System. (Academic Press, 2004).

There is NO Competing Interest. JGB is a member of the Management Board of the EBRAINS AISBL, Brussels, Belgium.

S1WHSSDratatlasv4Hierarchy.pdf
Supplementary File 1: Overview of atlas hierarchical atlas terminology, with full atlas terms and abbreviations

Download PDF

Journal Publication

published 02 Oct, 2023

Read the published version in Nature Methods →

Version 1

posted

You are reading this latest preprint version

Waxholm Space atlas of the rat brain: A 3D atlas supporting data analysis and integration

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Declarations

Methods

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1