An open access mouse brain flatmap and upgraded rat and human brain flatmaps based on current reference atlases

Abstract

Here we present a flatmap of the mouse central nervous system (CNS) (brain) and substantially enhanced flatmaps of the rat and human brain. Also included are enhanced representations of nervous system white matter tracts, ganglia, and nerves, and an enhanced series of 10 flatmaps showing different stages of rat brain development. The adult mouse and rat brain flatmaps provide layered diagrammatic representation of CNS divisions, according to their arrangement in corresponding reference atlases: Brain Maps 4.0 (BM4, rat) (Swanson, The Journal of Comparative Neurology, 2018, 526, 935-943), and the first version of the Allen Reference Atlas (mouse) (Dong, The Allen reference atlas, (book + CD-ROM): A digital color brain atlas of the C57BL/6J male mouse, 2007). To facilitate comparative analysis, both flatmaps are scaled equally, and the divisional hierarchy of gray matter follows a topographic arrangement used in BM4. Also included with the mouse and rat brain flatmaps are cerebral cortex atlas level contours based on the reference atlases, and direct graphical and tabular comparison of regional parcellation. To encourage use of the brain flatmaps, they were designed and organized, with supporting reference tables, for ease-of-use and to be amenable to computational applications. We demonstrate how they can be adapted to represent novel parcellations resulting from experimental data, and we provide a proof-of-concept for how they could form the basis of a web-based graphical data viewer and analysis platform. The mouse, rat, and human brain flatmap vector graphics files (Adobe Reader/Acrobat viewable and Adobe Illustrator editable) and supporting tables are provided open access; they constitute a broadly applicable neuroscience toolbox resource for researchers seeking to map and perform comparative analysis of brain data.

Keywords: brain atlases; brain flatmap; brain mapping; computer graphics; human; mouse; rat.

Figure 1.

A color-coded flatmap representation of major divisions of the vertebrate central nervous system. The flatmap template shown is based on one created previously for the adult rat that was itself based on rat brain development fatemaps (see text for details). The same overview template is used here for rat and mouse brain flatmaps; an updated human brain flatmap overview follows the same color scheme (see Figure 8).

Figure 2.

Flatmap representation of the nested divisional hierarchy of the rat central nervous system (CNS), following Brain Maps 4.0 (Swanson, 2018), and used here in RtBF5. The top left panel (a) shows the composite CNS divisional hierarchy, comprised of 143 parent divisions (above the level of gray matter region), 418 gray matter regions, and 122 subregions (for each side—left and right—of the CNS; excepting divisions that are cortical lamina). The tiles for each division are delineated, and semi-transparent to create an impression of the depth of nesting. Panels b-l represent the individual hierarchical divisions of the CNS: Panels b and c represent (respectively) gray matter regions, and subregions; panels d-l represent each of the 9 depths of CNS parent division. Yellow shading indicates the presence of a given division, whereas a black background indicates its absence. The two curved lines to the left of each panel represent the retina (a gray matter region division of the hypothalamus).

Figure 3.

Graphical upgrades to the flatmap representation of the nerves and ganglia of the peripheral nervous system (PNS), and central nervous system (CNS) white matter tracts (tracts). As shown in (a), different divisional groupings are represented by contrasting colors, with a different shade of the same color used to both indicate and differentiate PNS nerves and ganglia belonging to the same divisions, as shown in the key below (a). The flatmap representation is also enhanced by optional light (a), and dark modes (b, c). The dark mode shown in (b) presents an inner black background that delineates the shape of the underlying brain flatmap; whereas the dark mode shown in (c) presents an outer dark gray background, as an alternative contrasting schema (both dark modes can also be enabled together). As seen at higher magnification (inset box d in (a) shown in d), further enhancement of structural differentiation is provided by selective shading of nerves, the use of italics for nerve abbreviations [see comment above in Results], non-italicized text for ganglia abbreviations, and use of the same color for text and graphics belonging to the same divisional grouping (for the purposes of illustration, not all abbreviations are represented). An additional graphical enhancement (inset box e in (a) shown in e) is the use transparency to improve visual differentiation of closely apposed and overlapping lines (representing nerves and tracts). In (e), segments of several white matter tracts at the rostral end of the spinal cord are represented by blue lines. The lines are shaded longitudinally and are semi-transparent; the graphical enhancement can be readily appreciated by direct comparison with the same section in the most recent previous version (f) (Swanson, 2018). Abbreviations: cen2, second cervical nerve; GDIX, distal glossopharyngeal ganglion; GSC, superior cervical ganglion; GSNce2, second cervical spinal ganglia; IXt, glossopharyngeal nerve trunk; jn, jugular nerves.

Figure 4.

Circular symbols are used to represent cortical lamination on the rat and mouse brain flatmaps. A representative example for the rat brain flatmap is shown in (a) for 5 gray matter regions (red text) and 2 subregions (blue text to the left of two layer indicator series), including the olfactory bulb and some adjacent regions of the cerebral cortex. For sequentially numbered layers (for example, 1-2, 1-4, or 1-6, as shown), the layer indicators (white disks with blue outlines) are arranged in a corresponding left-to-right sequence. For individually named layers, each layer name abbreviation is placed next to its layer indicator. The rat and mouse brain flatmaps (Supporting Information 1) include layer indicators for all layered gray matter regions and subregions of the cerebral and cerebellar cortices. As shown in (b), the layer indicator symbols can be individually altered to represent data. In the example, a color scale is applied that corresponds to binned percentage tertiles. A zero value is indicated by a black symbol fill color, and the default white fill color indicates no data. To facilitate selection of layer markers to adapt, in the brain flatmap Adobe Illustrator file (Supporting Information 1) they are organized in layers named by the abbreviation of their corresponding gray matter region and arranged alphabetically. Abbreviations: AOA, anterior olfactory area; AOB, accessory olfactory bulb; e, external part; gl, glomerular layer; gr, granular layer; ILA, infralimbic area; ipl, inner plexiform layer; mi, mitral layer; MOB, main olfactory bulb; opl, outer plexiform layer; pr, principal part; TTd, tenia tecta dorsal part.

Figure 5.

Bilateral flatmaps representing several stages of embryonic development of the rat nervous system, and the adult structure, beginning with the neural plate (a) (truncated caudally for the illustration as indicated by the inset), and progressing through eight representative embryonic stages (b-i) to adult (j). The lower half of each flatmap (left side, as shown) represents the actual embryonic (e) stage (or the adult); whereas the upper half (right side, as shown) represents the developmental fate of structures at a later stage. Major structures are color-coded as indicated by the color-key (top left) that also shows the relationships of the represented gray matter subdivisions. The area of the major subdivisions is proportional to the embryonic volume of corresponding gray matter (see Fig. 14 in Alvarez-Bolado & Swanson, 1996). The figure is based on a previous version (see Supporting Information 5 in Swanson, 2018) of an earlier original (see Figure 17 in Alvarez-Bolado & Swanson, 1996). For the purposes of illustration, structure labeling in b-j is simplified (for other labels see RtDevBF2 in Ai file Supporting Information 1, or the PDF file for RtDevBF2 in Supporting Information 2). Abbreviations: ebp, epibranchial placodes; hp, hypophysial placode; olp, olfactory placode; opm, oropharyngeal membrane; OPV, optic vesicle; rlr, rostrolateral ridge; sopt, optic sulcus.

Figure 6.

Comparison of the number of gray matter divisions for each depth of division of the central nervous system (CNS) according to its structural divisional hierarchy arranged topographically in Rat Brain Maps 4.0 (BM4) (Swanson, 2018), compared for the rat (based on BM4), and the mouse (based on the Allen Reference Atlas, ARAv1) (Dong, 2007). The upper chart (a) compares 9 depths of CNS divisions above the level of gray matter region (parent divisions) beginning with CNS at depth 1. The lower chart (b) shows the total numbers for all gray matter divisions (All divisions), and for the gray matter division categories of parent division (Parent), gray matter region (Region), and gray matter subregion (Subregion; including sub-subregion divisions, but excepting divisions that layers of cerebral and cerebellar cortical regions that apply similarly to rat and mouse). In general, ARAv1 has fewer gray matter divisions in each category (b) and at each depth of CNS division (a) compared to BM4, primarily due to a more coarse-grained parcellation. The retina (1 gray matter region) and spinal cord (22 gray matter regions) were not included in ARAv1 but were included in BM4; however, BM4 numbers for these divisions are included with the ARAv1 numbers here (and on MsBFv1 where they are present) to avoid exaggerating differences that were not due to these omissions. Conversely, to avoid understating differences, subregions that are laminar divisions of gray matter regions in the cerebral cortex and elsewhere (and present in both rat and mouse), are not included. Despite fewer ARAv1 gray matter divisions compared to BM4, the overall distribution is quite similar, reflecting a high-level of similarity in the underlying parcellation as represented in the reference atlases (see Supporting Information 3, and text for details).

Figure 7.

Flatmap adaptability to novel parcellation schemas based on experimental data. Two examples, using MsBF1, illustrate how the present brain flatmaps can be adapted to represent novel parcellations. The first example (a) represents a refined parcellation for the caudoputamen (CP) based on network analysis of axonal connections from the cerebral cortex to the CP (Hintiryan et al., 2016). The second example (b) represents a refined parcellation for hippocampal regions based on a combination of gene-expression and network analysis (Bienkowski et al., 2018). The hippocampal divisions were mapped to the parcellation represented on MsBF1, and also with respect to atlas level contours for the cerebral cortex corresponding to ARAv1 (shown in purple on the right side of (b), that also shows the retrohippocampal region that is the other major division of the hippocampal formation). The insets at lower left show the general location of the regions shown in (a) and (b). The table at lower right shows the divisional hierarchy underlying each novel parcellation schema, and ARAv1 atlas levels for the hippocampal divisions. See text for additional information. Abbreviations shown in red text in (b) (upper right side) are defined in Supporting Information 3. Abbreviations for (a) and upper part of table: cd, central dorsal; cvl, central ventrolateral; cvm, central ventromedial; d, dorsal; dl, dorsolateral; dm, dorsomedial; ext, extreme; i, intermediate; im, intermedial; imd, intermedial dorsal (applies to CPi.dl.imd); imd, intermediate dorsal (applies to CPr.imd); imv, intermedial ventral (applies to CPi.vl.imv); imv, Intermediate ventral (applies to CPr.imv); l, lateral; ls, lateral strip; r, rostral; v, ventral; vm, ventromedial; vt, ventral tip. Abbreviations for (b) and lower part of table: CA1-3, Ammon’s horn (cornu Ammonis in Latin) Field 1, 2, or 3; CAldc, CA1 dorsal, caudal part; CAldr, CA1 dorsal, rostral part; CAlvv, CA1 ventral tip; CA3dd, CA3 rostral-dorsal tip; CA3vv, CA3 ventral tip; dc, dorsal caudal; dd, dorsal tip; DG, dentate gyrus; ic, intermediate caudal; id, intermediate dorsal; pod, dorsal polymorph layer; pov, ventral polymorph layer; ProSUB, prosubiculum; SUB, subiculum; SUBdd, dorsal subiculum, dorsal part (tip); SUBdv, dorsal subiculum, ventral part (tip); SUBvv, ventral subiculum, ventral part (tip).

Figure 8.

A flatmap representation of the adult human brain (a) showing on the upper half (right side of brain) Brodmann’s areal parcellation and associated numbering schema for divisions of the cerebral cortex, and on the lower half (left side of brain) primary sulci. The nested cortical subplate is outlined in green on the upper half. Also represented are major structural divisions of the brain, represented differently on either side for the purposes of illustration; their structural hierarchical relationships are indicated by a color-coded tree. Additional divisional parcellation is represented by the inset (b) for the basolateral amygdalar complex (BLX), and for selected divisions of the hippocampal formation. The present version of the human brain map (version 3) is based on an earlier (second) version (Swanson, 2018). See text for additional information. Abbreviations for (a): 6b, isocortical layer 6b; AOA, anterior olfactory area; CA, Ammon’s horn (cornu Ammonis in Latin); CBN, cerebellar nuclei; CLA, claustrum; COX, cortical amygdalar complex; DG, dentate gyrus; EP, endopiriform nucleus; IG, indusium griseum; INS, insular region; MB, midbrain; OB, olfactory bulb; PAL, pallidum; SBC, subicular complex; STR, striatum; TG, tegmentum; TT, tenia tecta. Abbreviations for (b): BLA, basolateral amygdala nucleus; BMA, basomedial amygdala nucleus; CA1-3, CA fields 1-3; d, dorsal; LA, lateral amygdala nucleus; m, medial; PARA, parasubiculum; POST, postsubiculum; PRE, presubiculum; SUB, subiculum; v, ventral.

Figure 9.

A conceptual brain flatmap visualization workflow. Initial input to the brain flatmap visualization workflow is tabulated experimental data representing measured values (automated pixel count, or cell count that is either manual or based on a thresholding algorithm applied to pixel count) corresponding to signal (for example a fluorescent reporter molecule) detected in brain tissue sections after histological processing, digital imaging, registration to a brain reference atlas (in this example a computer graphics version of the Allen Reference Atlas version 1 (ARAv1), Dong, 2007), and image processing to achieve optimal signal-noise ratio. Following data input, step 1 of the workflow analysis generates an annotated output recording data location by ARAv1 atlas level and by gray matter region (according to boundaries representing ARAv1 parcellation). In step 2, the data are aggregated across atlas levels for each gray matter region, to give an average value for each region; aggregation may also be performed across experiments (as indicated). In step 3, the aggregated region-specific data are visualized (for example as a heatmap as shown) on a scalable vector graphic implementation of the brain flatmap using a back-end coordinate system with tags applied to each coordinate-defined area to enable matching to gray matter regions. The example data shown here were reported previously (Zingg et al., 2018—Atlas Level maps (top left) reproduced from Figure 5, and flatmap visualization based on data for Animal 1 in Table 3); they represent sites of input to claustrum neurons that send monosynaptic projections to the retrosplenial area, determined by a combinatorial and conditional virus-based pathway-tracing strategy (Zingg et al., 2018). A monochrome heatmap represents positive data in five percentile bins (quintiles) calculated separately for connections ipsilateral and contralateral to the side targeted. The brain schematic at lower left represents major divisions of the central nervous system (see also Figure 1). The table at lower right lists cell counts for regions in the top 25% for total cell count, arranged from high to low, and showing counts for the targeted side (ipsilateral), contralateral side, and both sides (total). Background colors applied to the region abbreviations correspond to the color-coded CNS divisions below the table. For abbreviations see Supporting Information 3.