A neuroscientist's guide to using murine brain atlases for efficient analysis and transparent reporting

Abstract

Brain atlases are widely used in neuroscience as resources for conducting experimental studies, and for integrating, analyzing, and reporting data from animal models. A variety of atlases are available, and it may be challenging to find the optimal atlas for a given purpose and to perform efficient atlas-based data analyses. Comparing findings reported using different atlases is also not trivial, and represents a barrier to reproducible science. With this perspective article, we provide a guide to how mouse and rat brain atlases can be used for analyzing and reporting data in accordance with the FAIR principles that advocate for data to be findable, accessible, interoperable, and re-usable. We first introduce how atlases can be interpreted and used for navigating to brain locations, before discussing how they can be used for different analytic purposes, including spatial registration and data visualization. We provide guidance on how neuroscientists can compare data mapped to different atlases and ensure transparent reporting of findings. Finally, we summarize key considerations when choosing an atlas and give an outlook on the relevance of increased uptake of atlas-based tools and workflows for FAIR data sharing.

Keywords: FAIR data; brain atlases; brain-wide analysis; mouse brain; neuroinformatics; rat brain; reporting practices; spatial registration.

FIGURE 1

Navigating brain atlases to find anatomical locations. (A) Simplified version of the brain atlas ontology model (AtOM, Kleven et al., 2023a). The main elements of an atlas include the coordinate system, the reference image (here exemplified with a coronal platypus brain section; Mikula et al., 2007), the annotated brain regions, and the brain region names. The elements provide different entry points for navigating the atlas, through a spatial, semantic or visual reference. (B) Illustration of the three standard planes (horizontal, blue; sagittal, yellow; coronal, green) typically used to cut brain sections. (C) Illustration of the essential terminology typically used for indicating positions in the brain (e.g., the terms “rostral” and “caudal” to refer to positions towards the front and back of the brain, respectively). (D) Illustration of useful landmark regions in the murine brain [adapted from Bjerke et al. (2023)], with examples from the horizontal (D1), coronal (D2; Leergaard et al., 2018), and sagittal (D3) planes.

FIGURE 2

Using brain atlases for spatial registration, analysis, visualization and comparison of data. (A) Example of spatial registration of a histological section to the Waxholm Space (WHS) rat brain atlas. Landmark regions, such as the hippocampus (A1,A2), are used to find corresponding positions between the image and the atlas. Features in the images (A3), with spatial metadata from the registration, can be extracted (A3’). The example in (A) shows a histological image stained for parvalbumin neurons registered to the WHS atlas. (B) The principal workflow of combining atlas registration with extracted features illustrated in (A) can be used for different types of atlas-based analyses. (B1) 3D dot map visualization of corticostriatal, corticotectal, and corticopontine axonal projections originating from the primary somatosensory cortex (SS, red) and visual cortex (VIS, blue) cortical areas, extracted from anterograde tract tracing data (Oh et al., 2014) registered to the Allen mouse brain CCFv3-2017 (Ovsthus et al., 2022). (B2) Analysis of dopamine 1 receptor positive cell densities in olfactory regions of the mouse brain across five postnatal day (P) age groups [y axis values not shown, preliminary data extracted from images provided by Bjerke et al. (2022)]. (C) Visualization of customized regions from the Allen mouse brain CCFv3-2017. The left panel shows the entire atlas with the default color scheme. The middle panel shows a transparent view of the brain with regions of the hippocampal formation color coded to their corresponding region in the WHS rat brain atlas, facilitating cross-species comparisons. In the right panel, to better visualize the extent of individual regions, they are coded with contrasting colors, whereas the original atlas uses the same or highly similar colors. (D) Example of how co-registration of brain atlases supports comparison of data referenced in different atlases. The stereotaxic atlas by Swanson (1998) has been spatially registered to the WHS rat brain atlas (D1). Data that have been extracted and mapped to the two atlases can therefore be co-visualized in the same 3D space (D2). In this example, the red points are extracted from a previous study where retrograde projections from injections in the infralimbic cortex were represented with schematic drawing of terminal fields onto atlas plates from the Swanson atlas (Figure 8, data mirrored for comparison; Hoover and Vertes, 2007). The blue points are extracted from a public dataset showing the anterograde projections originating from the lateral orbitofrontal cortex (case F1 BDA; Kondo et al., 2022). AOB, accessory olfactory bulb; AON, anterior olfactory nucleus; CP, caudoputamen; COAa, cortical amygdalar area, anterior part; COApl, cortical amygdalar area, posterior part, lateral zone; COApm, cortical amygdalar area, posterior part, medial zone; DP, dorsal peduncular area; IL, infralimbic cortex; LO, lateral orbitofrontal cortex; MOB, main olfactory bulb; NLOT, nucleus of the lateral olfactory tract; PAA, piriform-amygdalar area; PG, pontine gray; PIR, piriform area; SC, superior colliculus; SS, somatosensory area; TH, thalamus; TR, postpiriform transition area; TTd, taenia tecta dorsal part; TTv, taenia tecta ventral part; VIS, visual area.