A Cell Atlas for the Mouse Brain
Erö C, Gewaltig M-O, Keller D and Markram H (2018) A Cell Atlas for the Mouse Brain. Frontiers in Neuroinformatics 12:84. doi: 10.3389/fninf.2018.00084
Summary
The document presents a dynamically generated 3D cell atlas of the whole mouse brain built from Allen Brain Atlas Nissl images, gene expression markers, and selected literature values. The authors estimate volumetric cell densities, generate 3D positions for cells across 737 brain regions, and classify them into excitatory and inhibitory neurons as well as astrocytes, oligodendrocytes, and microglia. The workflow uses image realignment, thresholding, overlap correction, and acceptance-rejection sampling to populate brain regions with cells, then validates the resulting regional counts and densities against published data and an automated point-detection approach. The atlas is designed as a public, expandable resource that can incorporate new markers, data sources, and future connectivity information.
Keywords
mouse brain
cell atlas
3D reconstruction
cell density
neurons
glia
astrocytes
oligodendrocytes
microglia
excitatory neurons
inhibitory neurons
Allen Mouse Brain Atlas
Nissl staining
gene expression
validation
Main claims
The study presents the first 3D whole mouse brain cell atlas with region by region estimates for major neuronal and glial cell types
Existing literature provides neuron counts for only a small fraction of mouse brain regions and reported values often differ substantially, motivating a data driven atlas that integrates multiple sources
The atlas reconstructs cell positions and densities for excitatory and inhibitory neurons, astrocytes, oligodendrocytes, and microglia across all 737 Allen Mouse Brain Atlas regions
The atlas is dynamic and can be refined as new staining, transcriptomic, and anatomical data become available
Whole brain scale analysis reveals nonuniform cellular organization, including extreme enrichment of cerebellar granule cell regions and relatively uniform glial distributions
Comparison with literature and automated counting shows generally good agreement while also highlighting persistent uncertainty in some regions such as the cerebellum