Cross-species gray-matter atlas spanning rodents, nonhuman primates, and humans.

Overview

The Common Hierarchical Atlas (CHA) provides a unified anatomical framework for cortical and subcortical gray-matter parcellation across multiple species. It was developed to support cross-species comparison in translational neuroscience by harmonizing homologous regions within a consistent hierarchical coordinate system. :contentReference[oaicite:1]{index=1}

The atlas is organized using species-specific minimal deformation templates (MDTs) and a common parcellation framework built on consistent tissue segmentation and literature-guided anatomical homologies. Primary species include:

• Human
• Rhesus macaque
• Marmoset
• Mouse

Rat coverage is also provided through atlas transfer from mouse using MRI-based alignment.

Key Features

• Population-based MDT reference spaces
• Cross-species consistent cortical and subcortical parcellation
• Hierarchical atlas design with multiple anatomical levels
• Harmonized regional definitions across rodents and primates
• Validation against species-specific atlases
• Freely available NIfTI atlas files for each species

Hierarchy

The CHA atlas is organized into three levels:

Level 0: Tissue Compartments

Level 0 defines major tissue classes derived from automated MRI segmentation:

• Gray matter (GM)
• White matter (WM)
• Deep gray matter (DGM)
• Cerebellum (CBL)
• Cerebrospinal fluid (CSF)

These compartment maps form the anatomical foundation for higher-level parcellations. :contentReference[oaicite:4]{index=4}

Level 1: Macro-Regions

Level 1 defines nine major cortical and subcortical divisions:

• Frontal (FRO)
• Parietal (PAR)
• Temporal (TEM)
• Occipital (OCC)
• Insular (INS)
• Cingulate (CIN)
• Olfactory (OLF)
• Basal ganglia (BG)
• Thalamic / hypothalamic system (THL)

These macro-regions are defined using consistent anatomical landmarks while allowing species-specific anatomical variation. :contentReference[oaicite:5]{index=5}

Level 2: Subdivisions

Level 2 provides finer anatomical parcels. There are:

• 26 Level-2 regions in primates
• 21 Level-2 regions in rodents

Some regions remain undivided in rodents due to anatomical differences or MRI segmentation limitations, including:

• Parietal lobe (PAR)
• Caudoputamen (CPu)

Included Level-2 Regions

Frontal

• FRO_Precentral (PreC)
• FRO_Premotor (PreM)
• FRO_Prefrontal (PreF)

Parietal

• PAR

Temporal

• TEM_Superior (TemS)
• TEM_Inferior (TemI)
• TEM_Medial (TemM)
• TEM_Hippocampus (Hipp)
• TEM_Amygdala (Amyg)

Occipital

• OCC_Lateral (OccL)
• OCC_Medial (OccM)

Insula

• INS_Anterior (InsA)
• INS_Posterior (InsP)

Cingulate

• CIN_Anterior (CingA)
• CIN_Posterior (CingP)

Olfactory

• OLF_Anterior (OlfA)
• OLF_Piriform (Pir)

Basal Ganglia

• BG_CaudoPutamen (CPu)
• BG_Accumbens (NAc)
• BG_Pallidum (Pal)

Thalamic / Hypothalamic

• THL_Thalamus (Thal)
• THL_Hypothalamus (Hypo)

Excluded Structures

The following structures were not included because reliable cross-species MRI segmentation was not feasible:

• Claustrum
• Substantia nigra
• Subthalamic nuclei

These exclusions reflect either MRI contrast limitations or lack of sufficiently consistent boundaries across species.

Coordinate Space and Templates

All atlas labels are defined in species-specific MDT spaces. Except for the human template, each MDT was generated specifically for this framework and aligned to the AC-PC axis with the origin set at the anterior commissure (AC). :contentReference[oaicite:9]{index=9}

Species Templates

• Human: ICBM152 Nonlinear Asymmetrical 2009a template
• Rhesus macaque: MDT generated from PRIME-DE T1-weighted MRI
• Marmoset: Brain/MINDS / MBM-derived reference with AC-centered coordinate adjustment
• Mouse: High-resolution diffusion MRI MDT
• Rat: MDT generated from ex vivo diffusion MRI and aligned to the common framework

:contentReference[oaicite:10]{index=10}

Construction Principles

Because cross-species homology is not always uniquely defined, the atlas prioritized three design principles:

1. Preserve regional definitions widely used in existing atlases
2. Maximize cross-species consistency where boundaries can be reasonably identified in MRI
3. Retain functional interpretability

In some cases, functionally meaningful groupings were favored over finer species-specific distinctions. For example:

• Superior and middle temporal gyri were grouped into a single superior temporal parcel for cross-species consistency
• Frontal subdivisions lacking clear homologues across species were not retained separately
• Olfactory parcels were preserved because of their importance in rodent and marmoset neurobiology

:contentReference[oaicite:11]{index=11}

Source Atlases Used

Level-2 homologies were defined with support from one or more established atlases, including:

• FreeSurferSeg (human)
• Brodmann areas (human)
• Allen Brain Atlas (mouse)
• CIVM atlas (rhesus macaque)
• Inia19 / NeuroMaps (rhesus macaque)
• Brain/MINDS atlas (marmoset)
• MBM atlas (marmoset)
• Paxinos atlas (marmoset)

When multiple references were available, boundaries were cross-validated across atlases.

File Availability

Atlas files are available for each species as:

<species>_CHA.nii.gz