Supported Atlases#

Coord2Region provides unified access to a wide range of neuroimaging reference spaces. Whether you are working with volumetric NIfTI files, cortical surfaces, or coordinate-based regions of interest (ROIs), the package handles downloading and caching automatically.

Quick Start: Listing Atlases#

You can check which atlases are available in your installed version using either the command line or Python:

CLI#

coord2region --list_atlases

Python#

from coord2region.fetching import AtlasFetcher

print(AtlasFetcher().list_available_atlases())

Usage Guide#

Anatomical Mapping: Use Harvard-Oxford or AAL to label coordinates with standard macroscopic anatomical names.
Functional Connectivity: Use Schaefer or Yeo networks when defining nodes for graph theory.
Surface Analysis: Use Aparc or HCP-MMP if your pipeline operates in FreeSurfer vertex space.
Custom Data: Provide a direct URL or a local file path if your atlas is not listed below.

Volumetric Atlases (MNI Space)#

These atlases are voxel-based (NIfTI) and aligned to the MNI152 template.

Anatomical & Cytoarchitectonic#

Identifier	Description & Defaults
`harvard-oxford`	Standard Probabilistic Atlas. Maps cortical and subcortical areas. Defaults: `atlas_name='cort-maxprob-thr25-2mm'`.
`aal`	Automated Anatomical Labeling. Standard macroscopic atlas (AAL3v2). Defaults: `version='3v2'`.
`juelich`	Cytoarchitectonic Atlas. Microscopic cell distribution maps. Defaults: `atlas_name='maxprob-thr0-1mm'`.
`brodmann`	Brodmann Areas. Classic cytoarchitectonic labels (BA1-BA52) via Talairach transform.
`pauli`	Subcortical Nuclei. High-resolution probabilistic atlas (2017).
`talairach`	Talairach Daemon. Digital version of the Talairach atlas (NIfTI).

Functional & Parcellations#

Derived from functional clustering (fMRI) for defining homogeneous regions.

Identifier	Description & Defaults
`schaefer`	Schaefer 2018. Gradient-weighted Markov Random Fields. Defaults: `n_rois=400`, `yeo_networks=7`, `resolution_mm=1`.
`yeo`	Yeo 2011 (Volumetric). Functional network parcellation. Defaults: `n_networks=7` (Liberal mask).
`basc`	BASC Multiscale. Bootstrap Analysis of Stable Clusters (2015).
`destrieux`	Destrieux 2009. Volumetric version of the `aparc.a2009s` parcellation.

Surface Atlases (FreeSurfer/MNE)#

Parcellations defined on the cortical surface (vertices). By default, these map to the fsaverage subject.

Standard FreeSurfer#

Identifier	Description
`aparc`	Desikan-Killiany. The default FreeSurfer gyral-based parcellation.
`aparc.a2005s`	Desikan-Killiany (Legacy). The older 2005 version of the atlas.
`aparc.a2009s`	Destrieux. Finer parcellation including both gyri and sulci.
`aparc_sub`	MNE Subdivision. A subdivided version of aparc for finer granularity.

PALS & OASIS#

Identifier	Description
`yeo2011`	Yeo 2011 (Surface). Surface-based 17-network parcellation.
`pals_b12_lobes`	PALS-B12 Lobes.
`pals_b12_orbitofrontal`	PALS-B12 Orbitofrontal structures.
`pals_b12_visuotopic`	PALS-B12 Visuotopic areas.
`oasis.chubs`	OASIS CHUBS labels.

Connectome Project (HCP)#

Identifier	Description & Requirements
`human-connectum project`	HCP-MMP 1.0. Glasser 2016 multi-modal parcellation. Important License Required: You must accept data usage terms. Run: `mne.datasets.fetch_hcp_mmp_parcellation(accept=True)` Or set env var: `COORD2REGION_ACCEPT_HCPMMP=1`

Coordinate Sets (ROIs)#

Spherical ROIs or centroids used in graph theory meta-analyses.

Identifier	Description
`power`	Power 2011. 264 ROIs derived from meta-analysis and functional connectivity.
`dosenbach`	Dosenbach 2010. 160 ROIs focused on sensorimotor and task control networks.
`seitzman`	Seitzman 2018. 300 ROIs including subcortical and cerebellar regions.

Custom Atlases (URLs & Local Files)#

If your desired atlas is not in the built-in list, you can provide a direct link or a local file path.

Using a Direct URL

You can pass a downloadable URL to a NIfTI file directly to the fetcher or CLI.

# Python
fetcher.fetch_atlas(atlas_url="https://example.com/my_parcellation.nii.gz")

Using a Local File

If you have already downloaded an atlas (or created your own parcellation), point Coord2Region to the file path.

# Python
mapper = AtlasMapper(atlas_file="/path/to/my_custom_atlas.nii")

# CLI
coord2region coords-to-atlas 30 -20 50 --atlas-file /path/to/my_custom_atlas.nii

Implementation Details#

The fetch_atlas() method returns a dictionary containing:

maps: Path to the NIfTI or annotation file.
labels: List of region names.
description: Metadata string describing the atlas version.
type: One of volumetric, surface, or coords.