MNI Structural maxprob thr25 1mm Atlas

Rotating Brain

Full Quality Version: Download MP4

Overview

The MNI Structural maxprob thr25 1mm brain atlas is a probabilistic macro-anatomical parcellation of the human brain in MNI152 standard space that divides the cortex and subcortex into discrete regions based on structural (sulcal-gyral and gross anatomical) landmarks, assigning each voxel to the label with the highest probability while retaining only voxels with at least 25% likelihood of belonging to a given region (the “maxprob thr25” criterion). It was constructed by nonlinearly transforming multiple individual T1-weighted structural MRI scans to the MNI152 template, manually or semi-automatically segmenting key anatomical structures in each brain, and then aggregating these segmentations to produce probabilistic maps that capture inter-individual variability in regional boundaries at 1 mm isotropic resolution. This atlas is widely used for region-of-interest definition, anatomical labeling of functional activation clusters, structural and functional connectivity analyses, and as a reference framework for cross-study comparisons and meta-analyses in neuroimaging, providing a standardized anatomical scaffold for integrating multimodal brain data.

Atlas Website

Overview generated by GPT-4o (2026).


Click any region name to view its full visualization and description.

ID Acronym Region
1 Caudate caudate_bilateral
2 Cerebellum cerebellum_bilateral
3 Frontal Lobe frontal lobe_bilateral
4 Insula insula_bilateral
5 Occipital Lobe occipital lobe_bilateral
6 Parietal Lobe parietal lobe_bilateral
7 Putamen putamen_bilateral
8 Temporal Lobe temporal lobe_bilateral
9 Thalamus thalamus_bilateral

Citation

Wali Sidiqyar*, Gaurav Rudravaram*, Elyssa M. McMaster, Trent M. Schwartz, Adam M. Saunders, Kurt G. Schilling, Bennett A. Landman "Introducing SPINS: A Shared Public Visualization Library of Neuroanatomical Structures." Medical Imaging with Deep Learning- short paper

This resource is licensed under CC0 1.0 Universal (Public Domain).