hoa-2 / subcorticalparcellations

Harvard-Oxford Atlas 2.0 Subcortical parcellations of Human Connectome Project (HCP) subjects

Harvard-Oxford Atlas 2.0: Subcortical parcellations

Release version: 2.0.0

Version 2.0.0 increases the number of subjects for the discrete brain segmentation to 100. The probablistic segmentation still includes only the first 50 subjects; it will be updated in the next minor release.

Introduction

The Harvard-Oxford Atlas 2.0 project is developing a state-of-the-art, high-resolution full brain MR atlas consisting of 200 high definition MR data sets from the Human Connectome Project (HCP) parcellated into 392 gray and white matter PUs using structural MRI and 189 white matter fascicles using diffusion MRI. Manual parcellation will be performed based on a developed neuroanatomical ontology using a principled and well-defined parcellation methodology.

Dataset

This dataset consists of 50 manual parcelations of MRI datasets from The Human Connectome Project (HCP).

Files in the dseg directory are discrete parcelations of the subject data. Image files are in nifti format and GZIP compressed. File numbers correspond to the HCP subject number from the HCP1200 dataset.

In the lut directory, the dseg.tsv file describes the label values, structure names, and suggested colors using the BIDS standard format. The same file is stored in ctbl format for use with 3D Slicer.

The probseg directory contains a probablistic segmentation of the atlas. See the README in that directory for more information.

Software

Datasets have been authored using 3D Slicer using custom modules developed by the Center for Morphometric Analysis (CMA) at Massachusetts General Hospital and Brigham and Women's Hospital in Boston.

Source files

The parcellation files are derived from 3D Slicer MRB file. Because of HCP data use agreements covering the underlying datasets, we cannot publicly post the MRB files that contain that data. They are available by request.

Citation

Rushmore R.J., Sunderland, K., Carrington H., Chen J., Halle M., Lasso A., Papadimitriou G., Prunier N., Rizzoni E., Vessey B., Wilson-Braun P., Rathi Y., Kubicki M., Bouix S., Yeterian E. and Makris N. (2022) Anatomically curated segmentation of human subcortical structures in high resolution magnetic resonance imaging: An open science approach. Front. Neuroanat. 16:894606. doi:10.3389/fnana.2022.894606

Sponsorship

NIH NIMH 5R01MH112748-04 High Resolution, Comprehensive Atlases of the Human Brain Morphology

License

This data is licensed under the 3D Slicer license, part B.

https://raw.githubusercontent.com/Slicer/Slicer/main/License.txt

Past version history

• v1.0.0: Initial release of the discrete segmentation of 50 subjects.
• v1.1.0: Initial release of the probability atlas for the subjects in v1.0.0.
• v1.1.1: Minor update of this README, no data files have changed.

subcorticalparcellations's Issues

Three subjects using wrong label numbers

In the dseg label files of the following three subjects:

192540
361941
523032

the following structures are labeled with the wrong label numbers:

Left-Inf-Lat-Vent
Right-Inf-Lat-Vent
Left-Hippocampus
Right-Hippocampus
Left-Amygdala
Right-Amygdala
Left-Caudate
Right-Caudate
Left-Putamen
Right-Putamen
Right-Thalamus
Left-Pallidum
Right-Pallidum
Left-Accumbens
Right-Accumbens
Right-VentralDC

For example, if you look at the right hippocampus, its label number is 20, indicating left hippocampus. similarly, the left hippocampus has label number 19, indicating right inferior lateral ventricle. Discovered when comparing against the FreeSurfer aseg.mgz file created for these three subjects.