Probabilistic MRI Brain Anatomical Atlases Based on 1,000 Chinese Subjects

Xing Wang, Nan Chen, Zuo ZhenTao, Rong Xue, Jing Luo, Yan Zhuo, Dinggang Shen, Kuncheng LiPLoS 2013 · DOI

Abstract

Brain atlases provide a standard reference coordinate system for neuroscience research. However, most existing human brain atlases are derived from Western populations, which may not accurately represent Eastern brain anatomy. This study constructed 10 new Chinese brain atlases for different ages and genders using MRI scans from 1,000 healthy volunteers (ages 18–70). Using the HAMMER registration algorithm, these population-specific templates capture the structural characteristics of the Chinese population, offering a more precise tool for basic neuroscience and clinical diagnosis in Eastern subjects.

Introduction: Addressing the Western Bias in Neuroimaging

In modern neuroscience, brain atlases serve as the fundamental "GPS" for localizing structural and functional data. For decades, the field has relied on templates like the Talairach atlas or the Montreal Neurological Institute (MNI) templates. While invaluable, these standards are primarily based on Western Caucasian subjects.

Significant anatomical variations exist across different ethnic populations due to genetic and environmental factors. Using Western-centric templates to analyze Eastern subjects often leads to "mislocalization" errors, where brain activity or lesions are incorrectly mapped. This study addresses this gap by developing a comprehensive set of 10 age- and gender-specific atlases based on a large cohort of 1,000 Chinese subjects.

Anatomical dimensions in axial, coronal, and sagittal views, highlighting the specific structural characteristics of the Chinese brain templates.

Methodology: High-Precision Registration with HAMMER

To create these atlases, the researchers employed the Hierarchical Attribute Matching Mechanism for Elastic Registration (HAMMER). Unlike traditional linear registration methods, HAMMER is a non-rigid algorithm designed to handle the complex, convoluted geometry of the human cortex.

The process involved:

  1. Attribute Vector Assignment: Every voxel is assigned a vector representing local geometric features.
  2. Hierarchical Alignment: The algorithm first aligns voxels with highly distinctive features (like major sulci) to establish a global framework before refining the registration for less distinctive areas.
  3. Template Selection: To minimize bias, the initial template for each group was selected from an individual whose brain volume was closest to the group's mean.

Results: Structural Insights and the 10-Atlas Set

The study successfully produced 10 probabilistic atlases, stratified by gender and five age brackets (18–30, 31–40, 41–50, 51–60, and 61–70).

The complete set of 10 probabilistic brain atlases, showing the progression across five age groups for both females (F) and males (M).

Key Findings:

  • Morphological Differences: Compared to MNI templates, Chinese brains were found to be generally shorter in length and height but wider relative to their length.
  • Volume and Aging: The data revealed distinct aging patterns. Male brain volumes peaked earlier (around age 50) and showed a more rapid decline compared to females, whose volumes tended to remain more stable until approximately age 55.
  • Sexual Dimorphism: While males exhibited larger absolute brain volumes across all age groups (e.g., ~1755 cm³ vs ~1660 cm³ in the 18-30 group), the rate of age-related atrophy was more pronounced in the male cohort.

Conclusion and Future Directions

These 10 atlases provide a more accurate anatomical reference for the Chinese population, potentially improving the precision of fMRI studies and image-guided neurosurgeries in Eastern countries.

Future work aims to transition from these "static" age-group snapshots to 4D probabilistic atlases that can model continuous brain changes over a lifespan. Additionally, integrating these anatomical templates with functional mapping (such as population-specific Brodmann areas) and utilizing higher-resolution 3.0T MRI data will further enhance the utility of these tools for clinical and basic research.

Summary updated

Abstract

Brain atlases are designed to provide a standard reference coordinate system of the brain for neuroscience research. Existing human brain atlases are widely used to provide anatomical references and information regarding structural characteristics of the brain. The majority of them, however, are derived from one paticipant or small samples of the Western population. This poses a limitation for scientific studies on Eastern subjects. In this study, 10 new Chinese brain atlases for different ages and genders were constructed using MR anatomical images based on HAMMER (Hierarchical Attribute Matching Mechanism for Elastic Registration). A total of 1,000 Chinese volunteers ranging from 18 to 70 years old participated in this study. These population-specific brain atlases represent the basic structural characteristics of the Chinese population. They may be utilized for basic neuroscience studies and clinical diagnosis, including evaluation of neurological and neuropsychiatric disorders, in Chinese patients and those from other Eastern countries.
BibTeX
@inproceedings{wang2013probabilistic,
  author = {Xing Wang and Nan Chen and Zuo ZhenTao and Rong Xue and Jing Luo and Yan Zhuo and Dinggang Shen and Kuncheng Li},
  title = {{Probabilistic MRI Brain Anatomical Atlases Based on 1,000 Chinese Subjects}},
  booktitle = {PLoS},
  year = {2013},
  doi = {10.1371/journal.pone.0050939},
}
Figure