Automated morphometric study of brain variation in XXY males

Dinggang Shen, Dengfeng Liu, Hong Liu, Liv S. Clasen, Jay N. Giedd, Christos DavatzikosNeuroImage 2004 · DOI

TL;DR

This study introduces an automated method for analyzing brain structure to identify subtle volumetric differences in individuals with an XXY chromosome arrangement (Klinefelter syndrome). The findings reveal significant reductions in gray and white matter volume in specific brain regions, offering a more precise understanding of the neurobiological underpinnings of XXY-related cognitive and behavioral variations.

What Problem Does This Paper Address?

Individuals with an XXY chromosome arrangement, a common genetic abnormality, often exhibit a range of physical, cognitive, and behavioral differences compared to typical males. While previous research has pointed to brain alterations in these individuals, studies have often relied on manual methods for measuring brain structures. These manual approaches are time-consuming, subjective, and can limit the number of subjects and brain regions that can be analyzed, potentially leading to less precise localization of differences and confounding results. This paper aims to address this by developing and applying an automated, voxel-based morphometric approach to identify and precisely locate brain structural variations in XXY males.

Methods

The core of this study is an automated, voxel-based morphometric protocol designed for analyzing brain structure. The method involves several key steps:

  • Non-brain voxel removal: An automated skull-stripping technique is used to remove non-brain tissue from MRI scans.
  • Tissue segmentation: Brain voxels are classified into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) using a fuzzy segmentation method that also corrects for image inhomogeneities.
  • Spatial normalization: A high-dimensional, mass-preserving spatial normalization algorithm called HAMMER (Hierarchical Attribute Matching Mechanism for Elastic Registration) is employed. This crucial step aligns individual brains to a standardized template, accounting for anatomical variability and enabling voxel-wise comparisons across subjects. The mass-preserving aspect ensures that local tissue density changes reflect actual volume changes without losing information.
  • Voxel-wise statistical analysis: Statistical parametric mapping (SPM) software is used to perform voxel-wise t-tests on the normalized tissue density maps, comparing the XXY group to the normal control group. This allows for the identification of regions with statistically significant differences in volume.

The novelty lies in the fully automated nature of the process, particularly the HAMMER algorithm for spatial normalization, which has been validated for its accuracy in aligning brain structures.

Datasets

The study utilized MRI data from:

  • 34 XXY males: Ranging in age from 5.3 to 19.2 years.
  • 62 normal XY male controls: Ranging in age from 5.3 to 20.8 years.

The data was acquired using a 1.5-T GE Signa scanner with specific T1-weighted imaging parameters. No new benchmark datasets were introduced; the study used existing subject data for analysis.

Key Insights and Ideas

  • Automated analysis unlocks finer detail: The automated, voxel-based approach allowed for a more precise spatial localization of brain volume differences compared to previous manual methods. This is crucial for understanding which specific neural circuits might be affected.
  • XXY brains show widespread atrophy: The study found significant reductions in gray and white matter volume in several brain regions in XXY males, suggesting a more pervasive structural impact than previously pinpointed.
  • Left hemisphere vulnerability: A greater number of atrophy clusters were observed in the left hemisphere, hinting at a potential specific vulnerability of this side of the brain in XXY individuals.
  • Structural changes correlate with known deficits: The identified atrophic regions align with known cognitive and behavioral characteristics of XXY individuals, such as language impairments (temporal gyri, insula) and issues with emotional regulation and spatial perception (limbic system, occipital/parietal lobes).

Key Results

  • Global volume differences: XXY males showed reduced overall gray matter (8.52% less) and white matter (8.05% less) volumes compared to controls. Conversely, ventricular CSF volume was significantly larger in XXY males (42.7% larger).
  • Localized atrophy: Pronounced and localized volume reductions were found in the insula, temporal gyri, amygdala, hippocampus, cingulate, and occipital gyri.
  • White matter involvement: While most significant differences were in gray matter, a cluster of atrophy was also identified in the right parietal lobe white matter.
  • No significant volume increase: No regions showed a significant increase in volume in the XXY group.
  • Statistical significance: Significance levels were set at P < 0.005 after correction for multiple comparisons.

Limitations and Future Work

  • Interpretation of negative results: The authors acknowledge that a negative result (no detected atrophy) does not definitively mean no difference exists, as statistical significance is influenced by normalization accuracy and variability.
  • Normalization accuracy: The accuracy of the spatial normalization process is critical. Errors in preprocessing steps like skull stripping and segmentation can affect the final results. Future work will focus on better understanding and quantifying these normalization errors.
  • Correlation with behavior: While the identified atrophic regions are hypothesized to relate to behavioral deficits, the study did not directly correlate specific structural changes with behavioral outcomes. Future research aims to investigate these correlations, for example, linking superior temporal gyrus atrophy to language comprehension difficulties.
  • White matter resolution: The study suggests that future research using diffusion tensor imaging (DTI) would be beneficial for a more detailed analysis of white matter structures.

Summary updated

BibTeX
@article{shen2004automated,
  author = {Dinggang Shen and Dengfeng Liu and Hong Liu and Liv S. Clasen and Jay N. Giedd and Christos Davatzikos},
  title = {{Automated morphometric study of brain variation in XXY males}},
  journal = {NeuroImage},
  year = {2004},
  doi = {10.1016/j.neuroimage.2004.08.018},
}
Figure