Tensor-based morphometry with stationary velocity field diffeomorphic registration: application to ADNI

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Computational Anatomy

Background:

• Tensor-based morphometry (TBM) relies on accurate inter-subject registration for analyzing anatomical differences.
• Previous TBM studies often overlooked the impact of registration parameter choices on results.
• Optimizing registration is crucial for reliable voxel-wise statistical analysis in neuroimaging.

Purpose of the Study:

• To evaluate the performance of TBM using stationary velocity field (SVF) diffeomorphic registration.
• To investigate the influence of registration parameters on TBM analysis of brain structure.
• To generate high-resolution brain atrophy maps for Alzheimer's Disease Neuroimaging Initiative (ADNI) data.

Main Methods:

• Employed stationary velocity field (SVF) diffeomorphic registration for non-rigid inter-subject image alignment.
• Systematically explored a wide range of registration parameters, including deformation smoothness and regularization.
• Applied TBM to a subset of subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) study.

Main Results:

• Achieved highly detailed anatomical resolution in brain atrophy maps.
• Demonstrated a high level of statistical significance in detected voxel-wise differences.
• Outperformed a previously used non-linear elastic registration method on the same dataset.

Conclusions:

• Optimized SVF-based TBM provides a robust method for detecting subtle brain structural changes.
• The choice of registration parameters significantly impacts the sensitivity and specificity of TBM analyses.
• This refined TBM approach offers improved capabilities for neurodegenerative disease research.