Related Experiment Video
Updated: Oct 4, 2025
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
Measuring variability of local brain volume using improved volume preserved warping.
Xuzhou Li1, Manli Huang2, Xuejun Hao3
1Key Laboratory of Brain Functional Genomics (MOE and STCSM), Shanghai Changning-ECNU Mental Health Center, Institute of Cognitive Neuroscience, School of Psychology and Cognitive Science, East China Normal University, Shanghai, China; Molecular Imaging and Neuropathology Division, Department of Psychiatry, Columbia University and New York State Psychiatric Institute, 1051 Riverside Drive, New York, NY 10032, USA.
New Voxel Preserved Warping (VPW) methods reliably measure local brain volume changes and variability, overcoming inconsistencies in traditional neuroimaging analysis. These VPW approaches ensure stable and dependable local brain volume variability assessments across different co-registration techniques.
Area of Science:
- Neuroimaging
- Medical Image Analysis
- Computational Anatomy
Background:
- Accurate measurement of local brain volume changes and variations (LBVCV) is crucial in neuroimaging studies.
- Current methods like Voxel Preserved Warping (VPW) and Jacobian determinants rely on local deformations but neglect global deformations during co-registration, leading to inconsistencies.
- Existing co-registration strategies lack a unified approach for global and local transformations, compromising the reliability of LBVCV measurements.
Purpose of the Study:
- To develop novel VPW approaches (VPWα and VPWβ) that integrate both global and local deformations for robust LBVCV measurement.
- To address the un-uniqueness and instability in LBVCV assessments caused by diverse co-registration strategies.
- To provide a registration-independent method for reliable local brain volume change analysis.
Main Methods:
- Proposed new VPW approaches (VPWα and VPWβ) that utilize general deformation concatenating global and local components.
- Validated the new VPW methods using simulated and real-world neuroimaging data.
- Compared the proposed methods against traditional VPW approaches using Automatic Registration Toolbox (ART) and Symmetric Image Normalization Method (SyN) registration.
Main Results:
- Experiments with simulated data showed that the new VPW methods reliably measure local brain volume changes and variability.
- Traditional methods produced inconsistent and potentially false findings in LBVCV maps.
- Real neuroimaging data from a schizophrenia study demonstrated high consistency of results using the proposed VPW methods, irrespective of the registration technique employed, unlike traditional methods.
Conclusions:
- The proposed VPWα and VPWβ methods offer a reliable and stable approach for measuring local brain volume changes and variability.
- These new methods are independent of specific registration techniques, resolving the unreliability issues associated with traditional LBVCV assessments.
- The developed VPW approaches can serve as dependable alternatives for LBVCV assessment in neuroimaging research.
