Related Experiment Video
Updated: May 31, 2026
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Statistical adjustments for brain size in volumetric neuroimaging studies: some practical implications in methods
Liam M O'Brien1, David A Ziegler, Curtis K Deutsch
1Department of Mathematics and Statistics, Colby College, Waterville, ME 04901, USA. lobrien@colby.edu
Psychiatry Research
|June 21, 2011
Summary
Statistical adjustment methods for volumetric MRI brain data vary. This study proposes a generalized approach with graphical tools to better analyze brain structure size, accounting for individual variability in neurological and psychiatric research.
Area of Science:
- Neuroimaging
- Biostatistics
- Neuroscience
Background:
- Volumetric magnetic resonance imaging (MRI) is crucial for identifying brain structure differences in neurological and psychiatric disorders.
- High individual variability in brain size complicates the detection of significant structural changes.
- Existing statistical methods for adjusting brain size have differing assumptions and limitations.
Purpose of the Study:
- To examine the theoretical basis of three common adjustment methods: proportion, residual, and analysis of covariance.
- To propose a generalized modeling strategy for adjusting volumetric MRI data for brain size.
- To provide graphical tools to aid researchers in understanding and applying these adjustment methods.
Main Methods:
- Theoretical examination of proportion, residual, and analysis of covariance methods.
- Application of these methods to a volumetric MRI dataset.
- Development of graphical tools for assessing method agreement and differences.
Main Results:
- The three common methods are specific cases of a proposed generalized approach.
- Significant differences exist in how these methods uncover relationships between brain structure volumes and head size.
- Graphical analyses effectively illustrate these differences and aid in uncovering potential relationships.
Conclusions:
- A generalized modeling strategy, incorporating graphical analyses, is recommended for adjusting volumetric MRI data.
- This approach helps researchers better understand and account for individual variability in brain size.
- The proposed method provides a robust framework for analyzing structural brain differences in disease research.
