Voxel-Based Morphometry
Voxel-Based Morphometry (VBM) · Also known as: VBM, grey matter morphometry
Voxel-Based Morphometry (VBM) is a whole-brain statistical technique for detecting local differences in gray matter volume or concentration from structural MRI. Introduced by John Ashburner and Karl Friston in 2000, VBM enables researchers to identify regional brain volume changes associated with disease, aging, learning, and other factors without requiring a priori region-of-interest definitions.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
VBM is appropriate for group studies comparing brain morphology, when sample sizes are moderate to large (20+ subjects per group), and when hypothesis about regional volumetric changes exist. VBM is less suitable for highly focal lesions (alternative: manual tracing or lesion-specific tools), single-subject clinical assessment, or extreme anatomical variation. Use VBM when whole-brain, data-driven exploration is desired.
Strengths & limitations
- Whole-brain, hypothesis-free approach; detects unexpected regional effects not defined a priori
- High spatial resolution enables localization of volumetric changes to specific anatomical structures
- Fully automated pipeline; minimal user intervention reduces bias in region definition
- Accommodates multiple outcome variables (volume, concentration, density) in unified framework
- Statistically robust when sample size is adequate; group differences can be detected with moderate N
- Registration accuracy is critical; systematic registration errors can create spurious group differences
- Subject to multiple comparison inflation; thousands of statistical tests require stringent correction
- Interpretation blurred by complex registration and smoothing; spatial extent and localization uncertain
- Assumes tissue probability is spatially stationary after registration; violates true at region boundaries
Frequently asked
What is the difference between gray matter volume and concentration in VBM?
Volume reflects absolute gray matter quantity after accounting for local brain size. Concentration reflects the proportion of gray matter at each voxel, independent of overall brain size. Both are useful; choose based on hypothesis. Volume is more interpretable for true anatomical differences; concentration is more robust to registration variability.
Do I need to include total intracranial volume (ICV) as a covariate?
It depends on your hypothesis. If interested in absolute volumetric differences (true atrophy), include ICV as a covariate. If interested in relative concentration independent of head size, omit ICV. Using VBM concentration (not modulated volume) automatically accounts for brain size differences.
Why is smoothing necessary in VBM?
Smoothing (kernel convolution) reduces noise and increases statistical power by assuming neighboring voxels are more similar than distant ones. Smoothing also shifts statistics closer to normality. Trade-off: too much smoothing reduces spatial resolution; too little increases multiple comparison burden.
What sample size do I need for VBM?
Minimum 15–20 subjects per group for adequate power; better practice is 30+. Power depends on effect size; clinical effects (e.g., Alzheimer's vs. controls) typically need 20–40 per group, while subtle effects (normal aging) may require 50+. Conduct power analysis before data collection.
Sources
- Ashburner, J., & Friston, K. J. (2000). Voxel-based morphometry—the methods. NeuroImage, 11(6), 805–821. DOI: 10.1006/nimg.2000.0582 ↗
- Good, C. D., Johnsrude, I. S., Ashburner, J., et al. (2001). A voxel-based morphometric study of ageing in 465 normal adult human brains. NeuroImage, 14(1), 21–36. DOI: 10.1006/nimg.2001.0786 ↗
How to cite this page
ScholarGate. (2026, June 3). Voxel-Based Morphometry (VBM). ScholarGate. https://scholargate.app/en/neuroimaging/voxel-based-morphometry
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Structural Equation ModelingResearch Statistics↔ compare
- Tract-Based Spatial StatisticsNeuroimaging↔ compare