Global Confounds in SPM Models

An SPM model can have controls for global factors with ANCOVA, grand mean scaling, and global normalization that is via ANCOVA or proportional.

Global scaling

In a nutshell: never use

Do not use. It aims to correct for drift over time. This is a remnant from the PET days.

Global normalization

In a nutshell: use for VBM OR add TIV as covariate, never for fMRI, usually never for quantitative MRI

This option controls for global values by either scaling the images or including the global mean as a covariate.

VBM: for analyzing regional gray matter volume with VBM, global normalization is about dealing with different brain sizes; bigger brains are likely to have bigger structures. This option corrects for global brain volume.

General: the default global calculation is the mean over the brain mask. In SPM, you can specify global values manually.

ANCOVA regressors in design

In a nutshell: never for fMRI, maybe for repeated measures VBM and quantitative MRI

In some designs (multiple groups, 2 sample or ANOVA), you can specify "ANCOVA-by-factor" regressors. The factor would usually be the group, and the covariate the global value (mean over the mask).

"These options allow different subjects to have different relationships between local and global measurements." If not repeated measures, it will do the group x global effects.

References

Dartmouth wiki: useful compilation about global scaling.