Related Experiment Video
Updated: Jul 28, 2026
09:33
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
Simulation of the effects of global normalization procedures in functional MRI
Maria Gavrilescu1, Marnie E Shaw, Geoffrey W Stuart
1Howard Florey Institute, University of Melbourne, Australia.
Neuroimage
|October 16, 2002
Summary
Global signal normalization in resting-state fMRI affects statistical sensitivity. Proportional scaling and ANCOVA methods reduce sensitivity and introduce artifacts, while masking and orthogonalization methods show promise but have limitations.
Area of Science:
- Neuroimaging
- Functional Magnetic Resonance Imaging (fMRI)
Background:
- Global signal normalization is a common preprocessing step in resting-state fMRI analysis.
- The choice of normalization method can impact statistical results and interpretation.
Purpose of the Study:
- To evaluate the impact of five global normalization methods on statistical sensitivity and false-positive rates.
- To compare grand mean session scaling, proportional scaling, ANCOVA, masking, and orthogonalization methods.
Main Methods:
- Simulated resting-state fMRI data with known activation patterns were used.
- Five distinct global normalization techniques were applied to the data.
- Sensitivity and false-positive rates were assessed for each method.
Main Results:
- Proportional scaling and ANCOVA significantly decreased statistical sensitivity.
- These methods also induced artifactual deactivation, even with low global signal correlation.
- Masking and orthogonalization methods demonstrated better performance but are condition-dependent.
Conclusions:
- The choice of global normalization method critically influences resting-state fMRI analysis outcomes.
- Proportional scaling and ANCOVA should be used with caution due to potential biases.
- Masking and orthogonalization offer alternatives, but their applicability requires careful consideration of experimental design.
