Suppose an experimental design with 4 different conditions of a fast event-related paradigm.
Each condition is presented in a single separate run. So there are 4 runs and each run is paired with a unique condition.
Suppose to analyze data with a single GLM.
My question is: given the different baseline for each run, the result of general linear contrasts between conditions can not be trusted? Why, exactly?
Is this true also if we use 3dMVM to analyze data after 3dDeconvolve?
given the different baseline for each run, the result of general linear contrasts between conditions can not be trusted?
As long as you perform voxel-wise mean scaling during preprocessing, the effect estimate per condition can be approximately interpreted as percent signal change relative to the baseline. Therefore, there is no problem comparing conditions across runs.
Is this true also if we use 3dMVM to analyze data after 3dDeconvolve?
There are quite a few programs in AFNI for group analysis. The choice depends on the specific scenario and the strategy for multiple testing correction.
The
National Institute of Mental Health (NIMH) is part of the National Institutes of
Health (NIH), a component of the U.S. Department of Health and Human
Services.