Hi,
We came across some strange effects when estimating voxel-wise regression models by the 2nd level Bayesian method implemented in SPM. Attached you find a document where we set up a small simulation study that demonstrates that these artifacts are false negatives.
FYI: We posted a message with regard to this issue on this list some time ago. However, this time we provide MATLAB code that can be used to reproduce our findings.
Best
Paul
|