I have already preprocessed my data (including spatial smoothing at 4 mm) and conducted a GLM analysis. However, I now need a statistical map with spatial smoothing at approximately 6 mm.
Would it be appropriate to apply additional spatial smoothing (e.g., 3dBlurToFWHM) to the individual-level beta maps (already smoothed at 4 mm) and then run 3dttest+ on these blurred beta maps?
Or would it be necessary to redo preprocessing and GLM with a 6 mm smoothing kernel from the start?
As an ad hoc preprocessing step, the impact of spatial smoothing on fMRI data remains poorly understood. In other words, there is no definitive answer regarding the differences between the two smoothing methods you mentioned. While they are certainly not algebraically equivalent, their exact differences are unknown unless tested and compared empirically.
So you need spatial smoothing *at* approximately 6 mm. Does that mean a 6 mm FWHM kernel was applied in a different case, or that the estimated smoothness of the data should be 6 mm? If the latter, you are probably already there. After applying a 4 mm kernel, the current smoothness is probably 6 mm or higher already. It does not start at zero, but around the voxel size.
Thank you, what I meant is the latter. I tried adding a additional 4mm spatial smooth on the beta map, eventually yield to a around 6mm spatial smoothness of the data. Results actually seemed similar so I decided to remain at my first preprocessed data with spatial smooth at 4 mm, without doing anything else. Thank you so much for all the help.
All the Best,
Zhiqing
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.