Dear SPMers,
I am analyzing data from fMRI studies by implementing connectivity techniques (CONN, DCM) in psychiatric populations.
I am trying to improve my preprocessing spatial pipeline in SPM.
I have a question about segmentation and explicit masking in fMRI specification.
1) I would like to create a binary brain mask generated from the segmented and normalized tissues with IMCALC (i1+i2+i3)>0.3. Do you think that CSF could be removed (ie. (i1+i2)>0.2)?
I would also change the default value in fMRI specification step for threshold from 0.8 to -Inf. Do you think that could be a good approach? Do you have any suggestions or comments about the two different thresholds (for brainmask in imcalc and explicit mask in fMRI specification)? I am afraid to be too inclusive and increase too much the number of comparisons.
2) Segmentation: entering the T1 previously coregistered to mean EPI and bias corrected in a previous segmentation, I want to segment the different tissues, I found that the better approach is to warp using unmodulated option, by providing better outcome for density. Is it right? However, is it possible to set Modulated+unmodulated option. Is it a better option?
Thank you and all the best,
Benedetta Vai.
Benedetta Vai, PhD
Postdoctoral research fellow
Department of Clinical Neurosciences
IRCCS Ospedale San Raffaele
20127, Milan, Italy
|