Thanks for the feedback! (I shoot another email to the marsbar mail list with another request...).
Yes, the percent signal change does not take into account other "regressors", nevertheless, I am under the impression that I have seen some papers that have used subtraction between percent signal change for the conditions of interest and I still find striking that it lacked correlation with the con image values for the same ROI.
Another source of discrepancy was from pulling the con values from a second level SPM.mat file, versus directly at an individual level from a con image (both using marsbar).
Please correct me if I am wrong, but I understand that if I pull the info from a second level it will give me information from the shared voxels across all subjects for that specific ROI, and that's why I may have found different values for the same con image depending on whether I pull those values from the second level SPM.mat file versus at an individual level, is this accurate?
Thanks tons!
V
|