Hi,

I was wondering what differences there could be between a glmfit regression performed on the whole surface, and the same done in an ROI.  I normally see very little difference, (other than the CWP being relative to that space when monte carlo is used for multiple comparisons correction) which is why I am surprised to see a big difference between ROI and whole surface resulting from the commands shown below.   The image attached is a medial snapshot of the unrestricted regression resulting from the first command shown below.  The second command, ROI-restricted, resulted in no activity in the ROI (no image attached because nothing to show).  I thought that maybe the smoothing had different effects in the ROI than in the unrestricted regression, so I also took out the smoothing as a test.  The results of this where just an unsmooth appearance of the whole surface version, and still nothing in the ROI version, so it can't be smoothing that causes these clusters to vanish.

Thanks in advance for your thoughts or help on this,
Rob

setenv SUBJECTS_DIR /space/ventzl/1/users/SUBJECTS_DIR
cd $SUBJECTS_DIR
source /usr/local/freesurfer/nmr-dev-env

# 1. showing that the ROI differs from the unrestricted version (nothing in ROI version)
mri_glmfit \
   --really-use-average7 \
   --y DTI_ADI_reg_glmfit/rh.FA.ADI_Com_ASD.fsfast.mgh \
   --glmdir DTI_ADI_reg_glmfit/rh.FA.ADI_Com_ASD.fsfast.glmdir \
   --fsgd fsgd_and_contrast_files/DTI_ADI_Com_age_ASD.fsgd \
   --C fsgd_and_contrast_files/DTI_ADI_age.mat \
   --surf average7 rh \
   --pca \
   --fwhm 4.6

mri_glmfit \
   --really-use-average7 \
   --y DTI_ADI_reg_glmfit/rh.FA.ADI_Soc_ASD.fsfast.mgh \
   --glmdir DTI_ADI_reg_glmfit/ROI/rh.FA.ADI_Soc_ASD.fsfast.glmdir \
   --fsgd fsgd_and_contrast_files/DTI_ADI_Soc_age_ASD.fsgd \
   --C fsgd_and_contrast_files/DTI_ADI_age.mat \
   --label average7/label/dACC_rACC_final-lh.label \
   --surf average7 rh \
   --pca \
   --fwhm 4.6

# 2. taking out smoothing: whole surface version is fuzzier, still nothing in ROI version
mri_glmfit \
   --really-use-average7 \
   --y DTI_ADI_reg_glmfit/rh.FA.ADI_Com_ASD.fsfast.mgh \
   --glmdir DTI_ADI_reg_glmfit/rh.FA.ADI_Com_ASD.fsfast.glmdir \
   --fsgd fsgd_and_contrast_files/DTI_ADI_Com_age_ASD.fsgd \
   --C fsgd_and_contrast_files/DTI_ADI_age.mat \
   --surf average7 rh \
   --pca

mri_glmfit \
   --really-use-average7 \
   --y DTI_ADI_reg_glmfit/rh.FA.ADI_Soc_ASD.fsfast.mgh \
   --glmdir DTI_ADI_reg_glmfit/ROI/rh.FA.ADI_Soc_ASD.fsfast.glmdir \
   --fsgd fsgd_and_contrast_files/DTI_ADI_Soc_age_ASD.fsgd \
   --C fsgd_and_contrast_files/DTI_ADI_age.mat \
   --label average7/label/dACC_rACC_final-lh.label \
   --surf average7 rh \
   --pca

---

-- 

Robert P. Levy, B.A.
Research Assistant, Manoach Lab
Massachusetts General Hospital
Charlestown Navy Yard
149 13th St., Room 2611
Charlestown, MA 02129
email: levy@nmr.mgh.harvard.edu
phone: 617-726-1908
fax: 617-726-4078

http://nmr.mgh.harvard.edu/manoachlab