Dear Cyril,
Sorry for this basic questions, but I should ask again to clear things to me.
"[0 0 1 -1] [0 0 -1 1] difference in the correlations"
Let's say the first group is brain image at the timepoint 1 and the second group is image at the timepoint 2.
we measured same clinical scores at two timepoints, so every subject has two images and two scores.
so, difference of clinical scores at two timepoints(pre-post) would represents behavioral improvement(or deterioration).
1-1. Then, would the contrast [0 0 1 -1] shows the areas of which cortical volume associated with behavioral improvement?
vise versa, the contrast [0 0 -1 1] shows the areas of which cortical volume associated with behavioral deterioration?
1-2. If so, with the contrast [0 0 1 -1] masked by [1 -1], I would say that more increase of cortical volume in the area is associated with better behavioral improvement. is it ok?
3. How about using the difference of effect size(y) between two timepoints as a covariate?(to perform sort of correlation with behavioral improvement) though I feel it's wrong way to do it.
many thanks in advance,
Hyojeong
Hyo-Jeong LEE ------------------------------------------------------------------------------
Associate professor
Department of Otolaryngology, Hallym University College of Medicine
896, Pyeongchon-dong, Dongan-gu, Anyang-si, Gyeonggi-do, 431-070
Republic of Korea
Phone: +82-31-380-3842/1670 Fax: +82-31-386-3860
e-mail: [log in to unmask]
-------------------------------------------------------------------------------------------------------
|