Hi all, sorry to post again, but I do need some help. I found myself some answers for my questions in the following post. Please correct me if I am wrong.
==================================
Hi all, I am rather new to the community. Hope I am not asking too simple questions.
I have a group of 161 subjects. I am interested in the age effect from the vbm results. I used the FEAT to generate the design.*. My questions are:
1. Do I put in an extra predictor of ones to remove the mean?
My ans: If I am interested in the effect of a continuous variable, I should put in a column of ones to remove the mean.
2. Is it necessary to put in -D for the randomise command, even in the fslvbm3b?
My ans: If I have put in extra column to demean in the previous question, I don't need to give -D in randomise. Because fslvbm3b is generated automatically, it is better I put a demean column in my design rather than change the script to add in -D.
3. I don't see integer in the .mat file for the age column, why?
My ans: The variable is demeaned.
Following are the contents of my design.con and design.mat
===================================
/ContrastName1 "Age"
/NumWaves 2
/NumContrasts 1
/PPheights 3.200000e+01
/RequiredEffect 1.401
/Matrix
0.000000e+00 1.000000e+00
===================================
/NumWaves 2
/NumPoints 161
/PPheights 4.535950e-30 3.200000e+01
/Matrix
4.535950e-30 -1.987578e+00
4.535950e-30 -1.987578e+00
4.535950e-30 -9.875776e-01
4.535950e-30 7.012422e+00
4.535950e-30 -8.987578e+00
4.535950e-30 -6.987578e+00
4.535950e-30 -4.987578e+00
4.535950e-30 -3.987578e+00
4.535950e-30 -7.987578e+00
4.535950e-30 -5.987578e+00
... ...
Any advice will be much appreciated.
Zheng Hui
|