On 02/24/2018 12:00 AM, Martin Juneja wrote:
Hi experts,
I am not sure if this has already been addressed in FS discussion.
If my fsgd file looks like following:
Class Male Class Female Variables Performance Age Input S1 Male 181 25 Input S2 Female 167 23 .... ....
I understand that the contrast matrix *[**0 0 1 -1 0 0]* represents whether there is a difference between the group (male/female)-performance slopesregressing out the effect of age. In other words, this matrix calculates whether two slopes i.e. one for males: thickness-performance slope and second for females: thickness-performance slope - are significantly different after regressing out the effect of age.
My questions are following:
(1) Is my above interpretation of contrast matrix*[**00 1 -1 0 0]* correct?
Yes
(2) If my interpretation is correct, then I was wondering what will be the contrast matrix if *I want to check whether there is **significant effect of age on e.g. thickness-performance slopes after regressing out the effect of gender.* And what will be the corresponding fsgd file for that, any changes in that?
This is an interaction between continuous variables, which can be a little tricky. You cannot do it with the model that you have. The typical way to do this is to create a third covariate which is the product of Performance*Age; you should remove the mean of Performance and Age before doing the product.
Thanks.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer