Paired analysis: use of covariates - Freesurfer

9 Dec 2013


      Dear FreeSurfer experts,
I'm trying to implement the steps you indicate in the Paired Analysis' wiki site, but I still don't get something: 
In this site, you show the next FSGD file: 
GroupDescriptorFile 1
Class Main
Variables Age
Input subject1pair   Main 30
Input subject2pair   Main 40
Input subject3pair   Main 50
Input subject4pair   Main 60
Accordingly, "Age" is a variable whose value is the same for each pair of subjects (i.e., it is a measure of between-pair differences, but does not contain information about differences found in subjects of the same pair). Nevertheless, I need to introduce some covariates of interest whose value would indicate within-pair differences (i.e., the difference between "subjectN" and "subjectNmatch", where "N" is the arbitrarily assigned pair number). I made the next FSGD, but I'm wondering it's not correct (is it?):
GroupDescriptorFile 1                        
Title trait1_trait2_paired                        
MeasurementName thickness                        
Class male
Class female                        
Variables Age trait1_diff trait2_diff                        
Input    subject1pair    female    24    150    1850
Input    subject2pair    male    41    -200    -2090
Input    subject3pair    female    48    50    -6508
Input    subjec4tpair    male    53    -100    -7100
Input    subject5pair    female    36    150    1950
.
.
.
In this FSGD, the last two variables ("trait1_diff" and "trait2_diff") would represent within-pair differences. As you can see, there are both positive and negative values, depending on the random order of individuals within a pair. Afterwards, contrast files such as "0 0 0 0 0.5 0.5 0 0" or "0 0 0 0 0 0 0.5 0.5" would be used to check whether "trait1_diff" or "trait2_diff" have an effect on cortical thickness, regressing out the effects of "gender", "Age" and either "trait2_diff" or "trait1_diff".
Could you please tell me if the FSGD table is correct? If it isn't, how to implement such an approach? (I.e., how to use both between-pair and within-pair variables?) Also, are the contrast files mentioned above correct?
Thank you in advance, 
Aldo