Dear FSL users;
I’m using TBSS for a study of 4 subject groups - control, MCIna, MCIa, and AD. I would like to assess the effect of age on FA values within each group using randomise and was wondering if my design is okay.
In summmary my model is as follows:
There are 8 EV’s
EVs 1-4 are for each group, and EVs 5-8, are for the demeaned age of each group.
I set up contrasts as follows:
0 0 0 0 1 0 0.5 0 con_age
0 0 0 0 0 1 0 0 mcina_age
0 0 0 0 0 0 1 0 mcia_age
0 0 0 0 0 0 0 1 ad_age
0 0 0 0 0 0.5 0.5 0 mci_age
The above tests for relationships with age within each group – is that reasonable?
Below, I test for the slope with age being steeper in one group than the other. Again, does it seem reasonable?
0 0 0 0 1 -1 0 0 con_age>mcina_age
0 0 0 0 1 0 -1 0 con_age>mcia_age
0 0 0 0 1 -0.5 -0.5 0 con_age>mci_age
0 0 0 0 1 0 0 -1 con_age>ad_age
0 0 0 0 -1 1 0 0 mcina_age>con_age
0 0 0 0 -1 0 1 0 mcia_age>con_age
0 0 0 0 -1 0.5 0.5 0 mci_age>con_age
0 0 0 0 -1 0 0 1 ad_age>con_age
0 0 0 0 0 0 -1 1 ad_age>mcia
0 0 0 0 0 -1 0 1 ad_age>mcina
0 0 0 0 0 -0.5 -0.5 1 ad_age>mci
0 0 0 0 0 0 1 -1 mcia_age>ad_age
0 0 0 0 0 1 0 -1 mcina_age>ad_age
0 0 0 0 0 0.5 0.5 1 mci_age>ad_age
0 0 0 0 0 -1 1 0 mcina_age>mcia_age
0 0 0 0 0 1 -1 0 mcia_age>mcina_age
0 0 0 0 -1 0 0 1 ad_age>con_age
I am unclear about F tests, and whether or not they would help me to find out more about how age is affecting FA values in the different groups.
Finally, I am running this model using the command:
fsl_sub randomise -i ../../../all_FA_skeletonised.nii -o age_effect -m ../../../mean_FA_skeleton_mask -d age_effect.mat -t age_effect.con -n 5000 --T2
Cheers;
Larry
|