Hello,
I have two group of subjects first group say (A) has subjects with brain
atrophies and second group say (B) has subjects which are controls.
I ran TBSS on them as recommended by FSL. I then run randomise with TFCE
option set.
I am trying to run unpaired t test between two groups and I am interested at
looking at differences in the FA tracts of group B > group A i.e. controls >
patients
Following is my design matrix e.g. .
Group Group A Group B
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
Contrast
A > B 1 -1
B > A -1 1
After completing the stats step (i.e. creating the design matrix and
running randomise) when I look at the result I get to see differences in
both p value images (uncorrected )
meaning
tbss_tfce_p_tstat1.nii.gz
tbss_tfce_p_tstat2.nii.gz
show differences. I am expecting to see differences between the two groups
in only one of these images which is Controls > Patients
I am looking at the images in fslview at with the following intensity range;
min (0.95) to max (1) for p < 0.05
So my question is , why would tbss stats show me differences in the opposite
direction meaning patients > controls when I know that this is not true ? I
am goofing up something big time
Kindly advise.
Thank you in advance
Sincerely,
Ankit Master
Programmer
Mayo Clinic
|