Hello Everyone
I apologise in advance if my questions are very naive, as I am new to DTI and do not comprehensively understand the mathematics behind TBSS.
I would like to do a single group TBSS analysis with about 100 subjects (all healthy individuals). I want to answer the question: does white matter vary significantly with certain demographic factors
I want to do a single group analysis and check for variation with 5 demographic factors and Age, Handedness & Gender. I want the demographic factors to be the covariates of interest, while the other 3 are covariate of no interest.
I have followed all the pre-processing and TBSS steps from the FSL website & set up my design matrix (using GLM) to have 8 columns and my contrast as follows:
1 0 0 0 0 0 0 0
-1 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0
0 -1 0 0 0 0 0 0
0 0 1 0 0 0 0 0
0 0 -1 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 -1 0 0 0 0
0 0 0 0 1 0 0 0
0 0 0 0 -1 0 0 0
After this I have run Randomize (as per the instructions on FSL website) & then I am looking at the results as
fslview $FSLDIR/data/standard/MNI152_T1_1mm mean_FA_skeleton -l Green -b 0.2,0.7 tbss_tfce_corrp_tstat1 -l Red-Yellow -b 0.99,1
I believe this shows 1-p, that is thresholded at a significant level of 0.01.
I get a lot of significant areas if I use this threshold but almost none is I make it more stringent
I want to know if I am doing the right thing? Is there any way to verify my procedure?
I would really be grateful if anyone here could offer any advice about this. Thank you for your advice and your time
best wishes
Prerona.
|