Hi FSLexperts, I would really appreciate your help in answering some of my questions.
I have managed to segment the subcortical shapes using FSL with the run_first_all tool and concatenated the bvars files.
Because of the nature of the experiment, which is 2 groups (8 controls vs 7 patients) and 2 time points (t1 and t2), I chose the 2 way mixed effect ANOVA as recommended in the FSL forums.
Is my design matrix and contrast correct to do shape analysis? The primary aim is the find out if shape changes in the patient group is significantly greater than the shape changes in the control group over time.
This is the design matrix I used
Subj time Grp EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8 EV9 EV10 EV11 EV12 EV13 EV14 EV15 EV16 EV17
1 t1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 t2 1 -1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
2 t1 2 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
2 t2 2 -1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
3 t1 3 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
3 t2 3 -1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
4 t1 4 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
4 t2 4 -1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
5 t1 5 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
5 t2 5 -1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
6 t1 6 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
6 t2 6 -1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
7 t1 7 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
7 t2 7 -1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
8 t1 8 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
8 t2 8 -1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
9 t1 9 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0
9 t2 9 0 -1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0
10 t1 10 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
10 t2 10 0 -1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
11 t1 11 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
11 t2 11 0 -1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
12 t1 12 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
12 t2 12 0 -1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
13 t1 13 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
13 t2 13 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
14 t1 14 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
14 t2 14 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
15 t1 15 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
15 t2 15 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
Contrasts used
C1 t1-t2 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
C2 t2-t1 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
I then ran first_utils using the following command.
first_utils --usebvars --vertexAnalysis -i L_putamen_all.bvars -o diff_con8_pat7_L_puta -d design_con8_pat7.mat --useReconMNI
From the results of the first_utils, I used fslsplit to separate the nii.gz file and did the subtraction of t1-t2 as follows
fslmaths subject1_t1 -sub subject1_t2 sub1_diff
fslmaths subject2_t1 -sub subject2_t2 sub2_diff
(repeat for subjects 3-15)
fslmerge -t t1_minus_t2 sub1_diff sub2_diff sub3_diff sub4_diff sub5_diff sub6_diff sub7_diff sub8_diff sub9_diff sub10_diff sub11_diff sub12_diff sub13_diff sub14_diff sub15_diff
Following this, I designed another design matrix to input my data into the two-sample t-test to run randomise, for which I use the following design matrix (Con refers to controls and Pat refers to patients)
Subj Grp EV1 EV2
1 1 1 0
2 1 1 0
3 1 1 0
4 1 1 0
5 1 1 0
6 1 1 0
7 1 1 0
8 1 1 0
9 1 0 1
10 1 0 1
11 1 0 1
12 1 0 1
13 1 0 1
14 1 0 1
15 1 0 1
Contrasts used
EV1 EV2
C1 Con>Pat 1 -1
C2 Pat>Con -1 1
randomise -i t1_minus_t2.nii.gz -o TwoSampT -d design_2Ttest_con8_pat7.mat -t design_2Ttest_con8_pat7.con -m diff_con8_pat7_L_puta_mask.nii.gz -T
Is it correct to use this matrix for the randomise step? Or should I continue to use the matrix from earlier for the 2-way Mixed Effect ANOVA.
I am not sure at this part which mask to input. I am using the nii.gz file that I created from fslmerge from the subtractions as the input but it does not have a mask. Instead, I used the mask generated from the earlier first_utils command.
I get 2 nii.gz files as the output. "TwoSampT_tfce_corrp_tstat1" and "TwoSampT_tfce_corrp_tstat2". Based on my steps above, am I right to understand that the file with tstat1 shows statistically significant hypertrophy of patients compared to controls and the file with tstat2 shows atrophy of patients vs controls?
I really appreciate your help in answering my questions, thank you!
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|