Dear FSL experts,
I have a question about adding a behavioral measure to my TBSS design matrix for correlation purposes.
In the current design matrix, I have two groups with gender included as a covariate. The behavioral measure I want to correlate is from a psychophysics experiment. Not all participants we have DTI data for in the current design matrix have participated in the behavioral experiment. As such, I’m sure it is necessary to re-run TBSS with these individuals removed so that I can correlated FA with their performance on the task.
Below are the details of the current design.mat and design.con files, along with my plan of correlating EV4.
EV1 = Group 1, EV2 = Group 2; EV3 = Gender (males = 0). EV4 will be the performance score from the psychophysics experiment.
This is the Design.mat inputs (this part works so I did not include all 79 entries of the matrix to save time).
/NumWaves 3
/NumPoints 79
/PPheights 1 1
/Matrix
1 0 1
1 0 0
1 0 1
1 0 1
1 0 1
1 0 0
1 0 1
1 0 0
1 0 0
1 0 0
1 0 1
1 0 0
1 0 0
1 0 1
1 0 0
0 1 0
0 1 1
0 1 1
0 1 1
0 1 1
0 1 0
0 1 1
0 1 1
0 1 0
0 1 1
0 1 0
0 1 0
0 1 0
This is the Design.con file
/NumWaves 3
/NumContrasts 4
/PPheights 1 1
/Matrix
1 -1 0
-1 1 0
0 0 1
0 0 -1
This is what I am thinking I need to do in order to correlate behavioral performance with FA values. Please let me know if there is an error in my approach.
In the design.mat file, I need to change the number of waves from 3 to 4, alter the NumPoints to reflect the final sample size, and add an EV4. EV4 will then be populated with the scores from their performance.
In the design.con file, I need to again make the number of waves 4 for consistency between the files. Number of contrasts would remain a 4 since I am correlating, and not contrasting, correct? What would I add to ensure a correlation is run between the EV1 and EV2 contrasts with EV4 after controlling for EV3?
Unless I have missed something, the literature on the FSL GLM page and other questions asked on the user listserv have not been able to clarify this issue for me, and any help would be greatly appreciated!
Thank you!
Victor
|