Hi Shannon, sorry for the delay. There are several ways you can do this. If you already have a slope for each subject, then just use that as input and use the simple two group FSGD file. If you really want to do the RM ANOVA, then there are two ways. The simpler way is to create a "stack" of all subjects for each visit (eg, v1, v2, and v3) using mris_preproc. Next compute paired differences d12 = v1-v2 and d13 = v1-v3 using fscalc (note that you do not need v2-v3). Next, create a new stack from d12 and d13 with mri_concat (eg, mri_concat d12.mgh d13.mgh --o d12+d13.mgh). Then create an FSGD file with a class for each group and two variables (d12 and d13)). The structure will look something like this:
Input subject1 G1 1 0 Input subject2 G1 1 0 ... rest of subjects for group 1 ... Input subject1 G1 0 1 Input subject2 G1 0 1 ... rest of subjects for group 1 ... Input subject1 G2 1 0 Input subject2 G2 1 0 ... rest of subjects for group 2 ... Input subject1 G2 0 1 Input subject2 G2 0 1 ... rest of subjects for group 2 ...
Run the analysis with DOSS. You can use the following contrast matrix to test for an interaction: [1 -1 -1 1]
Let me know if that makes sense or not doug
On 08/23/2012 09:39 PM, Shannon Kogachi wrote:
Hi, Is there a way to set up a repeated measures ANOVA and ANCOVA with two different groups (ex: study group v. controls) and each subject has more than two timepoints? Would it be possible to assign each Input in the fsgd file to two classes (ex: timepoint scan and group) or is there another way to set up the fsgd file and contrast matrices? I've already used qdec in freesurfer 5.1 to calculate rates of change and symmetrized percent changes between the timepoints (using long_mris_slopes) but now would like to perform a repeated measures test. Thank you!
-Shannon K.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer