On 11/29/2023 2:16 PM, Tejaswi D. Sudhakar wrote:

External Email - Use Caution

Hi,

I would like to run a group analysis on 4 groups (patientM, patientF, controlM, controlF) with 3 covariates (age, cogscore1, cogscore2).

I am running with dods - so I would have 16 total regressors. How would I create contrast matrices for this?

I have this so far:

0.5 0.5 -0.5 -0.5 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0.5 0.5 -0.5 -0.5 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0.5 0.5 -0.5 -0.5

Yes, that will test for a difference between patients and controls for offset, cogscore1, and cogscore2

Secondly, I created lh.train.thickness.10B.mgh since the initial run was not qcached. How can I check if there is an error with this file?

There is not really a procedure for this. I guess you could look at the rstd.mgh map in the output mri_glmfit. The rstd is the std dev of the residual error. If there are some regions that are much higher than others it could represent a problem, but I don't have much more guidance than that. You can also look at the mask.mgh for any areas that are 0 that should not (eg, the medial wall will be 0, but that should be the case)

Thanks in advance.

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer