You'll need a regressor for each of the 8 classes you describe below. You can use mri_glmfit to generate this (Xg.dat file) You'll need two more regressors for age, one for each diagnosis. If a subject (ie, row) is a control then the two values will be AGE 0. If the subject of the row is a patient, then the two values will be 0 AGE. You can then set up a Controls-Patients age (ie, interaction between dx and age) contrast like [0 0 0 0 0 0 0 0 1 -1]
On 05/24/2016 02:30 PM, Timothy Hendrickson wrote:
Freesurfer Support,
I'd like to create a design matrix for a group analysis outside of the DODS and DOSS models. I understand that in order to do this the -X flag must be used. However, I have been unable to find examples of how to do this.
I am hoping to reveal a difference in thickness or gyrification amongst a clinical population. The data set contains two factors: diagnosis, and study site and one covariate: age. Diagnosis has two levels: controls, and patients. Study site has four levels, one level for each location the data has been collected from.
What I would ideally like to do is:
Take into account offset differences amongst diagnosis and study site.
Allowing a difference in age slope amongst the diagnosis levels
Modeling the age slope as the same for the study site levels
My FSGD file is designed as follows
Class SITE 1-Control Class SITE 1-PATIENT Class SITE 2-Control Class SITE 2-PATIENT Class SITE 3-Control Class SITE 3-PATIENT Class SITE 4-Control Class SITE 4-PATIENT Variables age_at_scan
study site levels = 1,2,3 and 4 diagnosis levels = PATIENT and Control age_at_scan = covariate age
Any advice would be greatly appreciated.
Respectfully,
Tim
-- Timothy Hendrickson Department of Psychiatry University of Minnesota Mobile: 507-259-3434 tel:507-259-3434 (texts okay)
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer