I have and it's been very helpful. I'm not sure though where a correlation was specified in those examples. If I understand it correctly, I have to set my groups to 0 and my covariate to 1. The columns in my designamtrix look like this:
1male 1female 2male 2female 3male 3female 1male_behavior 1female_behavior 2male_behavior 2female_behavior 3male_behavior 3female_behavior
So if I want to test the correlation between thickness and behavior across all groups the contrast would be this: 0 0 0 0 0 0 1 1 1 1 1 1 or would it have to be 1/3 instead of 1 because I have 3 groups? Or even 1/6 because of my 3 groups with 2 levels each?
Thanks Clara
----- Ursprüngliche Mail ----- Von: "Douglas N Greve" greve@nmr.mgh.harvard.edu An: freesurfer@nmr.mgh.harvard.edu Gesendet: Dienstag, 31. Mai 2016 19:18:40 Betreff: Re: [Freesurfer] correlation with glm
have you seen our worked-out examples?
http://surfer.nmr.mgh.harvard.edu/fswiki/FsgdExamples
On 05/31/2016 03:39 AM, Clara Kühn wrote:
Dear FreeSurfer experts,
I would like to correlate my thickness data with behavioral measures. Can I do that with glmfit? If so, could you explain how to specify the contrast for that? My fsgd file looks like this:
GroupDescriptorFile 1 Class 1male plus blue Class 1female plus red Class 2male circle blue Class 2female circle red Class 3male triangle blue Class 3female triangle red Variables age_sc1 Mottier Input subject1 1female 51.61 10 Input subject2 1female 48.89 15 ...
Thank you! Clara _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer