Hi Dr. Greve, thanks for getting back to me. I was able to successfully run that as apart of the command however the stderr appears to be too small. The first 6 beta values range from 25000 to 30000 however the stderr is only 0.0003. From the raw results the within group std is about 2000 and with a sample size of ~60 per group I would have expected a stderr in the low hundreds.

On Wed, Apr 22, 2020 at 3:43 PM Douglas N. Greve <dgreve@mgh.harvard.edu> wrote:

When you run fast_fratio, include a 3rd output (so [F pvalues ces] = ...), then the stderr for the contrast is
stderr = sqrt(F./(ces.^2));
Select the contrast to isolate each group

On 4/21/2020 5:54 PM, Graduate Imaging wrote:
External Email - Use Caution

Hi FreeSurfer experts,

I ran a vertex wise analysis that consisted of six groups with two continuous covariates which yielded a cluster after mri_glm-fit. I've extracted the values from the ocn.dat file. I ran a post hoc analyses in matlab to determine which group is significantly different using the command below:

X = load('Group.Xg.dat'); design matrix from freesurfer
C = load('C.dat'); post hoc contrast(s)
Y = load('Whole.brain.cluster.dat'); values from the ocn.dat file
[beta rvar] = fast_glmfit(Y,X)
[F pvalues] = fast_fratio(beta,X,rvar,C)

This outputs an array of beta values where the first six would be the intercept. How would I calculate the standard error of the intercept for each group? Would it be rvar/(sqroot(n)).
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer