If you want to do a within group analysis, just declare one class. Correlation tests boil down to testing a continuous variable in a glm. I don't have anything which will compute a correlation coefficient, though it is a good idea. As for the # of subjects, that's a power question.

doug

Wang, Xin wrote:

Thank you, Doug. Do you have more examples of different designs? Could anyone answer the rest of my questions?

Thank you again,

Xin

From: Doug Greve [mailto:greve@nmr.mgh.harvard.edu] Sent: Fri 7/20/2007 2:45 PM To: Wang, Xin Cc: freesurfer Subject: Re: [Freesurfer] statistical designs

Check out

http://surfer.nmr.mgh.harvard.edu/docs/ftp/pub/docs/freesurfer.groupanalysis...

for some answers. For paired-t, you can specify --paired-diff to mris_preproc

Wang, Xin wrote:

...

Hello, group I want to learn more about designs of the general linear models for different statistical tests, such as paired T, ANOVA, and correlation tests. Could anybody suggest some examples of the designs for each model? I am facing a few problems of correlation analysis. If I want to do the within group correlation analysis, how can I do it without groups (or classes) as a regressor (if I only have one class and assign a 0 for the group in the contrast, does it work)? How can I specify the pearson's or Spearman's correlation in the analysis? How to know I have enough subjects for this correlation analysis?

Thank you in advance,

XIn Wang

---

---

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

-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422

In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting