On 05/23/2013 03:20 PM, Tudor Popescu wrote:
Hello,
I was wondering whether someone could please explain exactly how values are assigned to each column of the design matrix, when a C.dat contrast vector (e.g. [0, 0, 1, -1]) is created, in the "correlation" as opposed to in the "intercept" types of contrasts ...
Each column of the design matrix means something. Eg, the first column is the intercept for group 1, the 2nd is the intercept for group 2, etc. The exact meaning depends on how you have set up your design. If the column gets a 0, then it is effectively treated as a nuisance variable that is regressed out of the contrast. The +1s and -1s are set up to test your hypothesis. See the presentation on group analysis for more details.
I am asking this because I have a hard time understanding how does a contrast's automatically-created namesake folder (which includes the keywords 'Avg'/'Diff'/'X' and 'Intercept'/'Cor'), such as "lh-Diff-group1-group2-Cor-thickness-behaviouralMeasure", as well as its phrasing in QDEC ("Does the thickness-age correlation differ between group1 and group2?"), exactly correspond to that C.dat vector.
The folder name just needs to be unique. You should look at the C.dat for definitive meaning of the contrast.
Many thanks! Tudor
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer