That looks correct. The contrasts for the 2nd example will test whether the score is equal to 0 or not (and not an interaction between diagnosis and score). It is not wrong, but I just wanted to make sure you know what you are testing.
doug
On 12/17/2013 01:50 AM, Rujing Zha wrote:
Dear all, I want to design a group t-test analysis and correlation analysis for pial thickness. I have read the PPT of freesurfer.groupanalysis, and I wrote a specific design matrix and contrast for my data. However it is the first time that I design matrix in fsgdf by freesurfer. I didnot confirm whether I wrote is correct. I need someone to help me review it. I have two groups, four groups were classified as sex.(Is this necessary or correct?) Here is my design matrix in fsgdf for 2 group t-test: GroupDescriptorFile 1 Title lh_ttest Class con_male Class con_female Class pat_male Class pat_female Variables Age edu Input subjid1 con_male 19 10 Input subjid2 con_male 20 20 Input subjid3 con_male 20 20 Input subjid4 con_male 19 10 Input subjid5 con_female 20 20 Input subjid6 con_female 20 20 Input subjid7 con_female 19 10 Input subjid8 pat_male 20 20 Input subjid9 pat_male 20 20 Input subjid10 pat_male 19 10 Input subjid11 pat_female 20 20 Input subjid12 pat_female 20 20 DefaultVariable Age In this section, I just want to compare the patient group(class 3 and 4) and control group(class 1 and 2) in thickness controling the age and education by ANCOVA. Does this fsgdf implement ANCOVA? Here is my contrast for this: 0.5 0.5 -0.5 -0.5 0 0 0 0 0 0 0 0 Here is my design matrix in fsgdf for partial regression: GroupDescriptorFile 1 Title lh_regression Class con_male Class con_female Class pat_male Class pat_female Variables Age edu score1 score2 Input subjid1 con_male 19 10 20 30 Input subjid2 con_male 20 20 20 30 Input subjid3 con_male 20 20 20 30 Input subjid4 con_male 19 10 20 30 Input subjid5 con_female 20 20 20 30 Input subjid6 con_female 20 20 20 30 Input subjid7 con_female 19 10 20 30 Input subjid8 pat_male 20 20 20 30 Input subjid9 pat_male 20 20 20 30 Input subjid10 pat_male 19 10 20 30 Input subjid11 pat_female 20 20 20 30 Input subjid12 pat_female 20 20 20 30 DefaultVariable score1 In this section, I want to implement partial regression analysis(i.e. score1 and score2) by controling the age,edu and sex. Here is my contrast for score1: 0 0 0 0 0 0 0 0 0 0 0 0 0.25 0.25 0.25 0.25 0 0 0 0 Here is my contrast for score2: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.25 0.25 0.25 0.25 Any reply will be highly appreciated. Thanks. All the best. 2013-12-17
/Rujing Zha/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer