Hi every,
       Recently, I'm doing the VBM analysisI using the CAT 12 as a toolbox of SPM, and our subjects was divided 3 groups according to age, the covariate including gender and TIV. But i'm not good in the matrix, so I have a difficult problem in statistical analysis. For example, the CAT 12 manual says in Fig.1 in the attached file, is it meaning that there is not a covariate in two-sample and anova statistical model. I'm so confused, is it correct, but why I can add covariate in these two models in SPM GUI.
      Second, I set anova analysis for determine the effect of age on people and two covariate as the mention above, and estimate and review this model, the results as Fig.2 . If it is ok to analysis continuously, and in the F-test, the contrast is  " 1 0 -1 "( Fig.3 ) for comparing the group 1 and group 3 Fig.4. Significant differences can be found between group 1 and 3.  Likewise, the matrix is " 0 1 -1 "for group 2 and group 3. But when I only compare group 1 and 3 only, the differences is not significant as the Fig.5.
      So,  is it an error to define the contrast for F-test in " 1 0 -1", and is it an error for using covariates in ANOVA analysis ? And why the result is  so different between ANOVA and two-sample t-test?
      Please give me some suggestions and  solution.  Any reply will be appreciated from my deep heart.
Best regards,
xiaojing