If there is no effect in A vs C, then adding it in the Ftest will make it weaker than the t-test.
On 02/06/2017 11:54 AM, Leon Ernst wrote:
Dear Freesurfer experts,
I am comparing cortical thickness data of two patient groups (Group A; Group B) and one control group (Group C) without any covariates.
In a simple t-test I observe a significant difference in cortical thickness only between Group B and Group C in several clusters (corrected p!).
Now I am running an ANOVA ("Three Groups (One Factor/Three Levels), No Covariates") as documented in your FSGD examples.
In the output of the f-test (using the following matrix [1 -1 0; 1 0 -1]) the clusters from my t-test remain but do not survive the correction using "mri_glmfit sim" as below.
Do I have to correct the results of my f-test using mri_glmfit-sim? If yes, why do I reveal a significant group-difference in the corrected t-test and not in the corrected f-test?
mri_glmfit-sim \ --glmdir lh.gender_age.glmdir \ --cache 4 neg \ --cwp 0.05\ --2spaces
Thank you very much in advance!
Sincerely,
Leon
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer