On 05/18/2014 12:01 AM, Santosh Yadav wrote:
Dear FreeSurfer Experts,
I am running an analysis in Qedc that include 2 groups (Control and Heart Failure). I want to see the cortical thickness difference between these 2 groups, with age and gender as a co-variate, not interested in any correlation analysis.
In Qdec, Model factors have three options that include the fixed factor, continuous factor (Co-variate) and Nuisance. So groups (control and heart failure) belongs to fixed factor while age and gender belongs to both continuous factor and Nuisance.
When I am running this analysis in Qdec by selecting the model that include only group in the model the contrast vector in the Qdec terminal is 1, -1, and its fine.
But when I am selecting the model that includes age and gender as a nuisance, the contrast vector showing 1, -1, 0, 0, 0, 0.
So my question is why we have two extra 0, 0 in contrast vector, it should be 1 -1 0 0.
Each discrete factor has its own covariate regressor, so the interpretation of the 6 regressors is: Control Offset, HF offset, Control age, HF Age, Control Gender, Control Age. BTW, you should not be using Gender as a continuous variable, it should be a discrete variable. This means that you have 4 groups (Control-Male, Control-Female, etc) and one continuous variable.
Another question: When I finished the analysis in Qdec and got the state data tables, that have 18 files, with volume, area, curv and thickness from both sides. I am interested in thickness and its have four files with left and right. lh.aparc.a2009s and lh.aparc, I am not sure what file should i used for the analysis. I believe for left side qdec is using lh.aparc, Is I am right?
Sorry, what files are you talking about? Were they generated by qdec or recon-all? doug
With Regards,
Santosh K. Yadav
-- With Best Regards, Post Doctoral Fellow, Department of Anesthesiology
University of California, Los Angeles, CA 90034
Phone No. 808-5896940
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer