On 3/2/2020 7:30 PM, Hengameh Marzbani wrote:

External Email - Use Caution

Hi Freesurfers, I'm stuck with the centering (de-meaning) issue. I have two discrete variables,gender and patient-control, and 2 continuous variables which are memory and age.I want to look for group differences as well as differences in relation to the behavior,controlling for age and sex and the relevant covariates (i.e. eTIV, mean thickness, surface area ...). I want to do vertex-wise analyses by QDEC. I have this questions: 1- Should I demean all the covariates (age, eTIV, mean thickness,surface area)? How about the variables of interest (e.g Left- hippocampus) and the relevant covariates?

I would demean (and normalize) all of them I think.

2-As I want to control for sex. should I consider 0 for male and 1 for female (as discrete variable)?

Treat sex as a class.

3- For demeaning, should I subtract each value (i.e. each age) by the mean of all participant's ages of both patient and control groups? or I should consider each group separately (subtracting the patient's age by the mean of patient's ages and the control age by the mean of control's ages)?

All subjects together

3- How about the FSGD? Is it the same for FSGD?

Yes

Any suggestion is very welcome, thanks in advance for your answer. Hengameh

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer