On 02/04/2013 02:10 AM, Sean Hatton wrote:

Hi FreeSurfers,

These should be a simple questions but they have been bugging me.

1. I have two diagnostic groups that I want to compare controlling for gender and age. I am entering them in Qdec with diagnosis and gender as discrete factors and age as continuous (n=89). If I use DODS I get non-sig result after FDR, but I am not sure if this is modelling back to age 0 – I have demeaned the age to try to get the mean age value but get the same results. I read that to overcome this you can use DOSS where I do get a significant results after FDR, but the results are exactly the same as the age-thickness correlation (I.e. There is no diagnostic effect, it is all age-related). Can you advise the best way to control for age and gender.

I think you are doing the right thing. If you use DOSS, you should first use DODS and look for an interaction with age (ie, a contrast between the age slope for group 1 vs group 2). If that is significant, then you cannot go on to use DOSS. When you include age as a covariate, then the group effect will be the intercept at 0. If you demean, then it will be the group effect at the age mean. Make sure to demean with the mean computed from all subjects (not group specific).

1. Can you control for FDR over both hemispheres?

We don't have anything to do this.

1. After FDR, how can you adjust the heat bars so that you can approapriately compare between statical maps?

I'm not sure what you mean. doug

Thank you!

---

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