Re: [Freesurfer] Statistical analyses with QDEC

Hi Cico, you can do some of this in qdec, but it might be easier to use the "command line" stream of mris_preproc, mri_surf2surf (to smooth), mri_glmfit, and mri_glmfit-sim (correct for multiple comparisons). The commandline us just much more flexible. You will set up and FSGD file that will handle your categories and covariates. Note: sex is a category, so you have to set it up as a class in the FSGD file. Search our wiki for FSGD to find a lot of examples.

On 01/29/2016 11:25 AM, Francesco Cardinale wrote:

Dear Freesurfer experts, I have some questions about statistical analyses.

1. How to adjust for sex and age? I use the precompiled Freesurfer 5.3.0 MacOSX version, and its tool QDEC to compare cortical thickness between two groups. I would like to use age and sex as nuisance factors. No problem with the age, but I don’t see the sex in the list of usable variables if I code it as a categorical variable (male / female). On the opposite, If I code it as numbers (male = 1, female = 2) I can use it, but how can I declare that it is a factor and not a continuous numerical variable? I don’t want that “female” affects the fitting of the GLM with a value that means “male” * 2.
2. What about the use of other morphometric properties as nuisance factors? For example, how could I use sulcality or LGI as nuisance factors? Is it possible to use more than one scalar simultaneously?
3. Similarly to question 2, is it possible to use home-made *.mgh scalars as a nuisance factors? How?
4. Do the command-line tools allow such analyses better than QDEC? Related to my questions, which features do they add?
5. In this field of interest, are there any major advantages coming with Freesurfer 6.0?

Thanks a lot and best regards

Cico

---

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