Hi Mona, when you run fcseed-config you can specify to use -mean or -pca npca. -mean produces the mean waveform. -pca will run a PCA on all voxels and take the first npca components. In principle, using more components will fit more of the noise. But it can also make the design more poorly conditioned. In practice, I have not found it to make much of a difference (or even including or not including those regressors), especially for task-based analysis.

On 1/17/19 4:37 PM, Nasiriavanaki, Zahra wrote:

Dear Freesurfers

Hi

I am using white matter and CSF external nuisance regressors in my mkanalysis-sess.

However, when I make these regressors, I use -mean in fcseed-config command, and so the .dat files only have one column (the mean values, I assume).

I looked up in the freesurfer archive and I saw that It is recommended to use -nuisreg wm.dat 5 and -nuisreg vcsf.dat 5.

My question is that what difference it makes to use mean of these regressors vs using only 5 of the columns.

I appreciate if you could please reply whenever you had time.

Thanks

Mona

Zahra (Mona) Nasiriavanaki

Postdoctoral Research Fellow

Martinos Center for Biomedical Imaging

Massachusetts General Hospital

149 13th Street, 149-2615

Charlestown, MA, USA, 02129
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer