Problem with nuisance covariate regression

Submitted by Dina Dajani on Wed, 10/01/2014 - 18:21

Hello,

When I regress nuisance covariates (linear trends, Friston 24, and WM & CSF with 99% subject specific maps) the resulting image is completely blank with no brain. When I do this same nuisance covaraite regression but instead choose the WM & CSF a priori SPM maps, it works fine and I am able to finish the preprocessing correctly.  But, for my analyses I would prefer to use the subject-specific maps.  What may be going wrong here?  Thank you for any advice.

 

Best

Dina Dajani

Forums:

YAN Chao-Gan

Sat, 10/04/2014 - 02:58

Hi Dina,

Before reading my answers, please vote us at https://www.missionmainstreetgrants.com/business/detail/7733 (after logging into your facebook).

You should check the individual WM & CSF segmentation masks. As I said, this is theoretically better, but not always in practical.
 
Best,
 
Chao-Gan

Dina Dajani

Thu, 10/09/2014 - 15:00

Hello,

 

Which masks are used in the regression for WM & CSF?  There are the (1) warped masks, and two types of segmentation masks: (2) functional space and (3) thresholded maps

 

Thank you,

Dina Dajani

YAN Chao-Gan

Thu, 10/09/2014 - 15:10

How did you specified nuisance regression. Please paste a snapshot here.
Best,

Chao-Gan


Dina Dajani

Fri, 10/10/2014 - 19:59

 

YAN Chao-Gan

Fri, 10/10/2014 - 20:05

If so, then this is used: (3) thresholded maps

Please vote us at http://rfmri.org/vote

Thanks,

Chao-Gan

Dina Dajani

Mon, 10/06/2014 - 14:59

I have already checked my WM and CSF maps and masks and they look fine.