DPARSFA Nuisance Covariate Regression Problem

Submitted by lcfoster on Sat, 08/02/2014 - 05:26

 Hello,

 

We are currently attempting to pre-process some resting state data using DPARSFA and have encountered some difficulties with the Nuisance Covariates Regression step. When we originally ran the data all the way through DPARSFA the images came out very blurry and the brain was hard to distinguish from the background. After troubleshooting, we’ve determined that this only happens when we check the NCR option. We’ve tried just checking ‘white matter signal’ and/or ‘CSF signal’ but the problem still doesn’t resolve. Has anyone else run into this? Does anyone do the NCR step after they’re already run through all the other steps? I’ve posted our DPARSFA protocol below.

 

Time Points: 360

TR(s): 2

Remove First: 5

Slice Timing

Slice Number: 31

Slice Order: Sequential [1 – 31]

Reference Slice: 16

Realign

Reorient T1*

T1 Coreg to Fun

Affine Regularisation in Segmentation: European

Normalize

Normalize by DARTEL

Smooth

Default mask

Normalize to Symmetric Template

 

Thanks!

Submitted by jigongjun on Fri, 08/08/2014 - 12:34

 It does not seem there is any mistake in your DPARSFA paramters.

Your question : " ...the images came out very blurry and the brain was hard to distinguish from the background."
I think the blurry is reasonable, because the result images of regression are just the "residual" rather than "residual + mean".   

But, to make the qustion clear, you'd better upload the "Problem Image"