Print

Print


They do.

Peace,

Matt.

From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of Sally Grace <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Monday, December 18, 2017 at 5:00 PM
To: <[log in to unmask]>
Subject: Re: [FSL] How to add distortion correction to preprocessing pipeline using command line tools

Hi all, 

Thank you for your reply's. Kristian, I'll give what you suggested a go and see how it turn's out. Looking up the HCP pipelines is a fantastic suggestion! I'll let you know how it goes.

Cheers,
Sally. 

On Mon, Dec 18, 2017 at 9:16 PM, Kristian Loewe <[log in to unmask]> wrote:
Hi Sally,

I'm not an expert myself but I use just one 3D image as input to epi_reg, for example,

  epi_reg --epi=func_meanvol ... --out=epi2struct

Then, you can combine epi2struct_warp with the struct2standard warp (previously computed with flirt/fnirt) using convertwarp.

After that, I also combine this with each volume's transformation from mcflirt so that I can use applywarp to warp each volume of the time series directly to standard space with one transformation.

You could also take a look at the relevant scripts in the HCP pipelines, where this is being done in a similar way, I believe. I don't remember if they use fsl_prepare_fieldmap and epi_reg, though.

Cheers,
Kris


Quoting Sally Grace <[log in to unmask]>:

Hi experts,

I have previously analysed my EPI resting-state data with the following
pipeline using command line tools, but I would like to add distortion
correction and I am unsure at what stage to perform slice timer, motion
correction (mcflirt), and distortion correction.

*Old pipeline:*

   1. STRUCTURAL PREPROCESSING (Input is T1 image)
      1. fslreorient2std
      2. BET brain extraction on T1's
      3. FLIRT (register T1 to standard space and generate and affine)
      4. FNIRT (coregister T1 native to standard space)
   2. FUNCTIONAL PREPROCESSING (input is resting-state EPI)
      1. fslreorient2std
      2. slicetimer
      3. mcflirt
      4. fslmaths -Tmean to get a merged and average .nii using the mean
      across time
   3. functional FLIRT (-in = is fslmaths'd output, -ref = bet'd T1)
   4. applywarp based on native T1 (--in= mcflirt'd functional --premat =
   affine from flirt)


My question is, at what stage in the above pipeline could I perform
distortion correction? Below are my steps to perform distortion correcton
which works well. However, I would like to also perform slice timer and
motion correction so I have attempted to input the mcflirt'd output into
epi_reg but that does not work.

   1. *fsl_prepare_fieldmap*
   2. *epi_reg: *epi_reg --epi=ep2d_3_3iso_Resting.nii --t1=t1.nii.gz
   --t1brain=bet_brain.nii.gz --fmap=fieldmap.nii.gz
   --fmapmag=fieldmap_mag.nii --fmapmagbrain=fieldmap_mag_BET.nii.gz
   --echospacing=0.00035 --pedir=-y --out=epi2struc
   3. *structural flirt:* flirt -in myStructuralImage -ref
   myStandardSpaceImage -omat struct2standard.mat
   4. *functional flirt:*  flirt -in epi2struc.nii.gz -ref
   myStandardSpaceImage -out epi2standard -applyxfm -init struct2standard.mat

Any help is greatly appreciated. Thank you!

Kind regards,
Sally.



--

Sally Grace
BSc (Hons), PhD Candidate.
Centre for Mental Health
Faculty of Health, Arts and Design | Swinburne University of Technology
Hawthorn, 3122

(03) 9214 8415

[log in to unmask] (preferred) | [log in to unmask]