Hello,
We've explored this issue further and solved the root of the problem - I'm now hoping for some feedback on this.
To reiterate, I have noticed that an artifactual "slice-gridding" appears to the top (largely) and bottom slices of example_func images in FSL FEAT. This is for functional acquisitions with a relatively coarse resolution in the sagittal plane (6mm). Here is an example:
https://www.dropbox.com/home/FEAT?preview=example_func.png
We've realised this issue can be traced back to the epi_reg command. In attempting to co-register the functional to the high-res space (with information derived from the fieldmap), the resultant func2highres image has of course been warped. Yet, as a result this has caused a sharp-dropoff of signal within the top slices:
https://www.dropbox.com/s/10xygvo6f88brz4/example_func2highres.png?dl=0
Because of these sharp slice dropoffs, applying the inverse warp back to the example_func images hence creates the artifacts shown in the first image.
This is not apparent when not performing the field map unwarping:
https://www.dropbox.com/s/1l2otyfl96ri6r1/example_func_distorted.png?dl=0
---
So, I was wondering two things.
Firstly, is there an alternative option that we may perform in terms of co-registering the functional to T1 images? We have tried running mcflirt and then fugue sequentially, and then inputting the subsequent corrected data into FEAT.
Secondly, what will the effect of the slice-gridding be on the resultant functional images used in later analyses?
Inspection of the filtered functional images does not reveal these artifiacts, which is not surprising given the applywarp functions after mcflirt are different (than to those used after epi_reg in preprocessing stage 1).
We also ran ICA on the filtered functional images "with" and "without" distortion. Both the temporal and spatial characteristics of the ICA components are somewhat similar across the different preprocessing options?
Regards,
Alistair
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|