questions
questions > RE: Sorted DWI images
Nov 17, 2017 03:11 PM | Michael Harms
RE: Sorted DWI images
And for full context, here is the query that I had sent to Jesper:
Hi,
This question is best posed to the developer of 'eddy', Jesper Andersson (FSL list is most appropriate, but cc'ed here).
In particular, beyond just correcting for eddy-current induced distortion, 'eddy' corrects for motion between frames, slice-to-volume movement (a relatively new feature), and in the future will include an option to correct for movement-by-susceptibility interaction effects. I'm not sure if any of those 3 latter types of corrections involve any assumptions that the frames are ordered in their temporal order of acquisition. (e.g., if the input to 'eddy' is sorted by b-value, there may be large "movements" between consecutive frames, which 'eddy' may not correct as well as would be the case if the data were left in their temporal order of acquisition).
You can always leave the ordering untouched during DICOM to NIFTI conversion, run 'eddy', and then resort according to b-value at that point.
FWIW: Why does DKE require the input to be sorted according to b-value? Seems like that is an unnecessary restriction built into the tool.
Cheers,
-MH
Hi,
This question is best posed to the developer of 'eddy', Jesper Andersson (FSL list is most appropriate, but cc'ed here).
In particular, beyond just correcting for eddy-current induced distortion, 'eddy' corrects for motion between frames, slice-to-volume movement (a relatively new feature), and in the future will include an option to correct for movement-by-susceptibility interaction effects. I'm not sure if any of those 3 latter types of corrections involve any assumptions that the frames are ordered in their temporal order of acquisition. (e.g., if the input to 'eddy' is sorted by b-value, there may be large "movements" between consecutive frames, which 'eddy' may not correct as well as would be the case if the data were left in their temporal order of acquisition).
You can always leave the ordering untouched during DICOM to NIFTI conversion, run 'eddy', and then resort according to b-value at that point.
FWIW: Why does DKE require the input to be sorted according to b-value? Seems like that is an unnecessary restriction built into the tool.
Cheers,
-MH
Threaded View
| Title | Author | Date |
|---|---|---|
| Meg Spriggs | Nov 16, 2017 | |
| Chris Rorden | Nov 16, 2017 | |
| Meg Spriggs | Nov 16, 2017 | |
| Michael Harms | Nov 16, 2017 | |
| Chris Rorden | Nov 17, 2017 | |
| Meg Spriggs | Nov 17, 2017 | |
| Chris Rorden | Nov 17, 2017 | |
| Meg Spriggs | Nov 17, 2017 | |
| Michael Harms | Nov 17, 2017 | |