Dear Bastian,
the third row looks like a plausible fieldmap, and that is a good start. Reasons for fieldmap correction failing usually has to do with getting the fieldmap wrongly scaled in the first place (for example be entering the wrong echo time difference) or by applying it wrongly (giving the wrong dwell time or unwarp direction).
As for combining it with eddy there is currently no mechanism for directly doing that. What I would recommend in that case is to
1. Create the fieldmap and register it to first b=0 volume
2. Run eddy
3. Apply fieldmap on eddy-corrected data using fugue.
It is a little less ideal since it involves two resamplings, but that isn’t really that big of a deal.
Jesper
On 28 Apr 2016, at 18:01, Bastian David <[log in to unmask]> wrote:
> Dear FSL-Experts,
>
> I inherited an older DTI dataset and I am stuck with the preprocessing at the moment. The dataset (recorded on a Siemens Trio 3T) consists of 60 directions and 6 b0's, the bvecs are oriented in a half-sphere and only one PE-polarity was used in the acquisition (A>>P), so unluckily for me the opposite of everything you recommend to pay attention to during acquisition ;) Therefore, I cannot use topup for susceptibility distortion correction. For most of the subjects I simply run the most recent implementation of eddy - the FA maps after dtifit look alright.
> Nevertheless, during the acquisition somebody had the glorious idea of recording a gradient field map for the last 10 subjects, which I understand to be another approach of tackling susceptibility distortions. My question is: how?
> What I have done so far is preparing the field map using fsl_prepare_fieldmap. For the next steps, I found some old code of Matt Glasser ( https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0904&L=FSL&D=0&P=454499 ). Adapting these few lines led to warping of the DTI data, that doesn't seem correct (see attached screenshots). Furthermore, I attached a screenshot of the flirted field map and magnitude image. Do you have any idea about what might have gone wrong or how I could proceed? I am very grateful for every advice!
>
> Greetings from Germany,
> Bastian
> <before_correction.jpg><after_correction.jpg><flirted_fm.jpg><flirted_magnitude.jpg>
|