[RESOLVED] Hcp_pre_freesurfer fails due to field map with excessive phase range

Issues
21

Hi Layla,

The current problem of incorrect distortion correction of T1w images most likely lies in the parameters, either the hcp_t1samplespacing or hcp_unwarpdir. I have also looked at a few field map images and it seems that a few of them have significantly wrapped phase, which means they need to be unwrapped by including the hcp_echodiff parameter. I assumed that this parameter equaled 3 ms, which is the default value for Philips B0 map sequences, but I keep getting a different error regarding the units. I will look into this issue right now, because I think that phase unwrapping is necessary in your case if you want to perform field map correction.

As for the hcp_pre_freesurfer step, I think you can run this without field map correction because the structural images do not seem heavily distorted to me.

Do you know when approximately you will be able to run hcp_fmri_volume? I can try to resolve the issue that you have just mentioned. I think it would definitely help the quality of the results to correct BOLD images, because I have noticed they are more distorted than T1w.

22

Looking at the these B0 maps, it seems that they were not acquired the same way as other Philips B0 maps I have seen before. The range of values is completely different. These field maps are more similar to Siemens field maps, and the current HCP steps for Philips field map correction assume a different input format. I would suggest that you try to run the entire preprocessing pipeline without field map correction. In the meantime I will try to figure out why these scans do not fit the “standard” (the current default Philips B0 map image) and if you preprocessing result will not be satisfactory, then we can resolve the specific issue with these images.

23

Ok thanks! I’m currently running the whole preprocessing pipeline without field maps, I should QC on tomorrow afternoon, I’ll keep you in touch about the quality of the preprocessing.