Hi Mark,
>The output of b0calc by default is a fraction of the main field strength (if --b0 is left at 1). So to get it in Tesla just multiply by your scanner's field strength. You can then convert into rad/s using the gyromagnetic ratio.
If I understand you correctly, I should have set --b0=3 (the field strength in our case is 3 according to the DICOM file) when using b0calc? And to scale the output from b0calc by the gyromagnetic ratio you mentioned?
By the way, is the gyromagnetic ratio a computed value or something I can read off the DICOM files? According to "Correction of geometric distortion in fMRI data" by Peter Jezzard, NeuroImage 2012, he quoted a specific value for gyromagnetic ratio of 42.575 × 10 6 Hz/T for protons. Is this the value I should be using?
>That being said, it is very difficult to get a good result from b0calc unless you have been very careful to get an accurate air/bone segmentation. This is hugely challenging from MRI alone as they are both dark on MRI but have different magnetic properties. So the input image to b0calc needs to have a value of 1 for all tissue, including non-brain tissue and bone, but not air. Using the struct_brain from BET will definitely not be useful as it removes all the non-brain tissues but these are absolutely crucial for determining the magnetic field inside the head (which is what b0calc is trying to do).
If I read this correctly, you are saying the input to b0calc should include extracted brain mask, skull and non-brain tissue (as "1" in the input image)? However I am still not clear what other non-brain tissue are you referring to? Scalp?
Unfortunately there are no CT scans for our data set we will just need to carry out some of the labelling manually.
Thanks a lot for your help!
Louis.
|