Hello,
I am looking at a multi-shell diffusion imaging sequence for a Siemens scanner - the b values are 300, 1000, and 2000. I am testing out the sequence to see if it would work for tractography. Examination of the bvals file shows that there are b=5 shell (instead of the expected zero) and at the other shells, the b values vary by small amount (for example for the b=300 shell, the b values are between 295 and 305, for b=1000 shell the b values vary between 990 to 1010). Thus instead of having 8 volumes of data acquired at a b-value of 300, there are 8 volumes of data with varying b values averaging out to 300.
For the b value =0 images, the data was instead acquired with b=5.
Questions: (a) what is the advantage of varying the b value relative to having each volume of data acquired with the same b value? (b) Why have b=5 images instead of b=0? What is the advantage?
The b-value includes a double time integral of the square of the magnetic field gradient along a particular orientation. Therefore it necessarily includes more that just the diffusion weighting gradients. Because the refocusing gradients are usually very close to to imaging gradients, most imagining gradients only contribute negligibly to the b-value when diffusion weighting gradients are applied. However the crusher gradients that are placed around the inversion pulse usually have a fairly high amplitude and therefore do have a noticeable impact on the b-value. Your b-value of 5, instead of 0, is primarily due to those crusher gradients. Because the inversion pulse is not a perfect 180 everywhere, there will be both a FID and an echo after the 180 they will interfere with each other causing numerous artifacts if the crusher gradients are not present. The crusher gradients crush the FID caused by the 180.
Siemens scanners at least (I'm not sure about other vendors), calculate the true b-value for each image, including the crushers and other gradients. This is why your b-value of zero is actually 5 and your actual images for the other nominal b-values have scatter around the nominal b-values. most tractography software seem to deal with this type of data without any problems. But if the true b-value of the nominal b=0 images is too high, I have seem some packages complain.
- Badges
- Science topic
- Similar topics
- Condensed Matter Physics
- Zero