So I had a multi-shell protocol which I was supposed to start using for a new study soon. It consisted of 2 shells (b=1000, b=2000), with both shells having 64 directions, and full volumes acquired in both AP and PA directions. Voxel size 2mm isotropic.
Upon coming across fixel-based analysis, and reading the recommendation of a b value of 2500 or higher, I decided to tweak the protocol a little bit. I ended up increasing the second shell’s b value from 2000 to 2500.
After testing the new protocol, I noticed multiple voxels in the FA maps having nonsensically high FA values (>1) dispersed throughout the image. I went back to images which were acquired using the older protocol (second shell b=2000), and such voxels where almost absent (although I did find a few). I also noticed that denoising the images (as recommended in your documentation) reduced the prevalence of such voxels using both protocols.
My question is, should I be worried about the noise produced by the higher b-value (2500)? Is that what’s causing the erroneous voxels? Should I stick with the original 2 shells (b=1000, 2000)? I know the protocol is not optimal, I would ideally want less gradient directions in the b=1000 shell and more in the b=2000/2500 shell, but at this point in time I have to choose between the 2. Which one would you recommend, generally and specifically for using CSD and FBA?
Sorry for the long post, and I hope my question isn’t too amateurish. I’m a physician, so my education didn’t exactly prepare me for making these decisions, and I’m pretty much doing everything on my own. Thanks for the help!
I have a doubt about your question, to calculate the FA maps, are you using both shells or only the higher one? My understanding is than the tensor model is not accurate for multi-shell acquisitions, in most of the articles I read with multi-shell acquisitions the people use the lower b-value shell (+b0) to calculate the FA. I hope this helps.
Thanks for the reply. I used FSL to produce an FA map from both shells.
I’m less concerned about the FA map, I’m more interested in whether or not the higher b value is appropriate within the context of the protocol (given the relatively low number of directions and the significantly lower b-value of the first shell).
I think your sequence is good. If your 64 directions are uniformely sampled over the sphere, the number is good to perform any analisys. I would go for the 2500 bvalue. But it would be great if someone with good knowledge in acquisition protocols could comment on this.
I think @mblesac is right to suggest you should use the b = 1000 s/mm² shell for the FA calculation – that would make it more consistent with the rest of the literature, which is typically single-shell.
It’s also correct that the tensor fit can struggle a bit with multi-shell data, since the signal decay is non-mono-exponential with respect to b (this is after all the entire basis for diffusion kurtosis imaging). How much it struggles depends on the particulars of the acquisition, and crucially on the tensor fitting procedure (as touched on in this post).
These negative values are not uncommon – see this post for an explanation for how these can arise. But these are indeed more likely to occur with higher noise, which matches your experience. Gibbs ringing can also be a major contributor, particularly near the ventricles, due to the much larger CSF signal bringing the surrounding b=0 signal down. You could try using
mrdegibbs in addition to
dwidenoise, see if that helps further.
As to whether you need to worry about them, personally I don’t think so: I’d recommend you compute FA from the b = 1000 s/mm² shell anyway, so the choice of b-value for the higher shell shouldn’t make much of a difference (other than a minor increase in TE). Whether it might cause problems for your analysis is entirely dependent on what you plan on doing with this (I would personally not recommend doing much with FA values in the first place, but that’s just my opinion… ).
As to which b-value is best, this is actually a difficult to question to answer with any authority, as things stand. I would personally encourage you to use the higher b-value of 2,500 s/mm², since that has been shown to provide better contrast in the angular domain (helps with estimating fibre orientations). I would expect it probably also provides better contrast in the b-value domain, since that’s often the value used for e.g. diffusion kurtosis imaging, and is also close to the values we picked in the developing human connectome project. There’s also the arguments presented in the FBA papers, although these relate to the interpretation of the results, rather than sensitivity as such – thankfully the recommendations are broadly in line anyway.
Hope this helps answer your questions…?
Thank you so much for the input, I think I’ll go with the 2500 shell after all
Thank you so much for the detailed reply. I have decided to go with the higher b-value after all
And, yes, I believe you answered my question