Hi guys -
As some of you know, the NIH has sponsored a large MR data acquisition project, that is intending to make a lot of high quality diffusion and resting MRI data freely available online.
We are in the final stages of piloting the imaging protocol, and I was hoping to gather some advice about one particular issue.
After a lot of testing, we have decided to go with 3 shells of b-values, that will likely be at something like b=1,2,3k. We therefore have the following decision: do we set all three to the same TE/TR killing the SNR of the inner shell unnecessarily, or do we optimise each shell for SNR. Optimising TE for SNR gives us about a 10% SNR increase on the inner shell, so not insignificant. We also have the option of optimising TR on each shell separately, which will give us some extra volumes.
The question is, if we make these decisions, will it make it less likely that other people will use the data? If we acquire matched b=0 vols for each shell, then we in Oxford don't think it is too hard to process this data, but we don't want to make it difficult for people to run their own code on it, to to give everybody data that they can't use.
Can people see any major disadvantages in optimising each shell for TE or TR or both?
Thanks very much in advance
Tim
|