Question about b-values and its fitness to the software

1

Dear experts, I have a question regarding the b-values of our acquired data, the data we have has the following b-values:
5 5 1489.96 2994.94 1489.99 3009.96 1499.95 2989.96 1494.98 3004.94 1489.99 2999.95 1489.96 2984.93 1494.97 2984.99 1499.99 4.99999 3004.94 1489.97 2989.94 1499.97 2999.98 1499.98 3004.94 1494.98 2984.97 1504.98 2989.95 1499.96 2999.92 1490 3004.99 4.99999 1504.97 2994.96 1494.97 2984.96 1499.96 2989.94 1504.98 2999.95 1499.97 2984.94 1494.97 2994.94 1499.99 2999.94 1485 5 3004.99 1499.96 2989.93 1504.97 2994.96 1494.96 2984.92 1489.95 2994.92 1499.99 2999.99 1499.97 2984.96 1494.97 2984.96 4.99999 1504.98 3004.94 1499.99 2994.97 1505 3009.99 1495 2999.97 1495 2994.97 1494.99 2999.97 1504.99 2994.95 1494.98 4.99997 2984.95 1499.99 2989.96 1489.99 2995 1494.99 2999.97 1489.98 2979.99 1500 3004.95 1489.98 3004.94 1494.97 2990 1499.99 3004.96

And when I fit the data into Mrtrix3, it automatically picks three representative b-value: 5 1489 2989? (I do not quite remember…)

My question is does that affect the tracking accuracy and the result?

Appreciate your help,
Lei

2

Welcome Lei!

There’s a couple of concepts at play here that in retrospect should probably be included in our documentation page on DW scheme handling.

The short answer is that this is absolutely the expected behaviour and there’s nothing to be concerned about. The software is identifying those volumes that have “effectively the same” b-value, and grouping them into b-value “shells”, as many processing steps depend on data that are organised in such a fashion. The variations in b-values reported within those data are likely smaller in magnitude than the inherent noise level of the data, so all they really do is preclude one from saying “give me all volumes with b=3000” in a naive fashion.

Rather than giving a long answer here, I’ll make some additions to the online documentation page, so that they are easier to find for other interested parties. Edit: Preview documentation changes here.

Cheers
Rob

2 Likes