Hi
It is not about being able to use dtifit with high b values. The tool will certainly run the estimation and give you maps. The question is whether you should do that knowing that the tensor model is not good in predicting/fitting the signal for high b values.
Unless you know how to interpret estimates and their contrast from higher b values, and if you are after typical DTI estimates you can simply extract the b0s and b1000 entries from the data, bvals and bvecs files and use those as inputs to dtifit.
Cheers
Stam
> On 9 Aug 2016, at 12:51, Jiyang Jiang <[log in to unmask]> wrote:
>
> Hi FSL users,
>
> We want to generate FA and MD maps using HCP’s pre-processed multi-shell data with 3 b values. These pre-processed HCP datasets have bvals and bvecs and the 4D datasets are eddy-corrected already. It seemed that dtifit can handle this multi-shell data with 3 b values and successfully generated FA and MD maps. However, there was some discussion among the users, saying that the model doesn’t handle data with high b values well (3 b-values are: 1000, 2000 and 3000)? Can I confirm this with you please and if dtifit is not proper for this purpose, what should I do?
>
> Thanks,
> Jiyang
|