ss3t_csd_beta1 on single shell data

[Davi1990]

I have single shell (one b0 and 30 b1 volumes) data with low b-value (b1=1000). I applied common pre-processing steps (denoising, preproc and bias correction). I estimated the response function of each tissue through dhollander algorithm. Then, I performed dwi2fod with ss3t_csd_beta1 and this is what I got

Template tractography looks great and I am really satisfied about this new dwi2fod algorithm

Cheers

Davide

[thijsdhollander]

I have single shell (one b0 and 30 b1 volumes) data with low b-value (b1=1000).

Seems very common still: lots of the other feedback similarly shows low b-value data, e.g. b=1000. Going by previous feedback and experience though, it should work pretty well! The single b=0 image is probably the more risky aspect, but not per se a problem if that b=0 image has no issues of course.

I applied common pre-processing steps (denoising, preproc and bias correction). I estimated the response function of each tissue through dhollander algorithm.

All sensible; the preprocessing is reasonably "default" by now indeed.

Then, I performed dwi2fod with ss3t_csd_beta1 and this is what I got ..... Template tractography looks great and I am really satisfied about this new dwi2fod algorithm.

That looks like an excellent WM FOD template, which managed to guide the template tractography really nicely. 👌 The fact that some cortical features stand out reveals that it must've worked pretty well; this especially taking into account that it's template tractography, i.e., where the template itself is already an average of all subjects, and which therefore unavoidably shows more blurred features towards the cortex. For these features to still stand out, the WM-GM separation by SS3T-CSD must've generated a pretty good contrast, which then benefited the template construction as well as the subsequent tractography. Beautiful! 😍

Just to avoid confusion for other readers: ss3t_csd_beta1 is a stand-alone command, i.e. it's not actually called via dwi2fod (even though it can in principle be regarded as an algorithm to process DWI data into WM FODs and other tissue types of course). This might change in the future (or not; not per se the goal), but at the moment, this is the most convenient way to deal with usage and feedback by early adopters. So far, it looks like it's pretty robust out of the box though. 😎