ENH 5ttgen fsl: allow use of prior probability map in FAST

[hbraunDSP]

Is your feature request related to a problem? Please describe. My team is working on a tractography problem where the default 5TT mask as output by 5ttgen fsl is inadequate. The problem occurs in subcortical gray matter structures. We have solved the problem as a proof of concept by manually editing the 5tt mask after it is output from 5ttgen fsl but this is prohibitively time-consuming to do on every subject.

Describe the solution you'd like I would like the option to pass a custom "prior probability map" to FSL FAST as described here (see the -a and -P options).

Describe alternatives you've considered I will be replacing the 5ttgen fsl command in my pipeline with a custom script based on it. If it's wanted I can submit this new script as a modification of the existing command.

[Lestropie]

Hi @hbraunDSP,

1. Can you give further details / screenshots of where the current 5ttgen fsl behaviour is not acceptable? That could give some hints as to whether there are alternative solutions.

2. Are you working in a context where the 5ttgen hsvs algorithm is not applicable?

3. Generally, making the requisite modifications to the existing script is advised over "substituting" with something else. There's often little tweaks or gotchas resolved in the base script that can be lost if you try to redo things from scratch. But it depends on the details. If it's only adding use of the -a and -P options, requiring a linear registration to template space and a couple of command-line flags, that's a fairly modular extension.

4. Always happy to have external contributions. I can provide advice on what code to put where if necessary. There's also advice here on how to deal with the surrounding software engineering aspects.

Cheers Rob