Multishell ACT HSVS recon workflow uses SS3T response functions

[araikes]

@mattcieslak

In the process of testing things for all of the other issues over the last 24 hours, I noticed that mrtrix_multishell_msmt_ACT-hsvs uses dhollander for the response function estimation (using the SS3TDwiResponse) interface.

	 [Node] Setting-up "qsirecon_0_0_wf.sub-001_mrtrix_multishell_msmt_hsvs.sub_001_ses_01_dir_AP_run_001_space_T1w_desc_preproc_recon_wf.msmt_csd.estimate_response" in "/output/scratch/qsirecon_0_0_wf/sub-001_mrtrix_multishell_msmt_hsvs/sub_001_ses_01_dir_AP_run_001_space_T1w_desc_preproc_recon_wf/msmt_csd/estimate_response".
241025-10:53:46,553 nipype.workflow INFO:
	 [Node] Executing "estimate_response" <qsirecon.interfaces.mrtrix.SS3TDwi2Response>
241025-10:55:00,857 nipype.workflow INFO:

Given that msmt_5tt is a response function option, is there a specific rationale for using dhollander in the HSVS spec?

[mattcieslak]

Luckily this one isn't a bug! When we added ss3t to the recon workflows, we learned that there was an updated dhollander method implemented in the ss3t fork. That method should be available in the mrtrix that is now included in the qsirecon container, so we should switch to explicitly use the mrtrix3 version unless ss3t is explicitly being requested.

[araikes]

Great. Just observed that and wasn't sure if it was intentional or a by product of other processes.