Add additional step in init_preprocessing and init_recon_models scripts about o3d-code

[francopestilli]

Hi Paolo.
Andrew and I are working on this. We understand the first portion of the issue. But we are trying to understand what we need to practically to address the following: "The init_tracking* and init_dissection* script should refer to these images in such a way also the tractograms will be in the radiological orientation."

Does this simply mean that:
(1) the same (current) script will seemingly work this way? They will be able to use the new files?
(2) we need to changes something also in the ini_tracking_* and init_dissection_* scripts so that the scripts become compatible with the orientation of the new nifti files?

Thanks for clarifying.
Franco

@andnp @Paolopost

[andnp]

Started first part with commit 7185db6

[Paolopost]

The scripts to generate the tracking and dissection from the 'fe' structure need to refer the anatomy. We need to be sure that the anatomy file are the one with the proper orientation, it means the one already included in the o3d tree and not the source file of anatomy recovered from the filesystem.

[francopestilli]

Hi Paolo. Thanks for the clarification.
(1) This simply means that we need to change the init_tracking and init_dissection scripts. Correct?

(2) Do you think I need to recreate the FE structures because of this? The FE structures contain the tractograms and the coordinate frames. So I fear that after this change in orientation the FE structures might become out of synch with the anatomy files.

[francopestilli]

Hi @Paolopost is this last question clear to you?

[Paolopost]

I apologize the late reply.

(1) The init_tracking and init_dissection scripts generates "trk" files. "trk" files are embed info on reference anatomy, the association is established at run time with a given anatomy (not necessarily the original one). In this case the revision of the scripts should only take care that the nifti files given in argument as reference anatomy are those "flipped".

(2) The answer to this point is tricky. I'm not 100% sure that we don't need to recompute the FE structures. I believe that independently from this question it would be helpful to define a policy for the sanity check of the orientation of tractography.

[andnp]

As for (1) the init_tracking and init_dissection scripts do currently grab the reference anatomy from the output of the preprocessing pipeline. So no changes need to be made there.