Longitudinal DTI-TK pipeline & from DTI-TK to TBSS

[Sherry Chen]

Dear Gary and DTI_TK experts,

I have been trying to work out the longitudinal DTI_TK proprecessing pipeline. I just wanted to clarity a few things about the pipeline. 

i) If I used the IIT2 template, rather than the ixi_aging_template, in the bootstrap from the very beginning when create the intra_subj_template, do I still need to map the final template to standard MNI space (after group template creation and before tbss_4_prestats_iit_256)? The command I used for bootstrap is as follows: dti_template_bootstrap IITmean_tensor_256.nii.gz intra_subjs.txt

ii)  Is the version of IITmean template usded for DTI-TK processing the 256 × 256 × 256 version? When I use the final template for TBSS in FSL, should I convert it back to the FSL format?

iii) I have wrapped each time point of each subject to the group mean and generated the FA maps using TVtool, say I have two subjects and each with two time points. After running the steps above, I have got four files: subj1_time1_FA.nii.gz, subj1_time2_FA.nii.gz, subj2_time1_FA.nii.gz, and subj2_time2_FA.nii.gz. I then combined these four files to generate the all_FA. Is this correct? Or should I use the '_diffeo.nii.gz' files for each intra_subj_template, which are generated when create the group template, for the final all_FA?   

iv) I would like to do an ROI analysis afterwards, based on the JHU atlas, do I still need to do additional transformation? 

Sorry for this long and complicated email. Please let me know if anything is unclear.

Thanks,

Sherry