This is not entirely surprising. You have a b-value that’s quite high (b=3000) and eddy_correct will not work well with such high b-values, because is tries to register the b=3000 images to a b=0 image, and the b=3000 images do not have enough anatomical detail to be registered correctly. You can “set doeddy = 0" in your config file and first use another preprocessing tool that’s better suited for high-b images like eddy, then run TRACULA on the corrected images. The command that extracts motion measures (dmri_motion) will work with the output of either eddy_correct or eddy (run dmri_motion without arguments for more info).

From: <freesurfer-bounces@nmr.mgh.harvard.edu> on behalf of "Daumail, Loic Jean" <LDAUMAIL@mgh.harvard.edu>
Reply-To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Date: Thursday, July 26, 2018 at 1:58 PM
To: "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu>
Subject: Re: [Freesurfer] questions about Tracula and doeddy=1 or 0

Dear FreeSurfer experts,

I had sent several times this email since the last few months, and never obtained any reply. Would anyone be able to provide me little bit of insight please?

I am analyzing a dataset (Q-Ball) of subjects with TRACULA on FreeSurfer 6.0. b1=0, b2=3000 s/mm2 and diff directions = 90.

Surprisingly the trac-all  reconstruction with doeddy =1 was not as good as with doeddy = 0 on average on all the subjects, as in our dataset there were more tracts missing with doeddy=1 among all the subjects, which is a very curious result, shouldn't we get the reverse?
      ==> should I rely more on the doeddy=0 result then?
      ==> or should I modify the configuration parameters, such as adding more control points , ... and maybe others? 

Secondly, we are using the head motion parameters (computed with doeddy=1) to exclude subjects who moved too much. Is that appropriate? In other words, can we trust the motion parameters and exclude subjects on that basis?

Thank you for your time and consideration,

Sincerely,


Loïc Daumail
Graduate Intern at Athinoula A. Martinos Center for Biomedical Imaging MGH/HST
From: Daumail, Loic Jean
Sent: Thursday, July 12, 2018 4:50 PM
To: freesurfer@nmr.mgh.harvard.edu
Cc: gdesbord@nmr.mgh.harvard.edu
Subject: questions about Tracula and doeddy=1 or 0

Dear FreeSurfer experts,

I am analyzing a dataset (Q-Ball) of subjects with TRACULA on FreeSurfer 6.0. b1=0, b2=3000 s/mm2 and diff directions = 90.
Surprisingly the trac-all  reconstruction with doeddy =1 was not as good as with doeddy = 0 on average on all the subjects, as in our dataset there were more tracts missing with doeddy=1 among all the subjects, which is a very curious result, shouldn't we get the reverse?
      ==> should I rely more on the doeddy=0 result then?
      ==> or should I modify the configuration parameters, such as adding more control points , ... and maybe others? 

Secondly, we are using the head motion parameters (computed with doeddy=1) to exclude subjects who moved too much. Is that appropriate? In other words, can we trust the motion parameters and exclude subjects on that basis?

Thank you for your time and consideration,

Sincerely,


Loic