External Email - Use Caution        

Hello,

I just wanted to follow-up on my last message. I will be thankful if you can respond at the earliest time of your convince.

Thanks and all the best,
Ela

From: Salar Dini, Elaheh <elaheh.salardini@yale.edu>
Sent: Wednesday, January 26, 2022 2:29 PM
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Subject: Re: TRACULA Question
 
Dear Anastasia,

Thank you so much for the response. Interestingly, when I try to open the two reconstructed brains at the same time, they don't overlay. Am I correct to assume then that we have to run recon-all using version 7.2.0? I'm providing a screenshot of the simultaneously opened MRIs in the attached drobox link: MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.dropbox.com/s/c2gsp8h8jj7c93d/Screen%20Shot%202022-01-26%20at%201.40.23%20PM.png?dl=0



If yes, this brings me to another question: if the missing voxels from the BET.MASK are limited to ventricles or gray matter, and we don't correct these missing voxels, is TRACULA still able to run with precision and provide us with an accurate result? Since TRACULA is a probabilistic method and use prior knowledge of positions of anatomical structures with respect to each other, I wasn't sure that white matter reconstruction won't be affected by missing voxels in ventricle or gray matter. Our missing voxels in Alzheimer's Disease Patients are usually limited to ventricles or in extreme cases gray matter. If it is not affecting the white matter reconstruct and extracted metrics, we can skip the step that we manually fix the BET. Mask. If you suggest going ahead and manually correct the missing voxels from the BET mask, we will do it. 

Thank you


____________________________________________________________________________________________________________________________________________

From: Yendiki, Anastasia 
Sent:  Wed, 19 Jan 2022 19:57:32 -0800
Subject: Re: TRACULA Question




Hi Ela - When you overlay the aparc+aseg from 6.0 and the one from 7.2, is 
there a notable difference? Is there any chance that the brain is cut off? The 
error happened in the anterior commissure, that's a tract that does goes fairly 
close to the ventral surface of the brain, and in the part of the log file 
where the error happens I see some strange negative coordinates in the z 
direction.
Anastasia.


From: Salar Dini, Elaheh <elaheh.salardini@yale.edu>
Sent: Tuesday, January 18, 2022 5:05 PM
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Subject: Re: TRACULA Question
 

From: Salar Dini, Elaheh
Sent: Tuesday, January 18, 2022 4:55 PM
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Subject: TRACULA Question
 
Dear Anastasia,
I have run trac -all -prep in 2 different conditions:
a) prerequisite recon-all step was done using freesurfer 7.2.0
b) prerequisite recon-all step was done using freesurfer 6

The first time everything ran smoothly without an error, but when we used aparc+aseg resulted from running recon -all using version 6 of freesurfer, we got the following error for which I also attach the log file to this message: Segmentation fault (core dumped)

My question is that if I'm correct to assume that we can't run trac-all -prep using structural MRI segmentation done by previous versions of freesurfer? The reason I'm asking this is that I have a data from 100 subjects with Alzheimer's Disease. We have been collecting this data over the past 4 years and have done the recon -all step using freesurfer 6 for all of them as we acquired them. If trac-all can't use those segmentation, we have to do this step again for all these 100 subjects. 

I appreciate your help and insight.
Ela