Thanks for all the help and explanation, Anastasia.

Best - Don

From: freesurfer-boun...@nmr.mgh.harvard.edu 
[mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Wednesday, April 11, 2018 9:18 PM
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Subject: Re: [Freesurfer] newbie trac-all -bedp question


It would answer the question "if I were a voxel of the forceps major, how 
likely would I be to be in any particular location?"

________________________________
From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 
<freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>>
 on behalf of Krieger, Donald N. <krieg...@upmc.edu<mailto:krieg...@upmc.edu>>
Sent: Wednesday, April 11, 2018 2:10:50 PM
To: Freesurfer support list
Subject: Re: [Freesurfer] newbie trac-all -bedp question


Ok - I have that sum.

For the forceps major, it's 111,685.

If I normalize by that, i.e. divide each voxel value by that, then I certainly 
get something that looks like a probability distribution since the sum of the 
normalized voxel entries over the entire volume is 1.0 as you had said in an 
earlier post.

But the maximum probability for any single voxel is 255/111685, i.e 0.0028.

Is that to be interpreted as the probability that the forceps major passes 
through a voxel with that value?

That is awfully low ... what am I missing?



From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Wednesday, April 11, 2018 1:55 PM
To: Freesurfer support list 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Hi Don - You would normalize these numbers by their sum, not by their maximum 
(i.e., not by 255).



a.y

________________________________

From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 
<freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>>
 on behalf of Krieger, Donald N. <krieg...@upmc.edu<mailto:krieg...@upmc.edu>>
Sent: Wednesday, April 11, 2018 1:51:43 PM
To: Freesurfer support list
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Hi, Anastasia.



Yes, that makes it simpler.



I can see that each of the path.pd.nii.gz has a voxel entry for the same number 
of voxels. I'm looking at elmo.2008 for which that number is 1,294,336 which I 
presume is the volume of this person's brain as delivered by freesurfer.



Most of those voxels are set to 0.

I presume that means that the probability that the tract passes through that 
voxel is very near 0.0 .

For the forceps major, 7342 voxels have non-zero entries ranging from 1 to 255. 
Most of them are pretty low numbers so the average value for the 7342 non-zero 
voxels is about 16.

For a voxel that has an entry of 1, I presume the probability that the tract 
passes  through that voxel is near zero. For a voxel that has an entry of 127,  
is the probability that the tract passes through it 0.5 ?

And if the entry is 255, is the probability 1.0?



It's that mapping of voxel value to probability that I'm trying to get at. Or 
am I misunderstanding what the voxel value means?



Best - Don



From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Wednesday, April 11, 2018 12:46 PM
To: Freesurfer support list 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



The probability maps are independent for each tract. Each one just tells you 
how likely this one tract is to go through any given voxel, but it tells you 
nothing about other tracts. Does this make sense?



On Apr 11, 2018 12:00 PM, "Krieger, Donald N." 
<krieg...@upmc.edu<mailto:krieg...@upmc.edu>> wrote:

Thanks for getting back, Anastasia.



I'm not suggesting that it should be done differently than it is.



I'm just trying to understand how the voxel values map to probability. That's 
why I described the wrong headed way I was thinking about it ... so you could 
tell me where I'm going wrong.







From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Wednesday, April 11, 2018 11:54 AM
To: Freesurfer support list 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



This wouldn't work because these 18 tracts do not include all tracts in the 
brain.



On Apr 11, 2018 11:19 AM, "Krieger, Donald N." 
<krieg...@upmc.edu<mailto:krieg...@upmc.edu>> wrote:

I was thinking that for a single voxel, the total of the values assigned to the 
voxel over all the tracts would have a maximum of 255 and that 255 would 
correspond to a probability of 1.0.

I can see is likely wrong but for sure it's too simple since each voxel could 
contain crossing fibers from several tracts; hence the total probability 
assigned to the voxel over all the tracts might be greater than 1.0.

??



Surely there is some normalization applied to the values assigned to the voxels 
so they fall between 0 and 255.

Does that normalization vary from one subject to the next and perhaps even from 
one tract to the next?

And how is the normalization mapped to probability?



Thanks - Don



From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Wednesday, April 11, 2018 10:48 AM
To: Freesurfer support list 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



The volume would have to be normalized to a total sum of 1 for the voxel values 
to represent an actual distribution.

________________________________

From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 
<freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>>
 on behalf of Krieger, Donald N. <krieg...@upmc.edu<mailto:krieg...@upmc.edu>>
Sent: Wednesday, April 11, 2018 10:43:11 AM
To: Freesurfer support list
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Great - thanks - that's what I need.

I can use the probability for each pixel as a weight for the neuroelectric 
current count for that pixel to quantify the activity for the tract.



I assume that each voxel value maps directly to its probability, i.e. 0-255 --> 
0.0 - 1.0



Thanks again - best - Don



From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Wednesday, April 11, 2018 10:26 AM
To: Freesurfer support list 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Hi Don - No, dlabel/ contains files that are used in the pathway 
reconstruction, but the output of the reconstruction is saved in dpath/. 
There's a subdirectory in there for each pathway, and it contains a volume 
called path.pd.nii.gz. This is the probability that each voxel belongs to this 
pathway, in the subject's native diffusion space. As TRACULA is a probabilistic 
method, you get a probability associated with each voxel, rather than a set of 
voxel coordinates.



More info (potentially too much info) on TRACULA outputs:

https://surfer.nmr.mgh.harvard.edu/fswiki/trac-all#Outputdirectoriesandfiles



Best,

a.y



________________________________

From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 
<freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>>
 on behalf of Krieger, Donald N. <krieg...@upmc.edu<mailto:krieg...@upmc.edu>>
Sent: Wednesday, April 11, 2018 8:06:01 AM
To: Freesurfer support list
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Thanks for the additional information, Anastasia.

I will look for those files.



What I am after is a list of xyz coordinates for each of the 18 tracts. I see 
masks in the trctrain/trc0xy/dlabel/mni directories which appear to be what I'm 
looking for. They have names like

<roi>_AS_<str1>.<str2>.nii.gz

where str1 = bbr or flt and

str2 = roi1 or roi2

I can extract the lists of coordinates from them with mri_cor2label although I 
haven't checked these lists to make sure they are actually what we need.



I haven't found documentation on these files nor do I see them in the tutorial 
directories for the three elmo's.

I have bedbpostx running but do not expect it to produce them. I suppose it is 
trac-all -path which will do so but I thought I would ask my questions anyway.



  1.  Are these the files that I'm looking for?
  2.  If so, where can I find documentation?
  3.  What is the shortest path to generate them?



It strikes me that maybe it is the streamlines which best define the tracts 
rather than these volumetric masks. I still do not thoroughly understand where 
in the processing sequence the streamlines get assigned to specific tracts. ??



Thanks - Don





From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Tuesday, April 10, 2018 9:23 PM
To: Freesurfer support list 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Hi Don - If the output files described here show up in the end, all is well:



https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FDT/UserGuide#BEDPOSTX



Best,

a.y

________________________________

From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 
<freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>>
 on behalf of Krieger, Donald N. <krieg...@upmc.edu<mailto:krieg...@upmc.edu>>
Sent: Tuesday, April 10, 2018 4:52:57 PM
To: Freesurfer support list
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Great - thanks.

It's running.

I see several warnings which I presume can be ignored:

hostname: Name or service not known



Best - Don





From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 [mailto:freesurfer-boun...@nmr.mgh.harvard.edu] On Behalf Of Yendiki, Anastasia
Sent: Tuesday, April 10, 2018 4:24 PM
To: freesurfer 
<freesurfer@nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>>
Subject: Re: [Freesurfer] newbie trac-all -bedp question



Hi Don - See: 
https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg38004.html



Best,

a.y

________________________________

From: 
freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>
 
<freesurfer-boun...@nmr.mgh.harvard.edu<mailto:freesurfer-boun...@nmr.mgh.harvard.edu>>
 on behalf of Krieger, Donald N. <krieg...@upmc.edu<mailto:krieg...@upmc.edu>>
Sent: Tuesday, April 10, 2018 1:15:05 PM
To: freesurfer
Subject: [Freesurfer] newbie trac-all -bedp question



I am getting the following output with error on this command:



trac-all -bedp -s HDFT1001 -i ep2d_diff_SliceAcc_b1k_64_768x768.23



INFO: SUBJECTS_DIR is 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects

INFO: Diffusion root is 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects

Actual FREESURFER_HOME /media/portable/home/kriegerd/Contrib/Freesurfer_6.0

WARN: Running bedbostx locally - this might take a while

WARN: It is recommended to run this step on a cluster

bedpostx_mgh -n 2 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects/HDFT1001/dmri

/home/kriegerd/Contrib/Freesurfer/bin/bedpostx_mgh: 131: 
/home/kriegerd/Contrib/Freesurfer/bin/bedpostx_mgh: Syntax error: "(" unexpected



I apologize for not finding the answer in the list archive.

It's probably there.



I'm running under ubuntu 14.04.



bedpostx_mgh is a shell script which calls for /bin/sh.

On my machine, that is a symlink to /bin/bash

If I change it to be explicitly /bin/bash, it runs a bit further.

Here is the full output:



INFO: SUBJECTS_DIR is 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects

INFO: Diffusion root is 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects

Actual FREESURFER_HOME /media/portable/home/kriegerd/Contrib/Freesurfer_6.0

WARN: Running bedbostx locally - this might take a while

WARN: It is recommended to run this step on a cluster

bedpostx_mgh -n 2 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects/HDFT1001/dmri

subjectdir is 
/media/portable/home/kriegerd/Contrib/Freesurfer_6.0/subjects/HDFT1001/dmri

Making bedpostx directory structure

Queuing preprocessing stages

/home/kriegerd/Contrib/Freesurfer/bin/fsl_sub_mgh: 470: 
/home/kriegerd/Contrib/Freesurfer/bin/fsl_sub_mgh: Syntax error: "(" unexpected 
(expecting ";;")

Queuing parallel processing stage

0 slices processed

/home/kriegerd/Contrib/Freesurfer/bin/fsl_sub_mgh: 470: 
/home/kriegerd/Contrib/Freesurfer/bin/fsl_sub_mgh: Syntax error: "(" unexpected 
(expecting ";;")

Queuing post processing stage

/home/kriegerd/Contrib/Freesurfer/bin/fsl_sub_mgh: 470: 
/home/kriegerd/Contrib/Freesurfer/bin/fsl_sub_mgh: Syntax error: "(" unexpected 
(expecting ";;")

/home/kriegerd/Contrib/Freesurfer/bin/bedpostx_mgh: line 439: 26461 Terminated  
            ${subjdir}.bedpostX/monitor





Thanks - Don


_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

Reply via email to