Re: [Freesurfer] Extreme mean values for significant clusters (using GLM)

[Bronwyn Overs]

Hi all,

I am still a little confused as to how the individual values are 
calculated for the significant clusters in the GLM model. Is the cluster 
mapped backed on the individual brain and the mean thickness for that 
region is printed out for each individual?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to the 
NeuRA Magazine 


On 14/04/2015 9:29 am, Bronwyn Overs wrote:

Hi Doug,

Sorry which file should the mri_segstats command be in? Also, I am 
still unclear on how the individual values in cache.th13.abs.y.ocn.dat 
are calculated. Could you possibly explain this?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to 
the NeuRA Magazine 


On 13/04/2015 11:34 pm, Douglas Greve wrote:
I'm not sure what the problem is, but it occurs to me that computing 
the mean over the cluster is probably not what should be done. 
Instead, the total should be computed. In the header of that file, 
there should be a command line for mri_segstats. Cut and paste that 
into a shell and add --accumulate and see if the numbers get more 
reasonable.


doug

On 4/6/15 9:48 PM, Bronwyn Overs wrote:

Dear Freesurfer Mailing List,

I have just run a GLM analysis using mri_glmfit with area as the 
measure of interest. For each of my significant contrasts, I 
extracted the mean area for each subjects from the corresponding 
cache.th13.abs.y.ocn.dat file. I then plotted these values to 
examine the nature of the interaction. In doing so I identified 
three individuals who had extreme mean area values (up to 11 sd from 
the average of the mean area) for many of the significant clusters. 
For each of these individuals I cannot see any problems with the 
freesurfer .mgz files or the cortical segmentation.


So my questions are:
1. Are extreme mean values for significant clusters in GLM a 
problem? If so, do you have any idea what could be causing these 
values if the scans appear to be well segmented?
2. If the significant clusters are calculated at a group level, how 
are the individual mean values (in cache.th13.abs.y.ocn.dat) generated?

--

Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 





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




___
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.




___
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.


___
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. 

Re: [Freesurfer] Extreme mean values for significant clusters (using GLM)

[Douglas Greve]

No, it gets the information from the stack of data used as the --y input 
to mri_glmfit.

doug

On 4/19/15 7:05 PM, Bronwyn Overs wrote:

Hi all,

I am still a little confused as to how the individual values are 
calculated for the significant clusters in the GLM model. Is the 
cluster mapped backed on the individual brain and the mean thickness 
for that region is printed out for each individual?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to 
the NeuRA Magazine 


On 14/04/2015 9:29 am, Bronwyn Overs wrote:

Hi Doug,

Sorry which file should the mri_segstats command be in? Also, I am 
still unclear on how the individual values in 
cache.th13.abs.y.ocn.dat are calculated. Could you possibly explain 
this?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to 
the NeuRA Magazine 


On 13/04/2015 11:34 pm, Douglas Greve wrote:
I'm not sure what the problem is, but it occurs to me that computing 
the mean over the cluster is probably not what should be done. 
Instead, the total should be computed. In the header of that file, 
there should be a command line for mri_segstats. Cut and paste that 
into a shell and add --accumulate and see if the numbers get more 
reasonable.


doug

On 4/6/15 9:48 PM, Bronwyn Overs wrote:

Dear Freesurfer Mailing List,

I have just run a GLM analysis using mri_glmfit with area as the 
measure of interest. For each of my significant contrasts, I 
extracted the mean area for each subjects from the corresponding 
cache.th13.abs.y.ocn.dat file. I then plotted these values to 
examine the nature of the interaction. In doing so I identified 
three individuals who had extreme mean area values (up to 11 sd 
from the average of the mean area) for many of the significant 
clusters. For each of these individuals I cannot see any problems 
with the freesurfer .mgz files or the cortical segmentation.


So my questions are:
1. Are extreme mean values for significant clusters in GLM a 
problem? If so, do you have any idea what could be causing these 
values if the scans appear to be well segmented?
2. If the significant clusters are calculated at a group level, how 
are the individual mean values (in cache.th13.abs.y.ocn.dat) 
generated?

--

Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 





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




___
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.




___
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.




___
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://m

Re: [Freesurfer] Extreme mean values for significant clusters (using GLM)

[Bronwyn Overs]

Hi Doug,

Sorry I am not sure that I understand. Does the --y file contain all of 
the values for each individuals subject, mapped to the fsaverage image?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to the 
NeuRA Magazine 


On 20/04/2015 9:14 am, Douglas Greve wrote:
No, it gets the information from the stack of data used as the --y 
input to mri_glmfit.

doug

On 4/19/15 7:05 PM, Bronwyn Overs wrote:

Hi all,

I am still a little confused as to how the individual values are 
calculated for the significant clusters in the GLM model. Is the 
cluster mapped backed on the individual brain and the mean thickness 
for that region is printed out for each individual?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to 
the NeuRA Magazine 


On 14/04/2015 9:29 am, Bronwyn Overs wrote:

Hi Doug,

Sorry which file should the mri_segstats command be in? Also, I am 
still unclear on how the individual values in 
cache.th13.abs.y.ocn.dat are calculated. Could you possibly explain 
this?


Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 
Subscribe to 
the NeuRA Magazine 


On 13/04/2015 11:34 pm, Douglas Greve wrote:
I'm not sure what the problem is, but it occurs to me that 
computing the mean over the cluster is probably not what should be 
done. Instead, the total should be computed. In the header of that 
file, there should be a command line for mri_segstats. Cut and 
paste that into a shell and add --accumulate and see if the numbers 
get more reasonable.


doug

On 4/6/15 9:48 PM, Bronwyn Overs wrote:

Dear Freesurfer Mailing List,

I have just run a GLM analysis using mri_glmfit with area as the 
measure of interest. For each of my significant contrasts, I 
extracted the mean area for each subjects from the corresponding 
cache.th13.abs.y.ocn.dat file. I then plotted these values to 
examine the nature of the interaction. In doing so I identified 
three individuals who had extreme mean area values (up to 11 sd 
from the average of the mean area) for many of the significant 
clusters. For each of these individuals I cannot see any problems 
with the freesurfer .mgz files or the cortical segmentation.


So my questions are:
1. Are extreme mean values for significant clusters in GLM a 
problem? If so, do you have any idea what could be causing these 
values if the scans appear to be well segmented?
2. If the significant clusters are calculated at a group level, 
how are the individual mean values (in cache.th13.abs.y.ocn.dat) 
generated?

--

Kind regards,

Bronwyn Overs
Research Assistant

Neuroscience Research Australia

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
*M* 0411 308 769 *T* +61 2 9399 1883 *F* +61 2 9399 1265

neura.edu.au 

Follow @neuraustralia on twitter 
Follow NeuRA on facebook 





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




___
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
dispos

Re: [Freesurfer] adjust pial surface outward for abnormally thick cortex

[Jonathan DuBois]

Hi Bruce,

Thanks. It took some trial and error but I was able to fix the surface with
the reposition surface tool. However, it results in a fairly jagged surface
with some topology error. To correct this, should I use the smooth surface
tool in freeview or reprocess the surfaces with recon-all? Thanks agian

Jonathan
___
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.

Re: [Freesurfer] adjust pial surface outward for abnormally thick cortex

[Bruce Fischl]

I think the topology is constrained to be correct - are you sure there is an 
error? And I guess I would try smoothing in freeview
Cheers
Bruce



> On Apr 19, 2015, at 8:29 PM, Jonathan DuBois  
> wrote:
> 
> Hi Bruce, 
> 
> Thanks. It took some trial and error but I was able to fix the surface with 
> the reposition surface tool. However, it results in a fairly jagged surface 
> with some topology error. To correct this, should I use the smooth surface 
> tool in freeview or reprocess the surfaces with recon-all? Thanks agian
> 
> Jonathan 
> ___
> Freesurfer mailing list
> Freesurfer@nmr.mgh.harvard.edu
> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

___
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.

Re: [Freesurfer] Why weighted least squares?

[ye tian]

Dear Doug,

I don't understand what you mean by " the variance of the fMRI time series".
I fed Freesurfer only structural MRI data. Does that mean that Freesurfer
automatically assumes the data to be longitudinal? I really hope not.

Sincerely,
Ye


On Tue, Apr 14, 2015 at 8:11 PM, Douglas Greve 
wrote:

>  In theory, WLS is the optimal estimator in the case where variance
> differs across subject. The FS implementation actually computes
> "psuedo-WLS" in which the weights are just the inverse of the first level
> variance. A true WLS would use the random effect variance plus the lower
> level variance. The first  level variance is determined from the variance
> of the fMRI time series.
>
> doug
>
>
> On 4/13/15 6:42 PM, ye tian wrote:
>
> Hello Freesurfers,
>
>  I would like to be a little more specific about my question. Was
> heteroskedasticity a reason for Weighted Least Square (WLS)? If so, how did
> Freesurfer determine the weights?
>
>  Thank you!
>
>  Sincerely,
> Ye
>
>
>
>
> On Mon, Apr 13, 2015 at 4:59 PM, ye tian  wrote:
>
>> Dear Freesurfers,
>>
>>  Can someone please explain to me why WLS was chosen as the method of
>> regression?
>>
>>  Thank you very much!
>>
>>  Sincerely,
>> Ye
>>
>
>
>
> ___
> Freesurfer mailing 
> listfreesur...@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
>
>
>
> ___
> 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.
>
>
___
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.