marsbar-users — List for users of MarsBaR

[Marsbar-users] unable to buid roi from image (batch)

[Ilaria S. <is...@ma...>]

Dear All,

I am trying to batch script roi extraction from image (nii) to then perform ROI analysis over multiple ROIs and subjects.
So, I am trying to batch script roi extraction from image (nii) to then perform ROI analysis over multiple ROIs and subjects.

I am using the example lines provided:

fname = roi_list{1};
roi = maroi_image(fname);


I got an error early on while loading the roi:

maroi_image ERROR (line 34) params.vol = spm_vol(params);


-I have seen a couple of posts that resolved a similar issue by adding marsbar('on'). It didn't work for me.
-I tried to versions of matlab: 2016b and 2019a - same error

So I tried a workaround by using

fname = spm_vol(roi_list{i1}); % instead of fname = 'image.nii'
roi = maroi_image(fname);
roi = maroi_matrix(roi);


the script now runs without errors and roi.mat is generated.

However, when I attempt to visualize the rois from marsbar - roi definition - view
They do not appear in the gui in the folder where they are. Thus I understand that marsbar does not recognize them.

Does anyone have a clue?

I noticed that the there is one difference between the roi build via scripting and via gui:
the label field is empty:

val =

[maroi_matrix - 11 - binarized]

label: ''
   source: 'C:\path-to-roi\11_roi.mat'
 binarize: 1
roithresh: 0.5000
 spm_hold: 1


  dat: [158×190×136 double]
  mat: [4×4 double]
maroi: [1×1 maroi]


Can this be the problem?

Please help!!!
Thanks!
Ilaria


-------------------------------------------
Ilaria Sani, PhD
Postdoctoral Fellow, Freiwald Lab
The Rockefeller University
1230 York Ave., New York, NY 10065.
Phone: (212) 327 7699
Fax:     (212) 327 7698
Email: is...@ro...<https://rumail.rockefeller.edu/owa/redir.aspx?C=Kqhy3ohui0OLc6NBHS0gx_jC9yHmUtEIuvMTV8kpl1lEP6utIV_5LPRr35l0-k2eSQyLqrMn4XI.&URL=mailto%3assadagopan%40rockefeller.edu>

Re: [Marsbar-users] (no subject)

[Matthew B. <mat...@gm...>]

Hi,

Sorry - you just reminded me that you got 0.0355 for the contrast -
and that's a little different from the GUI version, but not
crazy-different.  So that might just be the difference between using
SPM con images at the second level compared to using MarsBaR at the
first level.

Cheers,

Matthew

On Mon, Jun 29, 2020 at 4:48 PM Matthew Brett <mat...@gm...> wrote:
>
> Hi,
>
> No - I'm afraid they are not the same - check the entry in the FAQ, in
> case it's helpful.
>
> Are you sure that -0.6894 is for the equivalent contrast to the one
> you loaded on the GUI?
>
> Cheers,
>
> Matthew
>
> On Mon, Jun 29, 2020 at 4:43 PM l.dunne--- via Marsbar-users
> <mar...@li...> wrote:
> >
> > Hello, thanks for your reply.
> >
> >
> >
> > I was under the impression that the code from the FAQ was equivalent to running in the GUI as from the video when I first tried it. You mention that they should be roughly equivalent but actually when I compare the values there is a big difference: the GUI gives -0.04396 while the code from the FAQ gives -0.6894. I’m not sure what I can do about this. Do you have any suggestions? I just want the mean betas from an ROI, and the mean %SC, ideally via code so I can repeat the steps easily. I assume the %SC code from the FAQ is fine but I did not try it from the GUI for a comparison.
> >
> >
> >
> > Thanks
> >
> > L
> >
> >
> >
> > Sent from Mail for Windows 10
> >
> >
> >
> > From: Matthew Brett
> > Sent: 29 June 2020 10:23
> > To: MarsBaR users list
> > Subject: Re: [Marsbar-users] (no subject)
> >
> >
> >
> > Hi,
> >
> > On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
> > >
> > > Hello
> > >
> > >
> > >
> > > I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.
> > >
> > >
> > >
> > > For the GUI I followed this tutorial. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
> > >
> > >
> > >
> > > -0.04396
> > >
> > > -0.06128
> > >
> > > -0.91203
> > >
> > >  ...
> >
> > The video shows the process of extracting the average contrast value,
> > for a single contrast, of two ROIs from an SPM second-level design.
> >
> > > For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:
> > >
> > >
> > >
> > > spm_name = ‘/sub-01/SPM.mat';
> > >
> > > roi_file = fullfile(regions_path, ‘my_ROI.mat’);
> > >
> > >
> > >
> > > % Make marsbar design object
> > >
> > > D  = mardo(spm_name);
> > >
> > > % Make marsbar ROI object
> > >
> > > R  = maroi(roi_file);
> > >
> > > % Fetch data into marsbar data object
> > >
> > > Y  = get_marsy(R, D, 'mean');
> > >
> > > % Get contrasts from original design
> > >
> > > xCon = get_contrasts(D);
> > >
> > > % Estimate design on ROI data
> > >
> > > E = estimate(D, Y);
> > >
> > > % Put contrasts from original design back into design object
> > >
> > > E = set_contrasts(E, xCon);
> > >
> > > % get design betas
> > >
> > > b = betas(E);
> > >
> > > % get stats and stuff for all contrasts into statistics structure
> > >
> > > marsS = compute_contrasts(E, 1:length(xCon));
> > >
> > >
> > >
> > > % The contrast of interest is at mars.rows{2}, so...
> > >
> > > BETAS = marsS.con
> > >
> > >
> > >
> > > The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:
> >
> > Here you have a slightly different analysis, which is where you have
> > taken the average signal across all the voxels of the ROI, for each
> > scan, then run the model on that average.  You'll get a slightly
> > different answer than the analysis in the video, which is using the
> > SPM per-subject contrast images, and averaging the contrast values
> > within the ROI.
> >
> > That said, if you can identify the contrast corresponding to your
> > second level analysis that you used in the GUI run, the value
> > shouldn't be too far off ...
> >
> > Cheers,
> >
> > Matthew
> >
> >
> > _______________________________________________
> > Marsbar-users mailing list
> > Mar...@li...
> > https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fmarsbar-users&amp;data=02%7C01%7Cl.dunne%40surrey.ac.uk%7C95f9e333140942a3df9708d81c0e1474%7C6b902693107440aa9e21d89446a2ebb5%7C0%7C0%7C637290194085567322&amp;sdata=35tfrZVTmrvmiiCdvIc967IGlBOZ2jfSm5BjAf7%2FKiI%3D&amp;reserved=0
> >
> >
> >
> > _______________________________________________
> > Marsbar-users mailing list
> > Mar...@li...
> > https://lists.sourceforge.net/lists/listinfo/marsbar-users

Re: [Marsbar-users] (no subject)

[Matthew B. <mat...@gm...>]

Hi,

No - I'm afraid they are not the same - check the entry in the FAQ, in
case it's helpful.

Are you sure that -0.6894 is for the equivalent contrast to the one
you loaded on the GUI?

Cheers,

Matthew

On Mon, Jun 29, 2020 at 4:43 PM l.dunne--- via Marsbar-users
<mar...@li...> wrote:
>
> Hello, thanks for your reply.
>
>
>
> I was under the impression that the code from the FAQ was equivalent to running in the GUI as from the video when I first tried it. You mention that they should be roughly equivalent but actually when I compare the values there is a big difference: the GUI gives -0.04396 while the code from the FAQ gives -0.6894. I’m not sure what I can do about this. Do you have any suggestions? I just want the mean betas from an ROI, and the mean %SC, ideally via code so I can repeat the steps easily. I assume the %SC code from the FAQ is fine but I did not try it from the GUI for a comparison.
>
>
>
> Thanks
>
> L
>
>
>
> Sent from Mail for Windows 10
>
>
>
> From: Matthew Brett
> Sent: 29 June 2020 10:23
> To: MarsBaR users list
> Subject: Re: [Marsbar-users] (no subject)
>
>
>
> Hi,
>
> On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
> >
> > Hello
> >
> >
> >
> > I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.
> >
> >
> >
> > For the GUI I followed this tutorial. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
> >
> >
> >
> > -0.04396
> >
> > -0.06128
> >
> > -0.91203
> >
> >  ...
>
> The video shows the process of extracting the average contrast value,
> for a single contrast, of two ROIs from an SPM second-level design.
>
> > For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:
> >
> >
> >
> > spm_name = ‘/sub-01/SPM.mat';
> >
> > roi_file = fullfile(regions_path, ‘my_ROI.mat’);
> >
> >
> >
> > % Make marsbar design object
> >
> > D  = mardo(spm_name);
> >
> > % Make marsbar ROI object
> >
> > R  = maroi(roi_file);
> >
> > % Fetch data into marsbar data object
> >
> > Y  = get_marsy(R, D, 'mean');
> >
> > % Get contrasts from original design
> >
> > xCon = get_contrasts(D);
> >
> > % Estimate design on ROI data
> >
> > E = estimate(D, Y);
> >
> > % Put contrasts from original design back into design object
> >
> > E = set_contrasts(E, xCon);
> >
> > % get design betas
> >
> > b = betas(E);
> >
> > % get stats and stuff for all contrasts into statistics structure
> >
> > marsS = compute_contrasts(E, 1:length(xCon));
> >
> >
> >
> > % The contrast of interest is at mars.rows{2}, so...
> >
> > BETAS = marsS.con
> >
> >
> >
> > The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:
>
> Here you have a slightly different analysis, which is where you have
> taken the average signal across all the voxels of the ROI, for each
> scan, then run the model on that average.  You'll get a slightly
> different answer than the analysis in the video, which is using the
> SPM per-subject contrast images, and averaging the contrast values
> within the ROI.
>
> That said, if you can identify the contrast corresponding to your
> second level analysis that you used in the GUI run, the value
> shouldn't be too far off ...
>
> Cheers,
>
> Matthew
>
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fmarsbar-users&amp;data=02%7C01%7Cl.dunne%40surrey.ac.uk%7C95f9e333140942a3df9708d81c0e1474%7C6b902693107440aa9e21d89446a2ebb5%7C0%7C0%7C637290194085567322&amp;sdata=35tfrZVTmrvmiiCdvIc967IGlBOZ2jfSm5BjAf7%2FKiI%3D&amp;reserved=0
>
>
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://lists.sourceforge.net/lists/listinfo/marsbar-users

Re: [Marsbar-users] (no subject)

[<l....@su...>]

Hello, thanks for your reply.

I was under the impression that the code from the FAQ was equivalent to running in the GUI as from the video when I first tried it. You mention that they should be roughly equivalent but actually when I compare the values there is a big difference: the GUI gives -0.04396 while the code from the FAQ gives -0.6894. I’m not sure what I can do about this. Do you have any suggestions? I just want the mean betas from an ROI, and the mean %SC, ideally via code so I can repeat the steps easily. I assume the %SC code from the FAQ is fine but I did not try it from the GUI for a comparison.

Thanks
L

Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986> for Windows 10

From: Matthew Brett<mailto:mat...@gm...>
Sent: 29 June 2020 10:23
To: MarsBaR users list<mailto:mar...@li...>
Subject: Re: [Marsbar-users] (no subject)

Hi,

On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
>
> Hello
>
>
>
> I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.
>
>
>
> For the GUI I followed this tutorial. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
>
>
>
> -0.04396
>
> -0.06128
>
> -0.91203
>
>  ...

The video shows the process of extracting the average contrast value,
for a single contrast, of two ROIs from an SPM second-level design.

> For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:
>
>
>
> spm_name = ‘/sub-01/SPM.mat';
>
> roi_file = fullfile(regions_path, ‘my_ROI.mat’);
>
>
>
> % Make marsbar design object
>
> D  = mardo(spm_name);
>
> % Make marsbar ROI object
>
> R  = maroi(roi_file);
>
> % Fetch data into marsbar data object
>
> Y  = get_marsy(R, D, 'mean');
>
> % Get contrasts from original design
>
> xCon = get_contrasts(D);
>
> % Estimate design on ROI data
>
> E = estimate(D, Y);
>
> % Put contrasts from original design back into design object
>
> E = set_contrasts(E, xCon);
>
> % get design betas
>
> b = betas(E);
>
> % get stats and stuff for all contrasts into statistics structure
>
> marsS = compute_contrasts(E, 1:length(xCon));
>
>
>
> % The contrast of interest is at mars.rows{2}, so...
>
> BETAS = marsS.con
>
>
>
> The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:

Here you have a slightly different analysis, which is where you have
taken the average signal across all the voxels of the ROI, for each
scan, then run the model on that average.  You'll get a slightly
different answer than the analysis in the video, which is using the
SPM per-subject contrast images, and averaging the contrast values
within the ROI.

That said, if you can identify the contrast corresponding to your
second level analysis that you used in the GUI run, the value
shouldn't be too far off ...

Cheers,

Matthew


_______________________________________________
Marsbar-users mailing list
Mar...@li...
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fmarsbar-users&amp;data=02%7C01%7Cl.dunne%40surrey.ac.uk%7C95f9e333140942a3df9708d81c0e1474%7C6b902693107440aa9e21d89446a2ebb5%7C0%7C0%7C637290194085567322&amp;sdata=35tfrZVTmrvmiiCdvIc967IGlBOZ2jfSm5BjAf7%2FKiI%3D&amp;reserved=0

Re: [Marsbar-users] (no subject)

[Sarah H. L. <sar...@gm...>]

This is really helpful too -- thank you!

On Mon, Jun 29, 2020 at 11:11 AM Matthew Brett <mat...@gm...>
wrote:

> Hi,
>
> By the way - check the FAQ for a tiny discussion of the differences of
> doing the ROI analysis at the first level (the batch option you've
> used) and extracting the data at the second level (the GUI procedure
> you've used):
>
>
> http://marsbar.sourceforge.net/faq.html#how-do-i-do-a-random-effect-analysis-in-marsbar
>
> Cheers,
>
> Matthew
>
> On Mon, Jun 29, 2020 at 10:22 AM Matthew Brett <mat...@gm...>
> wrote:
> >
> > Hi,
> >
> > On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
> > >
> > > Hello
> > >
> > >
> > >
> > > I’m trying to extract Beta weights for conditions but I seem to be
> getting different results between the GUI and the script and I don’t know
> why.
> > >
> > >
> > >
> > > For the GUI I followed this tutorial. This produced 25 values, one for
> each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
> > >
> > >
> > >
> > > -0.04396
> > >
> > > -0.06128
> > >
> > > -0.91203
> > >
> > >  ...
> >
> > The video shows the process of extracting the average contrast value,
> > for a single contrast, of two ROIs from an SPM second-level design.
> >
> > > For the batch, I followed the example code from the FAQ. For the first
> subject, which should have beta of -0.04396 as above, this was the code:
> > >
> > >
> > >
> > > spm_name = ‘/sub-01/SPM.mat';
> > >
> > > roi_file = fullfile(regions_path, ‘my_ROI.mat’);
> > >
> > >
> > >
> > > % Make marsbar design object
> > >
> > > D  = mardo(spm_name);
> > >
> > > % Make marsbar ROI object
> > >
> > > R  = maroi(roi_file);
> > >
> > > % Fetch data into marsbar data object
> > >
> > > Y  = get_marsy(R, D, 'mean');
> > >
> > > % Get contrasts from original design
> > >
> > > xCon = get_contrasts(D);
> > >
> > > % Estimate design on ROI data
> > >
> > > E = estimate(D, Y);
> > >
> > > % Put contrasts from original design back into design object
> > >
> > > E = set_contrasts(E, xCon);
> > >
> > > % get design betas
> > >
> > > b = betas(E);
> > >
> > > % get stats and stuff for all contrasts into statistics structure
> > >
> > > marsS = compute_contrasts(E, 1:length(xCon));
> > >
> > >
> > >
> > > % The contrast of interest is at mars.rows{2}, so...
> > >
> > > BETAS = marsS.con
> > >
> > >
> > >
> > > The plan was to loop over all subjects and export their betas for each
> condition, but when I view the BETAS variable I get the following:
> >
> > Here you have a slightly different analysis, which is where you have
> > taken the average signal across all the voxels of the ROI, for each
> > scan, then run the model on that average.  You'll get a slightly
> > different answer than the analysis in the video, which is using the
> > SPM per-subject contrast images, and averaging the contrast values
> > within the ROI.
> >
> > That said, if you can identify the contrast corresponding to your
> > second level analysis that you used in the GUI run, the value
> > shouldn't be too far off ...
> >
> > Cheers,
> >
> > Matthew
>
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://lists.sourceforge.net/lists/listinfo/marsbar-users
>

Re: [Marsbar-users] (no subject)

[Matthew B. <mat...@gm...>]

Hi,

By the way - check the FAQ for a tiny discussion of the differences of
doing the ROI analysis at the first level (the batch option you've
used) and extracting the data at the second level (the GUI procedure
you've used):

http://marsbar.sourceforge.net/faq.html#how-do-i-do-a-random-effect-analysis-in-marsbar

Cheers,

Matthew

On Mon, Jun 29, 2020 at 10:22 AM Matthew Brett <mat...@gm...> wrote:
>
> Hi,
>
> On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
> >
> > Hello
> >
> >
> >
> > I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.
> >
> >
> >
> > For the GUI I followed this tutorial. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
> >
> >
> >
> > -0.04396
> >
> > -0.06128
> >
> > -0.91203
> >
> >  ...
>
> The video shows the process of extracting the average contrast value,
> for a single contrast, of two ROIs from an SPM second-level design.
>
> > For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:
> >
> >
> >
> > spm_name = ‘/sub-01/SPM.mat';
> >
> > roi_file = fullfile(regions_path, ‘my_ROI.mat’);
> >
> >
> >
> > % Make marsbar design object
> >
> > D  = mardo(spm_name);
> >
> > % Make marsbar ROI object
> >
> > R  = maroi(roi_file);
> >
> > % Fetch data into marsbar data object
> >
> > Y  = get_marsy(R, D, 'mean');
> >
> > % Get contrasts from original design
> >
> > xCon = get_contrasts(D);
> >
> > % Estimate design on ROI data
> >
> > E = estimate(D, Y);
> >
> > % Put contrasts from original design back into design object
> >
> > E = set_contrasts(E, xCon);
> >
> > % get design betas
> >
> > b = betas(E);
> >
> > % get stats and stuff for all contrasts into statistics structure
> >
> > marsS = compute_contrasts(E, 1:length(xCon));
> >
> >
> >
> > % The contrast of interest is at mars.rows{2}, so...
> >
> > BETAS = marsS.con
> >
> >
> >
> > The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:
>
> Here you have a slightly different analysis, which is where you have
> taken the average signal across all the voxels of the ROI, for each
> scan, then run the model on that average.  You'll get a slightly
> different answer than the analysis in the video, which is using the
> SPM per-subject contrast images, and averaging the contrast values
> within the ROI.
>
> That said, if you can identify the contrast corresponding to your
> second level analysis that you used in the GUI run, the value
> shouldn't be too far off ...
>
> Cheers,
>
> Matthew

Re: [Marsbar-users] (no subject)

[Sarah H. L. <sar...@gm...>]

Thank you so much!

best
Sarah Hope

On Mon, Jun 29, 2020 at 8:46 AM l.dunne--- via Marsbar-users <
mar...@li...> wrote:

> Just realised I had made a mistake in my previous message  – the code from
> the FAQ gives a beta value of 0.0355 for the contrast (a main effect in the
> form of [1 0 0 0]) I was looking at.
>
>
>
> L
>
>
>
> Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986> for
> Windows 10
>
>
>
> *From: *Matthew Brett <mat...@gm...>
> *Sent: *29 June 2020 10:23
> *To: *MarsBaR users list <mar...@li...>
> *Subject: *Re: [Marsbar-users] (no subject)
>
>
> Hi,
>
> On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
> >
> > Hello
> >
> >
> >
> > I’m trying to extract Beta weights for conditions but I seem to be
> getting different results between the GUI and the script and I don’t know
> why.
> >
> >
> >
> > For the GUI I followed this tutorial. This produced 25 values, one for
> each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
> >
> >
> >
> > -0.04396
> >
> > -0.06128
> >
> > -0.91203
> >
> >  ...
>
> The video shows the process of extracting the average contrast value,
> for a single contrast, of two ROIs from an SPM second-level design.
>
> > For the batch, I followed the example code from the FAQ. For the first
> subject, which should have beta of -0.04396 as above, this was the code:
> >
> >
> >
> > spm_name = ‘/sub-01/SPM.mat';
> >
> > roi_file = fullfile(regions_path, ‘my_ROI.mat’);
> >
> >
> >
> > % Make marsbar design object
> >
> > D  = mardo(spm_name);
> >
> > % Make marsbar ROI object
> >
> > R  = maroi(roi_file);
> >
> > % Fetch data into marsbar data object
> >
> > Y  = get_marsy(R, D, 'mean');
> >
> > % Get contrasts from original design
> >
> > xCon = get_contrasts(D);
> >
> > % Estimate design on ROI data
> >
> > E = estimate(D, Y);
> >
> > % Put contrasts from original design back into design object
> >
> > E = set_contrasts(E, xCon);
> >
> > % get design betas
> >
> > b = betas(E);
> >
> > % get stats and stuff for all contrasts into statistics structure
> >
> > marsS = compute_contrasts(E, 1:length(xCon));
> >
> >
> >
> > % The contrast of interest is at mars.rows{2}, so...
> >
> > BETAS = marsS.con
> >
> >
> >
> > The plan was to loop over all subjects and export their betas for each
> condition, but when I view the BETAS variable I get the following:
>
> Here you have a slightly different analysis, which is where you have
> taken the average signal across all the voxels of the ROI, for each
> scan, then run the model on that average.  You'll get a slightly
> different answer than the analysis in the video, which is using the
> SPM per-subject contrast images, and averaging the contrast values
> within the ROI.
>
> That said, if you can identify the contrast corresponding to your
> second level analysis that you used in the GUI run, the value
> shouldn't be too far off ...
>
> Cheers,
>
> Matthew
>
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
>
> https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fmarsbar-users&amp;data=02%7C01%7Cl.dunne%40surrey.ac.uk%7C95f9e333140942a3df9708d81c0e1474%7C6b902693107440aa9e21d89446a2ebb5%7C0%7C0%7C637290194085567322&amp;sdata=35tfrZVTmrvmiiCdvIc967IGlBOZ2jfSm5BjAf7%2FKiI%3D&amp;reserved=0
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://lists.sourceforge.net/lists/listinfo/marsbar-users
>

Re: [Marsbar-users] (no subject)

[<l....@su...>]

Just realised I had made a mistake in my previous message  – the code from the FAQ gives a beta value of 0.0355 for the contrast (a main effect in the form of [1 0 0 0]) I was looking at.



L



Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986> for Windows 10



From: Matthew Brett<mailto:mat...@gm...>
Sent: 29 June 2020 10:23
To: MarsBaR users list<mailto:mar...@li...>
Subject: Re: [Marsbar-users] (no subject)



Hi,

On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
>
> Hello
>
>
>
> I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.
>
>
>
> For the GUI I followed this tutorial. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
>
>
>
> -0.04396
>
> -0.06128
>
> -0.91203
>
>  ...

The video shows the process of extracting the average contrast value,
for a single contrast, of two ROIs from an SPM second-level design.

> For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:
>
>
>
> spm_name = ‘/sub-01/SPM.mat';
>
> roi_file = fullfile(regions_path, ‘my_ROI.mat’);
>
>
>
> % Make marsbar design object
>
> D  = mardo(spm_name);
>
> % Make marsbar ROI object
>
> R  = maroi(roi_file);
>
> % Fetch data into marsbar data object
>
> Y  = get_marsy(R, D, 'mean');
>
> % Get contrasts from original design
>
> xCon = get_contrasts(D);
>
> % Estimate design on ROI data
>
> E = estimate(D, Y);
>
> % Put contrasts from original design back into design object
>
> E = set_contrasts(E, xCon);
>
> % get design betas
>
> b = betas(E);
>
> % get stats and stuff for all contrasts into statistics structure
>
> marsS = compute_contrasts(E, 1:length(xCon));
>
>
>
> % The contrast of interest is at mars.rows{2}, so...
>
> BETAS = marsS.con
>
>
>
> The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:

Here you have a slightly different analysis, which is where you have
taken the average signal across all the voxels of the ROI, for each
scan, then run the model on that average.  You'll get a slightly
different answer than the analysis in the video, which is using the
SPM per-subject contrast images, and averaging the contrast values
within the ROI.

That said, if you can identify the contrast corresponding to your
second level analysis that you used in the GUI run, the value
shouldn't be too far off ...

Cheers,

Matthew


_______________________________________________
Marsbar-users mailing list
Mar...@li...
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.sourceforge.net%2Flists%2Flistinfo%2Fmarsbar-users&amp;data=02%7C01%7Cl.dunne%40surrey.ac.uk%7C95f9e333140942a3df9708d81c0e1474%7C6b902693107440aa9e21d89446a2ebb5%7C0%7C0%7C637290194085567322&amp;sdata=35tfrZVTmrvmiiCdvIc967IGlBOZ2jfSm5BjAf7%2FKiI%3D&amp;reserved=0

Re: [Marsbar-users] (no subject)

[Matthew B. <mat...@gm...>]

Hi,

On Sun, Jun 28, 2020 at 6:55 PM <lew...@gm...> wrote:
>
> Hello
>
>
>
> I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.
>
>
>
> For the GUI I followed this tutorial. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:
>
>
>
> -0.04396
>
> -0.06128
>
> -0.91203
>
>  ...

The video shows the process of extracting the average contrast value,
for a single contrast, of two ROIs from an SPM second-level design.

> For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:
>
>
>
> spm_name = ‘/sub-01/SPM.mat';
>
> roi_file = fullfile(regions_path, ‘my_ROI.mat’);
>
>
>
> % Make marsbar design object
>
> D  = mardo(spm_name);
>
> % Make marsbar ROI object
>
> R  = maroi(roi_file);
>
> % Fetch data into marsbar data object
>
> Y  = get_marsy(R, D, 'mean');
>
> % Get contrasts from original design
>
> xCon = get_contrasts(D);
>
> % Estimate design on ROI data
>
> E = estimate(D, Y);
>
> % Put contrasts from original design back into design object
>
> E = set_contrasts(E, xCon);
>
> % get design betas
>
> b = betas(E);
>
> % get stats and stuff for all contrasts into statistics structure
>
> marsS = compute_contrasts(E, 1:length(xCon));
>
>
>
> % The contrast of interest is at mars.rows{2}, so...
>
> BETAS = marsS.con
>
>
>
> The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:

Here you have a slightly different analysis, which is where you have
taken the average signal across all the voxels of the ROI, for each
scan, then run the model on that average.  You'll get a slightly
different answer than the analysis in the video, which is using the
SPM per-subject contrast images, and averaging the contrast values
within the ROI.

That said, if you can identify the contrast corresponding to your
second level analysis that you used in the GUI run, the value
shouldn't be too far off ...

Cheers,

Matthew

Re: [Marsbar-users] Extracting ROI first level analysis

[Matthew B. <mat...@gm...>]

Hi,


On Mon, Jun 29, 2020 at 2:27 AM Sarah Hope Lincoln
<sar...@gm...> wrote:
>
> Hi,
>
> I've been asked to extract activity using an ROI on first level analyses (each subject) rather than looking at the group level. When I try to do the ROI extraction in MarsBar I end with 408 time points (the number of scans I have) --- how do I extract an ROI from a first level analysis for a specific contrast?
>
> I'm not even sure this makes sense as a question --- I need to define ROIs at the individual level.

I'm not sure what you mean by extracting for a specific contrast.
What you have now is the data that goes into estimating the
statistical model.   When you estimate the model, you get parameter
estimates - for example, for the amount of raw data change
attributable to a specific event.   Then, you use the contrasts to
select or combine these parameter estimates:

http://matthew-brett.github.io/teaching/glm_intro.html

So you have contrast values for specific contrasts, but not data for
specific contrasts.   The data is the stuff the model uses to make the
estimates that go into the contrasts.   You can pretty easily get the
parameter values and contrast values, using the Statistic Table option
on the Results menu.

Can you give more detail about what you want - in terms of data for a contrast?

Cheers,

Matthew

[Marsbar-users] Extracting ROI first level analysis

[Sarah H. L. <sar...@gm...>]

Hi,

I've been asked to extract activity using an ROI on first level analyses
(each subject) rather than looking at the group level. When I try to do the
ROI extraction in MarsBar I end with 408 time points (the number of scans I
have) --- how do I extract an ROI from a first level analysis for a
specific contrast?

I'm not even sure this makes sense as a question --- I need to define ROIs
at the individual level.

Thanks!
Sarah Hope

[Marsbar-users] (no subject)

[<lew...@gm...>]

<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"><head><meta http-equiv=Content-Type content="text/html; charset=utf-8"><meta name=Generator content="Microsoft Word 15 (filtered medium)"><style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0cm;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri",sans-serif;}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
MsoChpDefault
	{mso-style-type:export-only;}
@page WordSection1
	{size:612.0pt 792.0pt;
	margin:72.0pt 72.0pt 72.0pt 72.0pt;}
div.WordSection1
	{page:WordSection1;}
--></style></head><body lang=EN-GB link=blue vlink="#954F72"><div class=WordSection1><p class=MsoNormal>Hello</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>I’m trying to extract Beta weights for conditions but I seem to be getting different results between the GUI and the script and I don’t know why.</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>For the GUI I followed <a href="https://www.youtube.com/watch?v=ZwQoYD_JPg4">this tutorial</a>. This produced 25 values, one for each of my subjects, stored in `SPM.mars.Y.Y`. Example first 3 values:</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><div><table class=MsoNormalTable border=0 cellspacing=0 cellpadding=0 width=64 style='width:48.0pt;border-collapse:collapse'><tr style='height:15.0pt'><td width=64 nowrap valign=bottom style='width:48.0pt;padding:.75pt .75pt 0cm .75pt;height:15.0pt'><p class=MsoNormal align=right style='text-align:right'><span style='color:black'>-0.04396 <o:p></o:p></span></p></td></tr><tr style='height:15.0pt'><td nowrap valign=bottom style='padding:.75pt .75pt 0cm .75pt;height:15.0pt'><p class=MsoNormal align=right style='text-align:right'><span style='color:black'>-0.06128<o:p></o:p></span></p></td></tr><tr style='height:15.0pt'><td nowrap valign=bottom style='padding:.75pt .75pt 0cm .75pt;height:15.0pt'><p class=MsoNormal align=right style='text-align:right'><span style='color:black'>-0.91203<o:p></o:p></span></p></td></tr></table></div><p class=MsoNormal><span style='color:black'> </span> ...</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>For the batch, I followed the example code from the FAQ. For the first subject, which should have beta of -0.04396 as above, this was the code:</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>spm_name = ‘/sub-01/SPM.mat';</p><p class=MsoNormal>roi_file = fullfile(regions_path, ‘my_ROI.mat’);</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>% Make marsbar design object</p><p class=MsoNormal>D  = mardo(spm_name);</p><p class=MsoNormal>% Make marsbar ROI object</p><p class=MsoNormal>R  = maroi(roi_file);</p><p class=MsoNormal>% Fetch data into marsbar data object</p><p class=MsoNormal>Y  = get_marsy(R, D, 'mean');</p><p class=MsoNormal>% Get contrasts from original design</p><p class=MsoNormal>xCon = get_contrasts(D);</p><p class=MsoNormal>% Estimate design on ROI data</p><p class=MsoNormal>E = estimate(D, Y);</p><p class=MsoNormal>% Put contrasts from original design back into design object</p><p class=MsoNormal>E = set_contrasts(E, xCon);</p><p class=MsoNormal>% get design betas</p><p class=MsoNormal>b = betas(E);</p><p class=MsoNormal>% get stats and stuff for all contrasts into statistics structure</p><p class=MsoNormal>marsS = compute_contrasts(E, 1:length(xCon));</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>% The contrast of interest is at mars.rows{2}, so...</p><p class=MsoNormal>BETAS = marsS.con</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>The plan was to loop over all subjects and export their betas for each condition, but when I view the BETAS variable I get the following:</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>0.491596075488378</p><p class=MsoNormal>0.0354958799257487</p><p class=MsoNormal>0.947696271051007</p><p class=MsoNormal>-0.912200391125258</p><p class=MsoNormal>0.912200391125258</p><p class=MsoNormal>1.68487768525147</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>Of course, none of these values are -0.04396. </p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>What should I do about this? Ideally I want to extract percent signal change (I haven’t checked if the GUI gives different values for this) as well as the beta values.</p><p class=MsoNormal> </p><p class=MsoNormal>Thanks</p><p class=MsoNormal>L</p><p class=MsoNormal><o:p>&nbsp;</o:p></p><p class=MsoNormal>Sent from <a href="https://go.microsoft.com/fwlink/?LinkId=550986">Mail</a> for Windows 10</p><p class=MsoNormal><o:p>&nbsp;</o:p></p></div><div id="DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2"><br />
<table style="border-top: 1px solid #D3D4DE;">
	<tr>
        <td style="width: 55px; padding-top: 13px;"><a href="https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient" target="_blank"><img src="https://ipmcdn.avast.com/images/icons/icon-envelope-tick-round-orange-animated-no-repeat-v1.gif" alt="" width="46" height="29" style="width: 46px; height: 29px;" /></a></td>
		<td style="width: 470px; padding-top: 12px; color: #41424e; font-size: 13px; font-family: Arial, Helvetica, sans-serif; line-height: 18px;">Virus-free. <a href="https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient" target="_blank" style="color: #4453ea;">www.avast.com</a>
		</td>
	</tr>
</table><a href="#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2" width="1" height="1"> </a></div></body></html>

[Marsbar-users] %BOLD signal change in two separates run

[MOUTHON M. <mic...@un...>]

Hi everyone,
I have a theorical question regarding to the extraction of %BOLD signal change in Marsbar.
I have scanned two run of the same task in the MRI (both are included in the FFX model). I know that the %BOLD change is compute regarding the mean in the current scan/session. However, I would like to combine the %BOLD in the two scan/session  for one condition:

-          Is it correct to select the condition in the both session in the Marsbar interface or %BOLD change of each session has to be extract separately ?

-          If it is correct, how Marsbar deal the combination of the two session ? Mean ? Sum ?

-          If it is incorrect, how can I combine the %signal change of each session obtained separately ?

Thank you very much

Michaël Mouthon, PhD

Re: [Marsbar-users] array size limits : a possible workaround?

[Thierry C. <thi...@un...>]

Thanks, that rocks!

The code was simple as possible, I just added the line and it computes.

For information, here it is:

roi{1}=maroi_image(spm_vol(['/hpc/banco/DATA/Convers/ROIs/AllROIs.nii']));
roi{2}=maroi_image(spm_vol(['/hpc/banco/DATA/Convers/ROIs/mask.nii']));

for j=1:25
    j
    D  = mardo(fullfile(Rootfolder, Subject(j).name, 'Anal_trials',
'SPM.mat'));
    D = autocorr(D, 'none');
    for rois=1:2
        rois
        Y  = get_marsy(roi{rois}, D, 'mean');
        E = estimate(D, Y);
        for sess=1:4
            for trial=1:6
                results{rois}.subject{j}(sess,trial)= event_signal(E,
[sess, trial], 60);
            end
        end

    end;
end
save('BRAIN_activity.mat','results')


Thanks a VERY BIG LOT for this one, it made my afternoon....

Take care, Thierry C.
------------------------------
Thierry Chaminade <http://www.int.univ-amu.fr/_CHAMINADE-Thierry_?lang=fr>,
PhD
Institut de Neurosciences de la Timone <http://www.int.univ-amu.fr/>, CNRS
& AMU
Equipe Neural Bases of Communication <http://neuralbasesofcommunication.eu/>
Porteur du projet A*Midex  Physiologie Des Interactions Sociales
<https://www.youtube.com/watch?v=pDyiZBMJILM>

Faculté de Médecine - Bâtiment Neurosciences

27, Bd Jean Moulin

13385 Marseille Cedex 05

33 (0)4 91 32 40 32



Le mer. 1 avr. 2020 à 15:15, Matthew Brett <mat...@gm...> a
écrit :

> Hi,
>
> So - specifically - you want an event-related response from the whole
> brain?
>
> Could you share a script, that's causing the error?
>
> You could try turning off the 'whitening' that's causing your
> immediate error with:
>
> % Make marsbar design object
> D  = mardo(spm_name);
> % Set to no AR modelling
> D = autocorr(D, 'none');
>
> Cheers,
>
> Matthew
>
> On Wed, Apr 1, 2020 at 1:56 PM Thierry Chaminade
> <thi...@un...> wrote:
> >
> > Thanks Matthew for the fast and nice reply!
> >
> > I can see part of your answer, but not enough to implement it. What I
> want is 60-second-condition activity as extracted per MarsBar using SPM
> matrix as input (very simple code!).
> >
> > If I were to calculate something on my own directly from the input
> images, how would I take into account the hemodynamic response, the
> nuisance covariates in SPM, the filtering, etc... I Maybe extracting
> directly from SPM, without MarsBar, would be a solution? If you could
> briefly tell me how you would implement it (thinking as a reviewer: will
> this be considered as a valid  covariate if using a different method?) I
> would try to implement it.
> >
> > At present, I am going to see what I can get with SPM.
> >
> > Thanks a lot for your help,
> > Thierry C.
> >
> >
> > ________________________________
> > Thierry Chaminade, PhD
> > Institut de Neurosciences de la Timone, CNRS & AMU
> > Equipe Neural Bases of Communication
> > Porteur du projet A*Midex  Physiologie Des Interactions Sociales
> >
> > Faculté de Médecine - Bâtiment Neurosciences
> >
> > 27, Bd Jean Moulin
> >
> > 13385 Marseille Cedex 05
> >
> > 33 (0)4 91 32 40 32
> >
> >
> >
> >
> > Le mer. 1 avr. 2020 à 11:35, Matthew Brett <mat...@gm...> a
> écrit :
> >>
> >> Hi,
> >>
> >> On Mon, Mar 30, 2020 at 3:59 PM Thierry Chaminade
> >> <thi...@un...> wrote:
> >> >
> >> > Hello Users,
> >> >
> >> >
> >> > For obscure reasons*, I want to evaluate the response of the whole
> brain using MarsBar. When launching the script that normally works without
> any issue on small(er) ROIs, I have an error that I interpret as having too
> large an ROI
> >> >
> >> > Error using diag
> >> > Requested 333393x333393 (828.1GB) array exceeds maximum array size
> preference.
> >> > Creation of arrays greater than this limit may take a long time and
> cause MATLAB to
> >> > become unresponsive. See array size limit or preference panel for
> more information.
> >> >
> >> > I guess this makes sense, but is there a simple work around to bypass
> the limitation (and I guess also this step:
> >> >
> >> > Error in pr_estimate (line 219)
> >> >     q  = diag(sqrt(trRV./ResSS'),0); % spatial whitening
> >>
> >> Do you just need the data?  In that case, you might be better off with
> >> a lower-level tool, that iterates through the images and takes the
> >> mean (or whatever) of the data.
> >>
> >> Glad to hear MarsBaR is getting some use from its home town :)
> >>
> >> Cheers,
> >>
> >> Matthew
> >>
> >>
> >> _______________________________________________
> >> Marsbar-users mailing list
> >> Mar...@li...
> >> https://lists.sourceforge.net/lists/listinfo/marsbar-users
> >
> > _______________________________________________
> > Marsbar-users mailing list
> > Mar...@li...
> > https://lists.sourceforge.net/lists/listinfo/marsbar-users
>
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://lists.sourceforge.net/lists/listinfo/marsbar-users
>

Re: [Marsbar-users] array size limits : a possible workaround?

[Matthew B. <mat...@gm...>]

Hi,

So - specifically - you want an event-related response from the whole brain?

Could you share a script, that's causing the error?

You could try turning off the 'whitening' that's causing your
immediate error with:

% Make marsbar design object
D  = mardo(spm_name);
% Set to no AR modelling
D = autocorr(D, 'none');

Cheers,

Matthew

On Wed, Apr 1, 2020 at 1:56 PM Thierry Chaminade
<thi...@un...> wrote:
>
> Thanks Matthew for the fast and nice reply!
>
> I can see part of your answer, but not enough to implement it. What I want is 60-second-condition activity as extracted per MarsBar using SPM matrix as input (very simple code!).
>
> If I were to calculate something on my own directly from the input images, how would I take into account the hemodynamic response, the nuisance covariates in SPM, the filtering, etc... I Maybe extracting directly from SPM, without MarsBar, would be a solution? If you could briefly tell me how you would implement it (thinking as a reviewer: will this be considered as a valid  covariate if using a different method?) I would try to implement it.
>
> At present, I am going to see what I can get with SPM.
>
> Thanks a lot for your help,
> Thierry C.
>
>
> ________________________________
> Thierry Chaminade, PhD
> Institut de Neurosciences de la Timone, CNRS & AMU
> Equipe Neural Bases of Communication
> Porteur du projet A*Midex  Physiologie Des Interactions Sociales
>
> Faculté de Médecine - Bâtiment Neurosciences
>
> 27, Bd Jean Moulin
>
> 13385 Marseille Cedex 05
>
> 33 (0)4 91 32 40 32
>
>
>
>
> Le mer. 1 avr. 2020 à 11:35, Matthew Brett <mat...@gm...> a écrit :
>>
>> Hi,
>>
>> On Mon, Mar 30, 2020 at 3:59 PM Thierry Chaminade
>> <thi...@un...> wrote:
>> >
>> > Hello Users,
>> >
>> >
>> > For obscure reasons*, I want to evaluate the response of the whole brain using MarsBar. When launching the script that normally works without any issue on small(er) ROIs, I have an error that I interpret as having too large an ROI
>> >
>> > Error using diag
>> > Requested 333393x333393 (828.1GB) array exceeds maximum array size preference.
>> > Creation of arrays greater than this limit may take a long time and cause MATLAB to
>> > become unresponsive. See array size limit or preference panel for more information.
>> >
>> > I guess this makes sense, but is there a simple work around to bypass the limitation (and I guess also this step:
>> >
>> > Error in pr_estimate (line 219)
>> >     q  = diag(sqrt(trRV./ResSS'),0); % spatial whitening
>>
>> Do you just need the data?  In that case, you might be better off with
>> a lower-level tool, that iterates through the images and takes the
>> mean (or whatever) of the data.
>>
>> Glad to hear MarsBaR is getting some use from its home town :)
>>
>> Cheers,
>>
>> Matthew
>>
>>
>> _______________________________________________
>> Marsbar-users mailing list
>> Mar...@li...
>> https://lists.sourceforge.net/lists/listinfo/marsbar-users
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://lists.sourceforge.net/lists/listinfo/marsbar-users

Re: [Marsbar-users] array size limits : a possible workaround?

[Thierry C. <thi...@un...>]

Thanks Matthew for the fast and nice reply!

I can see part of your answer, but not enough to implement it. What I want
is 60-second-condition activity as extracted per MarsBar using SPM matrix
as input (very simple code!).

If I were to calculate something on my own directly from the input images,
how would I take into account the hemodynamic response, the nuisance
covariates in SPM, the filtering, etc... I Maybe extracting directly from
SPM, without MarsBar, would be a solution? If you could briefly tell me how
you would implement it (thinking as a reviewer: will this be considered as
a valid  covariate if using a different method?) I would try to implement
it.

At present, I am going to see what I can get with SPM.

Thanks a lot for your help,
Thierry C.


------------------------------
Thierry Chaminade <http://www.int.univ-amu.fr/_CHAMINADE-Thierry_?lang=fr>,
PhD
Institut de Neurosciences de la Timone <http://www.int.univ-amu.fr/>, CNRS
& AMU
Equipe Neural Bases of Communication <http://neuralbasesofcommunication.eu/>
Porteur du projet A*Midex  Physiologie Des Interactions Sociales
<https://www.youtube.com/watch?v=pDyiZBMJILM>

Faculté de Médecine - Bâtiment Neurosciences

27, Bd Jean Moulin

13385 Marseille Cedex 05

33 (0)4 91 32 40 32



Le mer. 1 avr. 2020 à 11:35, Matthew Brett <mat...@gm...> a
écrit :

> Hi,
>
> On Mon, Mar 30, 2020 at 3:59 PM Thierry Chaminade
> <thi...@un...> wrote:
> >
> > Hello Users,
> >
> >
> > For obscure reasons*, I want to evaluate the response of the whole brain
> using MarsBar. When launching the script that normally works without any
> issue on small(er) ROIs, I have an error that I interpret as having too
> large an ROI
> >
> > Error using diag
> > Requested 333393x333393 (828.1GB) array exceeds maximum array size
> preference.
> > Creation of arrays greater than this limit may take a long time and
> cause MATLAB to
> > become unresponsive. See array size limit or preference panel for more
> information.
> >
> > I guess this makes sense, but is there a simple work around to bypass
> the limitation (and I guess also this step:
> >
> > Error in pr_estimate (line 219)
> >     q  = diag(sqrt(trRV./ResSS'),0); % spatial whitening
>
> Do you just need the data?  In that case, you might be better off with
> a lower-level tool, that iterates through the images and takes the
> mean (or whatever) of the data.
>
> Glad to hear MarsBaR is getting some use from its home town :)
>
> Cheers,
>
> Matthew
>
>
> _______________________________________________
> Marsbar-users mailing list
> Mar...@li...
> https://lists.sourceforge.net/lists/listinfo/marsbar-users
>

Re: [Marsbar-users] array size limits : a possible workaround?

[Matthew B. <mat...@gm...>]

Hi,

On Mon, Mar 30, 2020 at 3:59 PM Thierry Chaminade
<thi...@un...> wrote:
>
> Hello Users,
>
>
> For obscure reasons*, I want to evaluate the response of the whole brain using MarsBar. When launching the script that normally works without any issue on small(er) ROIs, I have an error that I interpret as having too large an ROI
>
> Error using diag
> Requested 333393x333393 (828.1GB) array exceeds maximum array size preference.
> Creation of arrays greater than this limit may take a long time and cause MATLAB to
> become unresponsive. See array size limit or preference panel for more information.
>
> I guess this makes sense, but is there a simple work around to bypass the limitation (and I guess also this step:
>
> Error in pr_estimate (line 219)
>     q  = diag(sqrt(trRV./ResSS'),0); % spatial whitening

Do you just need the data?  In that case, you might be better off with
a lower-level tool, that iterates through the images and takes the
mean (or whatever) of the data.

Glad to hear MarsBaR is getting some use from its home town :)

Cheers,

Matthew

[Marsbar-users] array size limits : a possible workaround?

[Thierry C. <thi...@un...>]

Hello Users,


For obscure reasons*, I want to evaluate the response of the whole brain
using MarsBar. When launching the script that normally works without any
issue on small(er) ROIs, I have an error that I interpret as having too
large an ROI

Error using diag
Requested 333393x333393 (828.1GB) array exceeds maximum array size
preference.
Creation of arrays greater than this limit may take a long time and cause
MATLAB to
become unresponsive. See array size limit or preference panel for more
information.

I guess this makes sense, but is there a simple work around to bypass the
limitation (and I guess also this step:

Error in pr_estimate (line 219)
    q  = diag(sqrt(trRV./ResSS'),0); % spatial whitening

Thanks for letting me know, I regularly use MarsBar and am very happy of
this tool, working in Marseille actually ;-)

Best regards, Thierry C.

* Reasons are obscure as I would never have done this. But in our
statistical analysis of our ROI data, ROIs covering the whole brain
(brainnetome parcellation) we wanted to have a "global signal" covariate,
and it was a first choice to use the same approach as for the individual
ROIs, hence running MarsBar but on the combined ROIs.

------------------------------
Thierry Chaminade <http://www.int.univ-amu.fr/_CHAMINADE-Thierry_?lang=fr>,
PhD
Institut de Neurosciences de la Timone <http://www.int.univ-amu.fr/>, CNRS
& AMU
Equipe Neural Bases of Communication <http://neuralbasesofcommunication.eu/>
Porteur du projet A*Midex  Physiologie Des Interactions Sociales
<https://www.youtube.com/watch?v=pDyiZBMJILM>

Faculté de Médecine - Bâtiment Neurosciences

27, Bd Jean Moulin

13385 Marseille Cedex 05

33 (0)4 91 32 40 32

Re: [Marsbar-users] missing ROIs when I do Select Data - Extract ROI data (full options)

[Matthew B. <mat...@gm...>]

Hi,


On Mon, Mar 2, 2020 at 3:58 PM Vera Tsogli <Bar...@ui...> wrote:
>
> Hello,
>
> Within SPM I did a 2nd level analysis which produced a result of an
> activation over 9 clusters. After saving every cluster as a mask (witin
> xjview) I converted them to .mat files in marsbar and then I did: Select
> Data - Extract ROI data (full options).
> The problem is that in the brief summary that is produced I see only 7
> regions instead of 9. Is there any reason that marsbar  omits 2 of the
> nine regions?

I don't know what happened there - but - did you save the clusters
within MarsBaR?  I'm afraid I don't know xjview.

I assume that you do end up with 9 _roi.mat files?

Cheers,

Matthew

[Marsbar-users] missing ROIs when I do Select Data - Extract ROI data (full options)

[Vera T. <Bar...@ui...>]

Hello,

Within SPM I did a 2nd level analysis which produced a result of an
activation over 9 clusters. After saving every cluster as a mask (witin
xjview) I converted them to .mat files in marsbar and then I did: Select
Data - Extract ROI data (full options).
The problem is that in the brief summary that is produced I see only 7
regions instead of 9. Is there any reason that marsbar  omits 2 of the
nine regions?

Thank you!
Vera

[Marsbar-users] The error of percent signal change

[Безматерных Д. <dmi...@ya...>]

Hello, my dear community. I have a question about percent signal change. Is there the way to evaluate the error of this estimate? I mean the scatter of this value. I hope that my question is correct.

-- 
With best regards, Dmitrii Bezmaternykh

[Marsbar-users] importing ROI mask and contrasts from the batch

[hatice ş. <hat...@gm...>]

Hello,

I'm quite new to MarsBar and here's my problem:

I have predefined functional ROI's for each subjects (as .nii) that I
created with wfu pickatlas and SPM. I would like to use all the voxels
inside these ROI's to extract beta values via MarsBar.

on GUI: ROI import looks alright (it lets me choose .nii files) but when I
want to import my contrasts MarsBar only looks for _mres.mat files, which I
don't have (I have SPM.mat). (weird enough, it let me choose SPM.mat file
once! I think I'm missing out something here.)

on Batch: Spm contrasts import looks alright:
spm_name = '/my/path/SPM.mat';
but this time I cannot use my ROI's because it asks for a mat file.
roi_file = '/my/path/my_roi.mat';

Do you have any solutions for this?

Preferably I'd like to use batch because I have 5 ROIs and 25 subjects. I
tried turning my .nii ROI's into .mats from MARSBar GUI, but I need all the
peak clusters and I need to combine them again. this works fine on GUI but
again, I don't know how to extract all the peak clusters and combine them
on the batch. I saw a probable solution on the email list:

% Which cluster contains the max?
multiple_maxes = max_coord_mm' * ones(1, size(XYZ, 2));
max_index = find(all(XYZ == multiple_maxes, 1));

The solution above would require me to extract max_coord_mm from all my
ROIs and all my subjects,  but I think there must be a more straightforward
way than separating my ROI mask and combining it again.

Kind regards,
Hatice

Re: [Marsbar-users] %signal change contrasts

[Mellick, W. <me...@mu...>]

Thanks for the quick reply, Matthew! And yes, I see what you mean. Appreciate your clarification on the topic.

Best,
Will

-----Original Message-----
From: Matthew Brett [mailto:mat...@gm...] 
Sent: Monday, January 27, 2020 10:49 AM
To: MarsBaR users list <mar...@li...>
Subject: Re: [Marsbar-users] %signal change contrasts

CAUTION: External

Hi,

On Mon, Jan 27, 2020 at 3:45 PM Mellick, William <me...@mu...> wrote:
>
> Dear Matthew,
>
> I've reviewed your past prior Q&As on %signal change extraction and had a follow-up question.
>
> In prior exchanges with users, you clarified that if you want to investigate the difference between two block-design conditions (i.e., positive pictures > neutral pictures) then you need to extract %signal change for each specific condition (i.e., 1) positive pictures > implicit baseline and 2) negative pictures > implicit baseline) and then subtract the two values outside of marsbar (i.e., in SPSS).
>
> So my specific question is whether or not there is any problem with doing this "double" subtraction (subtracting each condition from baseline, then subsequently subtraction the two)? If you could, please elaborate on your response just so I can understand your logic. Many thanks in advance!

Well - you have no choice but to do this - in order to calculate the signal change, you have to calculate it relative to something.  This is just as true for the betas in the regression model, in MarsBaR or
SPM.   And of course, it's not going to cause a problem algebraically
because

(PP - IB) - (NP - IB) = PP - NP

if you see what I mean ...

Cheers,

Matthew


_______________________________________________
Marsbar-users mailing list
Mar...@li...
https://lists.sourceforge.net/lists/listinfo/marsbar-users

Re: [Marsbar-users] %signal change contrasts

[Matthew B. <mat...@gm...>]

Hi,

On Mon, Jan 27, 2020 at 3:45 PM Mellick, William <me...@mu...> wrote:
>
> Dear Matthew,
>
> I've reviewed your past prior Q&As on %signal change extraction and had a follow-up question.
>
> In prior exchanges with users, you clarified that if you want to investigate the difference between two block-design conditions (i.e., positive pictures > neutral pictures) then you need to extract %signal change for each specific condition (i.e., 1) positive pictures > implicit baseline and 2) negative pictures > implicit baseline) and then subtract the two values outside of marsbar (i.e., in SPSS).
>
> So my specific question is whether or not there is any problem with doing this "double" subtraction (subtracting each condition from baseline, then subsequently subtraction the two)? If you could, please elaborate on your response just so I can understand your logic. Many thanks in advance!

Well - you have no choice but to do this - in order to calculate the
signal change, you have to calculate it relative to something.  This
is just as true for the betas in the regression model, in MarsBaR or
SPM.   And of course, it's not going to cause a problem algebraically
because

(PP - IB) - (NP - IB) = PP - NP

if you see what I mean ...

Cheers,

Matthew

[Marsbar-users] %signal change contrasts

[Mellick, W. <me...@mu...>]

Dear Matthew,

I've reviewed your past prior Q&As on %signal change extraction and had a follow-up question.

In prior exchanges with users, you clarified that if you want to investigate the difference between two block-design conditions (i.e., positive pictures > neutral pictures) then you need to extract %signal change for each specific condition (i.e., 1) positive pictures > implicit baseline and 2) negative pictures > implicit baseline) and then subtract the two values outside of marsbar (i.e., in SPSS).

So my specific question is whether or not there is any problem with doing this "double" subtraction (subtracting each condition from baseline, then subsequently subtraction the two)? If you could, please elaborate on your response just so I can understand your logic. Many thanks in advance!

Best,
Will