Lars Kuchinke | 1 Apr 13:45
Picon
Picon
Favicon

combining events in marsbars stats

Hi,

i just wonder what happens if one defines new events that contain two or 
more events
(regressors?) from one session for the % signal change analysis
does this mean that a time course and a mean %signal change will be 
computed where
all regressors contribute to it or what is the difference to defining 
single events...

or in other words is there a possibility to compute % signal change and 
time courses
for predefined contrasts...?

Thanks.
Lars.

Dr. Lars Kuchinke
Freie Universität Berlin
Allgemeine Psychologie
Habelschwerdter Allee 45
14195 Berlin

kuchinke <at> zedat.fu-berlin.de
www.fu-berlin.de/allgpsy
Tel:  +49 (0)30 838 55776

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
(Continue reading)

Tom Jenkins | 11 Apr 15:22
Picon
Picon

Fetching data error message

Dear MarsBar experts,

I would be very grateful for your advice on 2 issues.

1. This error message (attached). I am running MarsBar with SPM5, latest 
updates, Matlab 7.3.
I built a region of interest at -43 -70 -13, sphere, radius 6.2mm and 
everything looks ok when I view it.
However when I try to extract the data, it gives the following error.
I have tried rebuilding the region, and re-estimating the SPM.mat the 
data is extracted from but the problem persists.

2. I was able to specify a different region of interest at -22 -26 -4 
and extract y values from the matlab window for mean signal change. 
However when I then changed to a second region and specified a different 
ROI, the extracted numbers were exactly the same. I have double and 
triple checked that I specified everything correctly. When I tried 
closing SPM and reopening it in between specifying ROIs, it did give 
different numbers. Has anyone else come across this and do you think I 
can trust the extracted data?

Many thanks for your help,

Tom

~ - SPM5: marsbar toolbox  -----------------------------------------

        Toolbox directory prepended to Matlab path:
(Continue reading)

Matthew Brett | 11 Apr 15:51
Picon
Gravatar

Re: Fetching data error message

Hi,

>  1. This error message (attached). I am running MarsBar with SPM5, latest
> updates, Matlab 7.3.
>  I built a region of interest at -43 -70 -13, sphere, radius 6.2mm and
> everything looks ok when I view it.
>  However when I try to extract the data, it gives the following error.
>  I have tried rebuilding the region, and re-estimating the SPM.mat the data
> is extracted from but the problem persists.

It looks like the sphere does not overlap your functional data - try
changing the 'default structural' in the view options to one of your
functional images and check if they overlap by viewing the ROI on the
functional image.

>  2. I was able to specify a different region of interest at -22 -26 -4 and
> extract y values from the matlab window for mean signal change. However when
> I then changed to a second region and specified a different ROI, the
> extracted numbers were exactly the same. I have double and triple checked
> that I specified everything correctly. When I tried closing SPM and
> reopening it in between specifying ROIs, it did give different numbers. Has
> anyone else come across this and do you think I can trust the extracted
> data?

Hmm.  How did you change to the second region?  Did you extract the
data and then plot it?  Did the data look the same for the two regions
when you did that?

Best,

(Continue reading)

Jiti K | 16 Apr 06:52
Picon

ROI from SnPM

Hi,
I want to extract ROI from an SnPM analysis.
I found in the archive a similar request to which Brett had suggested the following function.
mars_cluster_img2rois.m
However, I failed to download the attachment.
Can I request for that function?
Thanks,
Jiti

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Marsbar-users mailing list
Marsbar-users <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/marsbar-users
Matthew Brett | 16 Apr 09:50
Picon
Gravatar

Re: ROI from SnPM

Hi,

On Wed, Apr 16, 2008 at 4:52 AM, Jiti K <jiti2005 <at> gmail.com> wrote:
> Hi,
> I want to extract ROI from an SnPM analysis.
> I found in the archive a similar request to which Brett had suggested the
> following function.
> mars_cluster_img2rois.m
> However, I failed to download the attachment.

Isn't it in your marsbar distribution?

Best,

Matthew

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
Jiti K | 17 Apr 04:37
Picon

Re: ROI from SnPM



On 4/16/08, Matthew Brett <matthew.brett <at> gmail.com> wrote:
Hi,


On Wed, Apr 16, 2008 at 4:52 AM, Jiti K <jiti2005 <at> gmail.com> wrote:
> Hi,
> I want to extract ROI from an SnPM analysis.
> I found in the archive a similar request to which Brett had suggested the
> following function.
> mars_cluster_img2rois.m
> However, I failed to download the attachment.


Isn't it in your marsbar distribution?

       Thanks Brett for your reply but I can't find it with in marsbar. Jiti


Matthew

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Marsbar-users mailing list
Marsbar-users <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/marsbar-users
Conor Wild | 17 Apr 17:17
Picon

A Simple ROI Question

Hi all,

I'm a new MarsBar user, and have a quick question for you:

I wish to perform an ROI analysis of my functional data using ROIs
defined by painting anatomical features of the subjects' structural
images (these are saved as img files).  Of course, the structurals
(1x1x1) are in a different space than the functionals (3x3x4) - should
I resize my images and generated ROI images to be the same size?  Or
can I just create Marsbar ROIs from my 1x1x1 images and directly use
those in my analyses?

Thank you for any input!

Sincerely,

Conor Wild
Centre for Neuroscience Studies
Queen's University
Kingston, ON, Canada

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
Matthew Brett | 17 Apr 17:34
Picon
Gravatar

Re: A Simple ROI Question

Hi,

>  I wish to perform an ROI analysis of my functional data using ROIs
>  defined by painting anatomical features of the subjects' structural
>  images (these are saved as img files).  Of course, the structurals
>  (1x1x1) are in a different space than the functionals (3x3x4) - should
>  I resize my images and generated ROI images to be the same size?  Or
>  can I just create Marsbar ROIs from my 1x1x1 images and directly use
>  those in my analyses?

You should be OK with the 1x1x1 ROIs.  I think!

Matthew

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
Matthew Brett | 18 Apr 12:30
Picon
Gravatar

Re: ROI from SnPM

Hi,

> > > mars_cluster_img2rois.m
> > > However, I failed to download the attachment.
> >
> >
> > Isn't it in your marsbar distribution?
> >
> >        Thanks Brett for your reply but I can't find it with in marsbar.

No, sorry, you're right, I think that's been renamed to mar_img2rois.m.

To import clusters, you can - I think - just follow the GUI options
from the ROI definition -> Import menu, choosing 'cluster image', or
you can call the mars_img2rois function with the image and various
other arguments - flag probably just 'c'.

I'm afraid I don't know the SnPM interface very well, or how they
store their designs - it may be difficult to import SnPM designs
directly through marsbar - because marsbar only understands SPM
designs,

Sorry about that.

Matthew

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
jaclyn billington | 24 Apr 16:30
Picon
Picon
Favicon

stats table batch

 

 

Dear Users,

 

I’m trying to create a script to do Multiple ROIs on Multiple Subjects and create stats table for each ROI/ Subject. I’m using SPM5. I keep getting the following error and I’m not sure if I’m somehow over writing my mars estimated design, I hope someone can point me in the right direction. My script is also posted below, in case I’m doing something blindingly obvious wrong…

 

Thank you in advance,

 

Jac

 

 

ERROR

 

 

??? Reference to non-existent field 'betas'.

 

Error in ==> pr_stat_compute at 39

betas = SPM.betas;

 

Error in ==> mardo_2.compute_contrasts at 43

[marsS.con marsS.stat, marsS.P, marsS.Pc] = ...

 

Error in ==> mardo.stat_table at 29

[marsS] = compute_contrasts(marsD, Ic);

 

Error in ==> v_batch_mars at 116

[rep_strs, marsS, marsD, changef] = stat_table(marsD, 1:length(xCon))

 

 

 

 

SCRIPT

 

 

clear all

 

% Start marsbar to make sure spm_get works

 

marsbar('on')

 

SubsSPMs={'C:\xxx\results\SPM.mat';...

     'C:\yyy\results\SPM.mat’;...

};

 

r1={'C:\xxx\MT+_roi.mat'    

 'C:\xxx\MT_roi.mat';...

};

 

r2={'C:\xxx\MT+_roi.mat'};    

 'C:\xxx\MT_roi.mat';...

};

 

 

for z=1:length(SubsSPMs)

   

 

    spm_name =SubsSPMs{z};

disp (['Subject  ' ,num2str(z), '                ...loading SPM']) 

   

% Make marsbar design object

marsD  = mardo(spm_name);

 

 rnam=strcat('r',num2str(z));

    eval ([ 'rr = ' rnam ';'])

   

    for i=1:length(rr)

        roi_file = rr{i}

 

% Make marsbar ROI object

R  = maroi(roi_file);

 

% Fetch data into marsbar data object

Y  = get_marsy(R, marsD, 'mean');

 

% Get contrasts from original design

xCon = get_contrasts(marsD);

 

% Estimate design on ROI data

E = estimate(marsD, 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));

 

% create stats table. IC - indices for constrast to be displayed  MarsD -

% Marsbar design structure

 

%[rep_strs, marsS, marsD, changef] = stat_table(marsD, Ic)

[rep_strs, marsS, marsD, changef] = stat_table(marsD, 1:length(xCon))

 

    end

   

end

 

 

 

 

 

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Marsbar-users mailing list
Marsbar-users <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/marsbar-users

Gmane