Print

Print


Hi Wen Wei,
 
Maybe this is not a perfect solution for your question, I guess you can do like this :
 
1. get a jobfile just select one subject as the source image through spm's  GUI, named "tempjob.mat";
 
2.  write a matlab script to modify the jobfile
 
       cd  /xxx/data    % specify your source image's directory and read their names
       subjects = struct2cell(dir);
       [m.n] = size (subjects);  
 
       load tempjob.mat       %load the jobfile and add in the all source images
       for i = 3:n
            subjects_name = subject{1,i};
            jobs{1,1}.spatial{1,1}.normalise{1,1}.estwrite.subj(1,i+1).source = subjects_name;
            jobs{1,1}.spatial{1,1}.normalise{1,1}.estwrite.subj(1,i+1).resample = jobs{1,1}.spatial{1,1}.normalise{1,1}.estwrite.subj(1,1).resample ;
            jobs{1,1}.spatial{1,1}.normalise{1,1}.estwrite.subj(1,i+1).wtsrc = subjects_name;
        end 
       save alljob.mat   jobs                     %save the modified jobfile
 
 
3. load the modified jobfile in spm's GUI and run it.
 
help this can help you!
 
bin lv
 
 
 
 
2008-06-05

The Key Lab of Complex Systems and Intelligence Science,
Institute of Automation,
Chinese Academy of Sciences. 

发件人: Wen, Wei
发送时间: 2008-06-04  16:30:28
收件人: [log in to unmask]
抄送:
主题: Re: [SPM] SPM5: multiple file selection for "normalization"?
-
Hi Karsten,
If there are n subjects then you need to "define" all these n subjects 
and then run it. I was thinking of putting all the source images under 
one directory and select them all and then run the who process in one 
go. Otherwise, if this n=500, then you need to do this 500 times which 
would be very tedious. Do I miss some tricks here?
Many thanks,
Wei Wen
Karsten Specht wrote:
> Dear Wei Wen,
>
>   
>> Is there a way of selecting multiple files (as source) in normalization 
>> process (using the same template target) ?  Smoothing etc can be done in 
>> that fashion but not normalization it seems.
>>     
>
> you should select only one source per subject, but in SPM2 as well as SPM5, you have the possibility to define a set of subjects, which should be normalized to the same template.
>
> Karsten
>
> -- 
> -------------------------------------------------------------
> Karsten Specht, PhD
>
> Department of Biological and Medical Psychology
> & National Competence Centre for functional MRI
> University of Bergen
> Jonas Lies vei 91
> 5009 Bergen
> Norway
> Tel.: +47-555-86279
> Fax: +47-555-89872
[log in to unmask] 
> http://fmri.uib.no/
>
>