Dear Daniel,
Yes, you can generate a DCM based on simulated data. What DCM gets from the SPM.mat are the stimulus regressors and timing information. This can be fiddly to do yourself – so I would recommend first creating an SPM.mat . You don’t need any fMRI data – click on Batch in the main SPM window, then in the menu at the top, select SPM->Stats->fMRI model specification (Design Only).
With your SPM.mat, you can now use spm_dcm_create(‘my_test_dcm’) to create a DCM with simulated data.
Best,
Peter
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Daniel Gutierrez Barragan
Sent: 04 March 2015 14:27
To: [log in to unmask]
Subject: [SPM] DCM on simulated data
Hello SPM experts/community,
I was wondering if there was a way of feeding a DCM model with simulated data (fMRI BOLD)? I know that the DCM module requires an SPM.mat and VOI.mat (xY, Y) files to be specified, but if I have time courses (disregarding any 1st level analysis and preprocessing) for some predefined seed; the seeds'- geometry, size, and location -, should'n I be able to run DCM with that? I have been browsing through the DCM and VOI routines in order to see what is the essential information that a DCM needs, but I have found lots of variables that are used and found in the VOI files produced from the SPM.mat, which simulated data doesn't have.
My question would be if I should:
a. Try to create fake SPM.mat? or
b. Modify which DCM routine and variables?
In advance I greatly appreciate your help.
Best regards,
--
Daniel Gutiérrez Barragán
CIMeC
PhD. Student, CBS - University of Trento.
