rDCM : how to make a predefined DCM structure for synthetic data

[code2023-source]

Hi Fraessle,

Sorry to brother you. I would like to ask you some questions about rDCM. In the rDCM toolbox, I don't konw how to build the predefined DCM structure. I see a DCM_LargeScaleSmith_model1.mat, it contains many parameters, but how to make it? Do I need to create a DCM structure for synthetic data by using SPM? Could you tell me the related instructions or helper function?

Thank you for your kindness!

Best regards,

Aixiao

[StefanFraessle]

Hi Aixiao,

My apologies for the delay in responding. Somehow, your question must have gone completely unnoticed. In case this is still an open question:

Yes, at the moment you would need to specify the DCM structure by using SPM routines or by writing your own little MATLAB script that sets up the DCM structure. In principle, the DCM structure is nothing but a MATLAB struct that contains the necessary fields (which you can see in the example DCM_LargeScaleSmith_model1.mat file).

If you want to use SPM to set up the model, you could in principle use the GUI which will guide you through the specification of the DCM structure step-by-step. If you are completely new to DCM, this might be a good start to get used to the structure. For this purpose, please also refer to the SPM manual, which you can download here.

Having said this, for whole-brain models (as used in rDCM), the GUI is very tedious. Hence, you might want to write your own MATLAB script that utilizes the respective SPM routines. If you want to get an idea of how to specify a DCM structure from scratch in your MATLAB script, you could check out the attention-to-motion example provided on the SPM homepage. Specifically, you would like to download the dcm_spm12_batch.m file which contains a code example and pointers to useful SPM routines.

As a final note: I'm thinking about providing a little helper function in a future release of the rDCM toolbox which could facilitate the specification of whole-brain models in rDCM.

I hope this is helpful and, again, my apologies for this long delay.

With very best wishes,
Stefan