rDCM example for empirical data

[pfdamasceno]

Hi, I was wondering if you guys could provide a pointer on how to run rDCM given an empirical fMRI.

This would be similar to thetapas_rdcm_tutorial.m tutorial but without the step of reading a pre-defined DCM file.

Do I first need to create a DCM struct? And, if so, is there a helper function that assists in this step?

Thanks. Pablo

[StefanFraessle]

Hi Pablo,

My apologies for the terribly delayed response - your question must have somehow skipped my attention.

In case this is still relevant (and you haven't figured it out by yourself in the meantime anyway): Yes, your intuition is correct that estimating rDCM for empirical data is essentially the same as for synthetic data, as shown in the tapas_rdcm_tutorial.m function. Obviously, the synthetic data generation step would fall away - instead, you would directly load the DCM.mat that contains the model and the empirical data.

We do not provide a helper function to generate the DCM structures because rDCM is designed to read in the standard DCM structures from SPM. Hence, you can use SPM to create your DCMs and then feed it directly into the estimation routine of rDCM: tapas_rdcm_estimate.m

I hope this helps. Let me know if you have any other questions or problems.

Cheers Stefan

[zenkavi]

Reviving this old thread instead of opening a new issue:

Is there any documentation on the DCM_LargeScaleSmith_mode1.mat file used in tapas_rdcm_tutorial.m. Looking at the model I understand it's a 50 node network with 25 inputs but I can't trace where/how/why it was generated.

[BPaton]

The dataset (if I a remembering it correctly) is described here:

https://www.fmrib.ox.ac.uk/datasets/netsim/

including a link to the NeuroImage paper that goes into parameter recovery for various network discovery algorithms.

Thanks.

Bryan.

On Tue, 23 Feb 2021, 07:11 Zeynep Enkavi, [email protected] wrote:

Reviving this old thread instead of opening a new issue:

Is there any documentation on the DCM_LargeScaleSmith_mode1.mat file used in tapas_rdcm_tutorial.m. Looking at the model I understand it's a 50 node network with 25 inputs but I can't trace where/how/why it was generated.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/translationalneuromodeling/tapas/issues/59#issuecomment-783642957, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABYA5W5A5DVM7EBXKIVJAGLTAK26BANCNFSM4HT7T6OQ .

[StefanFraessle]

Hi Zeynep, hi Bryan,

Thanks for the question and for the quick help - great to see the TAPAS community helping each other out :)

Maybe just to add some further details (in the hope that this is helpful): As Bryan already pointed out, the network structure resembles the one introduced by Smith and colleagues in their NeuroImage paper from 2011. The idea of the model is to - very loosely and superficially - account for the small-world architecture of the human brain. That is, the model comprises clusters (or cliques) of strongly connected regions which are then linked via a few long-range connections. This yields a network that combines both high clustering and short path length, which is the defining property of a small-world network.

Based on the proposal by Smith and colleagues, we had used the model in one of our initial publications on rDCM (here). This is why the model is the exemplary network architecture used in tapas_rdcm_tutorial.m of the rDCM toolbox.

I hope this helps. If you have any further questions, please let me know.

All the best, Stefan

[zenkavi]

Thank you both for your prompt and informative responses!