Dear Doug,
Simple answer is, yes you can compare (with BMS) models with different
input setup.
HTH,
Chris
Le 10/01/2013 9:17, Doug Davidson a écrit :
> Dear SPM list,
>
> Is it possible to compare DCMs with different exogenous input (one
> versus two) using BMS?
>
> In our experiment, subjects heard either left ear or right ear
> stimulation (i.e., monaural tones) and we recorded MEG. We would like
> to construct different DCM models for the evoked response. The models
> include two sources: left and right temporal cortex. I would like to
> compare models with different specifications of the input: One DCM has
> a single input to both left and right sources (DCM.C=[1 1]'), while a
> second DCM has one input to the left and another input to the right
> (DCM.C=[1 0; 0 1]). We are trying to test whether the dual input model
> will provide a better account of the response.
>
> I have fit these models, and both seem to capture the main data
> pattern, but I would like to compare them with BMS. For a given
> condition (e.g., left ear stimulation), I construct both the unitary
> input model and the dual input model. The only difference is the
> specification of the input. Can I use BMS to compare these different
> input patterns?
>
> If I do the same for the other condition (e.g., right ear
> stimulation), can I then add this second condition as another session
> in the BMS, or is it better to run a separate BMS for each?
>
> Thanks a lot for your help.
>
> -Doug
>
> Doug Davidson
> Basque Center on Cognition, Brain and Language
> Basque Country, Spain
>
|