Dear Pinar,
When comparing families of models that differ w.r.t. a key parameter, you should use family-level inference (Penny et al 2010). It is specified via BMS in the batch interface.
Penny WD, Stephan KE, Daunizeau J, Rosa MJ, Friston KJ, et al. (2010) Comparing Families of Dynamic Causal Models. PLoS Comput Biol 6(3): e1000709. doi:10.1371/journal.pcbi.1000709
Best wishes
Martin
> On 14 Jun 2014, at 22:47, "Pinar Pnr" <[log in to unmask]> wrote:
>
> Hi dear all,
>
> I am working on fmri data and I want to compare some models of DCM. There are 2 groups of models for one subject. I want to discriminate if the difference of two groups of models evidence is more than 3.
> Can I integrate the log evidences of the models of each group and compare them?
|