Dynamic causal modelling -- cross-validation group index

, ,

Hi SPM experts,

We want to model the difference in dynamic processes between depression patients and health controls.

We have done the PEB and got several significant connectivity. Next, we want to check the predicted ability of the significant connectivity using cross-validation.

However, we are not sure if the results would be different when we used the group index of "1 -1“, for the two groups or if using the de-mean index of “-0.7 0.7”. Thus, we would like to ask if the group index should be de-meaned for cross-validation? Thanks in advance.