Hi SPM users,
I am running a simple experiment in DCM to observe the effective connectivity between 4 regions and compare the models for best fit. When i check the results, I understand few things looking at them. But the more I go deeper on the results, it seems that I can narrow down my questions ion two concrete questions:
1- When I check the 'C-probability' and 'B- probability' for the effect of condition across six subjects. It seems all the C and B probability are 'stuck' around 0.5. According to my understanding from SPM conference video and basic understanding of DCM, the probability should be minimum 0.9 CI to support the Direct effect in Hz. My question is, how should i trust or 'report' these data sets, if the probability are only around 0.5. Does my direct effect turns to be valid if the C and B probabilities are around 0.5.
2- Many times, the modulatory effect and direct effects show negative values. I do understand the modulatory effects and direct effects are calculated based on differential equations, but does negative effect depict anything which is supposed to be carefully noted?
It will be really helpful if anyone can answer these questions.
Thanks and regards
|