When we took an EEG data from subjects, does the subject have to be in "neutral state" i.e there is no anger, no sadness, no happiness ...etc. and they also in situation that they neither do any logical operations nor any concerns in their life events?
Often when I read in the literature, they assume that the subject is in this situation above! For example if someone want to elicit the emotion state on the subject from an EEG signal , it's feature extraction process depends on the variance of the signal, and it's assumed that all variance in the EEG is for only the emotions. The brain process a lot of processes, and the signal will contains Emotion , logical operation , the usually processes to organize human functions ... etc .
So what is the rational judgment on this experiment? Assuming that the subject so clever and operated with experimenter, how much will he really control on his mind behavior?
In short : how can we depend on parameters (i.e variance) for evidence when this parameter does not only depend on the behavior that we want to make a decision with?
I know that when we deal with event related potentials (ERPs) , we make the parameters highly dependent on that event , but this is does not apply on inner-voice research.