For time-locked/ phase locked events both functional and effective connectivity are have been well documented and reported. What if the EEG signals are continuous and some one is interested to continuously analyze the connectivity patterns and/or use them as features/input of a classifier.