TF measures were extracted in two ways: (i) for evoked theta activity, decompositions were performed using averaged ERP data, measuring brain activity phase locked to the stimuli, while attenuating nonphase-locked (i.e., induced) activity, and (ii) for induced theta activity, the trial-averaged ERP waveform was first subtracted from each single trial data, with the residual being transformed to the TF domain. The resultant single trial TF surfaces were then averaged across trials to produce a TF representation of the event-related nonphase-locked TF activity. With these methods, one evoked TF representation and one induced TF representation were produced for each electrode site for each subject. To confirm the non-stimulus-phase-locked nature of the induced theta activity, intertrial coherence (ITC), a measure of the extent to which phase locking occurs across trials, was also calculated for induced theta at each electrode for each subject.
