Abstract

The traditional view holds that event-related potentials (ERPs) reflect fixed latency, fixed polarity evoked responses that appear superimposed on the 'background EEG'. The validity of the evoked model has been questioned by studies arguing that ERPs are generated at least in part by a reset of ongoing oscillations. But a proof of phase reset that is distinct from the 'artificial' influence of evoked components on EEG phase-has been proven difficult for a variety of methodological reasons. We argue that a theoretical analysis of the assumptions and empirical evaluation of predictions of the evoked and oscillatory ERP model offer a promising way to shed new light on mechanisms generating ERPs that goes well beyond attempts to prove phase reset. Research on EEG oscillations documents that oscillations are task relevant and show a common operating principle, which is the control of the timing of neural activity. Both findings suggest that phase reorganization of task relevant oscillations is a theoretical necessity. We further argue and show evidence that (i) task relevant oscillations exhibit a typical interactive and task relevant relationship between pre- and poststimulus power in the theta and alpha frequency range in a way that small prestimulus power is related to large poststimulus power and vice versa, (ii) ERP (interpeak) latencies and (iii) ERP amplitudes reflect frequency characteristics of alpha and theta oscillations. We emphasize that central assumptions of the evoked model cannot be substantiated and conclude that the ERPR model offers a new way for an integrative interpretation of ongoing and event-related EEG phenomena.