In this work we show the application of a measure of entropy defined from the wavelet transform, namely the wavelet entropy (WS), to the study of event-related potentials (ERPs). WS was computed for ERPs recorded from nine healthy subjects with three different types of stimuli, among them target stimuli in a cognitive task. A significant decrease of entropy was correlated with the responses to target stimuli (P300), thus showing that these responses correspond to a more "ordered" state than the spontaneous EEG. Furthermore, we propose the WS as a quantitative measure for such transitions between EEG ("disordered state") and ERP ("ordered state").