This study measured the changes in the spectrum of the EEG (electroencephalogram) and in the event-related potentials (ERPs) as subjects detected an improbable target in a train of standard stimuli. The intent was to determine how these measurements are related, and to what extent the ERPs might represent phase-locked changes in EEG rhythms. The experimental manipulations were the stimulus modality (auditory or visual), the discriminability of the target, and the presence or absence of distraction. The ERPs showed sensory-evoked potentials that were specific to the modality and a target-evoked P300 wave that was later in the visual modality than in the auditory, and later and smaller when the discrimination was more difficult. The averaged EEG spectrograms showed that targets increased the frontal theta activity, decreased posterior and central alpha and beta activity, and decreased the central gamma activity. The scalp topography of the changes in the alpha and beta activity indicated a posterior desynchronization specific for the visual task and occurring with both targets and standards and a more widespread desynchronization for targets in either modality. Increased phase synchronization occurred during the event-related potentials, but modeling demonstrated that this can be seen when an evoked potential waveform is simply added to the background EEG. However, subtracting the spectrogram of the average ERP from the average spectrogram of the single trials indicated that phase-resetting of the background EEG rhythms can occur during the ERP. The idea that the ERPs and the EEG rhythms "share generators" can explain these findings.