The analysis of multiple habituation profiles of single trial evoked potentials

Presents a new statistical technique for analyzing multiple habituation profiles of single-trial evoked potentials (EPs). The method is based on a model of the essential characteristics of the habituation process--trial-dependent modulation of amplitude and phase of the underlying brain responses to a repeated stimulus--while recognizing the presence of unsystematic background variation inherent to single-trial registrations. Without assumptions about the time course of habituation, a set of single-trial EPs is decomposed into 1 or more underlying EPs, their habituation profiles across trials, and the corresponding true EPs. The analysis proceeds by spectral analysis, and its validity is illustrated with applications to simulated data.