The relationship between prestimulus EEG and visual-evoked potentials

This study was designed to determine the extent to which the electrical brain response evoked by a visual flash stimulus could be predicted from the power spectrum of the 1 second prestimulus EEG. The main purpose of the study was the investigation of how conscious brain states, as indexed by the prestimulus EEG affect the subsequent response to an external stimulus.; Seven male adult subjects endured up to three 70 minute experimental sessions. Each session consisted of 1000 discrete flashes presented with a pseudo-random interstimulus interval of between 2 and 6 s. EEG data were collected from the POz electrode site referenced to linked ears with a sampling rate of 250/s. Power in each of the classical EEG bands (delta, theta, alpha and beta) was extracted from each prestimulus 1 s epoch along with the frequencies of the two predominant spectral peaks. Data trials were sorted based on power levels in each frequency band. Trials were placed in bins for signal averaging and five averaged poststimulus EP features were extracted from each bin.; Multiple regression analysis was used to relate each EP measure to the prestimulus power spectral features on both a within- and between-subjects basis. The between-subjects results reveal that the prestimulus relative power spectrum accounted for up to 45% of overall evoked response amplitude variance, and much less for both N1 (13%) and P2 (22%) peak latency. Low prestimulus alpha power was associated with smaller amplitude EPs than high prestimulus alpha power.; In addition, for those trials in which there was substantial prestimulus alpha power, the alpha phase angle at the moment of stimulation was measured. The alpha cycle was partitioned into octants. Selectively averaged EPs were computed for each octant and correlations were calculated within both the N1 and P2 latency ranges between each octant averaged EP and one consisting of all data trials. The correlograms show that there is a set of preferred phase angles (corresponding to the positive zero cross) in which N1 exhibits a maximum. The correlograms for P2 did not in general exhibit preferred phase angles. These results suggest an intimate relationship between alpha and the particular N1 component accessed in this paradigm, and that P2 is influenced by the state of prestimulus alpha as well.