The scalp topography of P300 in the visual and auditory modalities: a comparison of three normalization methods and the control of statistical type II error.

This study was designed to replicate recent findings suggesting that the P3 component of the event-related potential is dependent on the modality of the eliciting stimulus. When assessing this research hypothesis two methodological problems are of special interest: first, the amplitudes have to be normalized, due to problems with the model of the analysis of variance; second, special care has to be taken regarding the beta error, which is the probability of falsely accepting the null hypothesis of a statistical test. A possible modality independence is associated with the acceptance of a null hypothesis. The first problem was assessed by using different normalization procedures and comparing their results. The second was solved by controlling the beta error. Results for P3 amplitudes from two sessions in which 61 subjects performed in each session an auditory and a visual oddball task (EEG measured at 11 locations) showed no influence of modality on the P3 elicited by the rare, task relevant, stimulus. Influences of modality were observed for the P3 elicited by the frequent stimulus. As it is quite unlikely that P3 generating sources are strongly active during the processing of the frequent stimulus, this effect is possibly due to a component overlap from the vertex potential.