On the influence of task relevance and stimulus probability on event-related-potential components.

Fifteen subjects were presented with series of tones. Any one tone was either loud or soft, and in any one series the probability of one tone intensity was either 0.9 or 0.1. Subjects were instructed to count the frequent tones or to count the rare tones. The stimuli were also presented while the subjects were solving a word-puzzle. Event-related potentials (ERPs) were recorded from 9 scalp locations (F3, C3, P3, FZ, CZ, PZ, F4, C4, P4) referred to linked mastoids. ERP components were measured with a Principal Components analysis and the relations between these measures and the independent variables were evaluated with the ANOVA procedure. This paradigm allowed an evaluation of the effect of stimulus probability, stimulus relevance, and task relevance on the waveform of the ERPs. We conclude that the P350 component is enhanced whenever the eliciting stimulus is both rare and in some sense relevant to the subject's task and the degree of enhancement is greatest when the rare--relevant tone is loud. A "slow wave" component which follows P350 is related to the same variables but has a scalp distribution quite different from that of P350. The slow wave shows a progressive shift in polarity from negative to positive from the frontal to the parietal sites, while the P350 is of nearly equal amplitude (and positive) at the central and parietal sites and has a smaller (positive) and amplitude at FZ. A third prominent component, negative in polarity, peaking at about 210 msec, is most pronounced following rare stimuli, whether or not they were task relevant. The amplitude of N210 tended to be largest at the frontal electrode. This study then demonstrates that when suitable measurement techniques are used, multiple endogenous ERP components can be observed, each related to distinct aspects of cognitive behavior.