Changes in P3 waves with event repetition: long-term effects on scalp distribution and amplitude.

Event-related brain potentials to tachistoscopically presented events were recorded from adults 24--36 years of age. Subjects counted the number of target events (P = 0.12) randomly interposed in sequences of background events (either P = 0.88 or P = 0.76). In some sequences, slides bearing the letter A were targets and those bearing the letter B were backgrounds; the reverse was true in the other sequences. Also interposed (P = 0.12) in these sequences of targets and backgrounds were one of two types of non-target events: in some sequences non-targets were slides (termed dims) bearing any one of the letters C to Z and having one-tenth the luminance of the target and background slides, while in the others, they were slides bearing 'novel' patterns, each consisting of a different, quasi-random, unrecognizable color pattern. The scalp amplitude distribution of P3 waves to novels were initially frontal, but, with repeated presentations of novel events, they became parietal. The scalp distributions of P3 waves to targets and dims were both initially parietal and did not change with repeated presentations of target and dim events. P3 amplitudes to targets were unaffected by repeated presentations of target events. However, P3 amplitudes to dims and novels were affected. P3 amplitudes to dims decreased linearly at Fz, Cz and Pz with repetition of dims. P3 amplitudes to novels decreased at Fz and increased at Pz with repetition of novels. An abbreviated analysis of possible effects of sequential event probabilities on P3 amplitudes and latencies to novels, dims and targets was performed and no effects were found. It is suggested that P3 waves may be affected by at least 4 factors: the facility with which events may be categorized by subjects, the importance attached to events (via task instructions, subjective assessment of the contents of events, etc.), a priori category probability, and sequential event structure.