The effects of serial order in long sequences of auditory stimuli on event-related potentials.

The generalizability of the model relating P300 amplitudes to subjective probabilities, developed by Squires et al. (1976) for random series of events, was examined with respect to long sequences of repetitions, which had been restricted in randomness, allowing only for sequences of between 4 and 12 frequent clicks. Subjects were asked to silently count either the rarer of two clicks, presented with a probability of .10, or light stimuli occurring with the same temporal distribution. Within these limits there was an increase in P300 amplitude not only to the rare clicks, but also—contrary to predictions from the model—to the frequent non-target clicks following longer series of repetitions, provided that clicks had to be counted. A plausible interpretation might be that the longer the series of repetitions in long non-random sequences with low predictability, the more the subjects become involved in the “stimulus evaluation” of both kinds of events. A similar increase across serial position was found for the N100 component to frequent clicks when the auditory modality was defined as task relevant, and was interpreted as progressive focusing of selective attention. For the rare clicks there occurred a decrease in a slow Negative Shift with peak amplitude at 220 msec after long series of non-targets, possibly reflecting a facilitative effect of the focused attention on decision processes related to target detection. There were no differences between normals and chronic alcoholics with respect to the above-mentioned effects.