To investigate a human sequential learning Nissen & Bullemer (1987) developed a serial reaction time task. This task has been used in dual-task situations to evaluate the role of attention in acquisition of the event sequence. In this article, we pointed out some problems of performance measures of learning, such as reaction times and error rates. The ambiguity of the relationship between performance measures and psychological processes supposed to occur in the task results in the disagreement among the interpretations of the results obtained in a variety of experiments. We suggest the usefulness of psychophysiological measures that could directly reflect our brain's activities, and introduce a several studies where some electrophysiological measures were devised to explore the explicit and implicit aspects of sequential learning.