In the serial reaction time task subjects must follow a marker by pressing the key which corresponds the current location of the marker. If the transition between marker locations is not random but repeating, the subject acquires the skill to reproduce the sequence by anticipating the next location. Subjects even accelerate if they are not aware of the fact that the sequence is repeating. This skill learning is portrait as implicit learning since subjects are not aware of the fact. To make the repetitions less obvious, the repeating sequence can be alternated by random sequences. In some experiments, the repeating sequence itself is diluted by including one or several random items between each element of the repeating pattern.

The delay between one location and the next seems to influence the learning as well. The shorter the interval the more direct subjects can learn the association from one to the next location.