Figure 1. Experiment design and reaction times (RTs) across training blocks.

(a) Serial reaction time (SRT) task apparatus. The experiment consisted of 1225 trials, divided in 25 blocks of 49 trials. In each trial the volunteer should press the button on the mouse devices corresponding to the stimulus location. (b) Each volunteer of Experiment 1 performed one of six sequence types displayed. The sequences could be either repetitive (deterministic) or probabilistic, shown on left or right, respectively. (c) Experiment 1: RTs across training blocks for each group of volunteers. All volunteers performed random sequences on blocks 1 to 4 and 23 to 25; from blocks 5 to 22 they performed one of the sequences presented in Fig. 1b. Results are denoted by the mean of the median RT for each volunteer in each training block; bars denote standard error of the mean (N = 8). Notice that the RTs increased with the complexity of the sequence. (d) Experiment 2: median RTs across training blocks for each session. Experimental design was the same of Experiment 1, except for (i) a wide variety of sequences was used (see Supplementary Table S1), and (ii) only one volunteer performed all the sequences in different sessions (one session per day). Notice that the RTs cannot be explained only by the session number.