Figure 1.

The cued Alternating Serial Reaction Time (ASRT) task and experimental procedure. (a) Pattern and random trials were presented in an alternating fashion. Pattern trials were marked by a picture of a dog’s head, and random trials were marked by a picture of a penguin. (b) An example of the sequence structure. Numbers indicate pattern trials, and ‘r’ indicates a randomly selected location out of the four possible ones. The alternating sequence makes some runs of three consecutive trials (labeled as triplets) more probable than others, called high-probability and low-probability triplets, respectively. Among high-probability triplets, the last element of the triplet can be either pattern or random. Based on this, we could determine pattern triplets that are always of high probability (orange shading in panel B and orange font in panel C) and random high-probability triplets (blue shading in panel B and blue font in panel C). Among low-probability triplets, only random low-probability triplets can occur (green shading in panel B and green font in panel C). (c) The underlying learning processes measured by the task. Statistical learning is calculated by contrasting the accuracies or RTs on the random high and random low trials (blue vs. green, the right column of the table). Sequence learning is quantified by contrasting the accuracies or RTs on the pattern and random high trials (orange vs. blue, the top row of the table). The table presents the calculation of learning processes on RT data. (d) The design of the experiment. The experiment was composed of three sessions. The Learning Phase consisted of four epochs (one epoch contained 5 blocks, and each block consisted of 85 stimuli), followed by a 5-h offline period then the two-epoch-long Testing Phase on the same day. The Retesting Phase with four epochs was administered ca. one year later. Figure 1A, 1B, and 1C are adapted from Nemeth, et al. 21, Fig. 1D is adapted from Kóbor, et al. 27.