Individual differences in statistical learning

Abstract

Statistical learning (SL) is an essential learning mechanism which enables humans to extract probabilistic regularities from the world. Even though previous studies have examined quality of SL, they overlooked quality of learning and efficiency of learning in SL. Moreover, the ultimate quality of learning with training, that is, the potential of learning has been known to be dissociated with the efficiency of learning. Therefore, in the present study, we elucidated the potential as well as the efficiency of SL separately and investigated which processes of executive functions mainly exerted on them. Specifically, we quantified the efficiency and the potential of SL through mathematical modeling, using participants’ performances in alternating serial reaction time (ASRT) task and correlating them with individuals’ executive functions such as set shifting, updating, and inhibition. In results, low efficiency of SL was closely related to good inhibitory function whereas potential of SL was not associated with any of the executive functions. Our results, via a novel approach of mathematical modeling, shed lights on the overarching role of inhibition in the efficiency of SL.