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.
Table Of Contents
I. Introduction 1 II. Methods 2.1 Participants 6 2.2 Procedure 6 2.3 Neuropsychological tests 2.3.1 Word fluency test (category and letter) 7 2.3.2 Counting span test (forward and backward) 7 2.3.3 Corsi-block test (forward and backward) 8 2.3.4 Wisconsin card sorting test (WCST) 8 2.3.5 Stroop test 9 2.3.6 Attention network test 9 2.3.7 Go/Nogo test 10 2.4 Alternating serial reaction time (ASRT) task 11 2.5 Data analysis 2.5.1 Investigation of participants’ performance in SL 15 2.5.2 Modeling of the SL score 16 2.5.3 Response bias 18 2.5.4 Correlation analysis 19 III. Results 3.1 Success in SL: Higher accuracy and faster RT in Random-High than Random-Low 20 3.2 Decrease of accuracy in Random-Low triggered by response bias 20 3.3 Modeling SL score for the investigation of individuals’ potential and efficiency of SL 24 3.4 A significant correlation between the efficiency of SL and inhibition across all the participants 26 IV. Discussion 4.1 A negative correlation between the efficiency of SL with an inhibitory control 28 4.2 A significance of response bias in SL 30 4.3 Limitations 31 4.4 Conclusions 32 V. References 33