Exploring the Neurocognitive Bases of Statistical Learning

Statistical Learning (SL) involves the extraction of organizing principles from a set of inputs. Recent advances in SL suggest that SL is a componential construct. To better characterize the componential nature of SL, a strategy may be to turn to literature regarding memory and learning. The current study sought to extend the literature by further characterizing the componential nature of. Aim 1 examined the effect of instruction type (explicit, implicit) on direct and indirect (explicit, implicit) indices of visual statistical learning (VSL) performance. Several studies have suggested explicit instructions shift engagement of additional explicit memory resources improving performance. There were no differences in indirect or direct measures of VSL performance. However, the relationship between direct and indirect measures of VSL was affected by instructional condition suggesting the processes underlying VSL may have been affected. Aim 2 examined the relationship between VSL performance and implicit and explicit memory/learning. Further, Aim 2 examined whether instructional condition affected the relationship between VSL and multiple memory systems. The relationship between the direct measure of VSL and explicit and implicit memory was inconsistent. However, the direct measure patterned similarly across explicit and implicit memory (positive relationship, not affected by instructional condition). The relationship between the indirect measures of VSL and multiple memory systems was similarly inconsistent, but had a similar patterning in the significant cases (positive relationship in the explicit condition, but negative relationship in the implicit condition). This suggests the indirect measure of VSL was affected by the instructional condition to differentially emphasize aspects of memory systems. In addition, In recent years, several methodological issues have been identified regarding measures established in the literature. To address the inconsistencies in the findings and these concerns, the psychometric properties of the established measures were examined. Exploratory Aim 3 sought to improve upon the processing of these measures using advanced statistical methods and provide recommendations regarding best practices for individual differences analyses. In Exploratory Aim 4, the first set of results were revisited in an exploratory manner using the insights gained from the updated measures. Implications for the characterization of the componential nature of SL were discussed