What Is Implicit About Implicit Learning?
Presentation Type
Poster
Abstract
Categorization is a fundamental cognitive process that underlies our ability to understand items as distinct and related. Present evidence suggests that we make categorical decisions by means of rules or boundaries between categories. One prominent theory, COVIS (Competition between Verbal and Explicit Systems), expands upon decision bound theory by assuming separate, competitive rule-based (RB) and information-integration (II) category learning systems, mediated mainly by the prefrontal cortex (PFC) and striatum, respectively. The RB system relies on easily verbalized one-dimensional rules while the II system requires integration of more than one stimulus dimension. Present fMRI neuroimaging and behavioral data suggest that this dual-system learning occurs, but it remains unclear as to how these decision boundaries develop over time (i.e., whether participants are aware that they are learning) and the types of rules participants apply to both RB and II tasks (i.e., whether participants are applying the “correct” rules). Participants in this study will categorize lines varying in length and orientation as belonging to one of two categories while corrective feedback shapes learning. Participants in the RB condition need only attend to line length but integrate both length and orientation in the II condition. The manipulation introduced here is the addition of a scale by which participants will attribute their categorical decisions to guess, intuition, rules, or memory, with responses serving as representations of participants’ consciousness of their learning when compared to accuracy across trials. Additionally, we will measure participants’ neural responses in PFC using a 20-optode fNIRs cap designed by NIRx (the NIRsport). fNIRs measures the relative changes in concentrations of oxy- and deoxy-hemoglobin (the same biomarker as fMRI) using LED light. Ideally, we expect an increase in accuracy across trials for both tasks. PFC activation should decline over trials in the RB condition in which participants learn and apply the easily verbalizable rule but remain stable while participants attempt to learn the more difficult, multi-dimensional rule in the II condition. In the RB task, we anticipate participants’ confidence to be higher as it relates to the easily acquired rule, their consciousness of which will be reflected in improved accuracy over trials. In the II task, however, it is likely that participants will believe they are guessing; with improved accuracy, this would suggest that learning is occurring but participants are unaware of it. Changes in participants’ relative hemodynamic response over trials should reflect whether they learned the appropriate rule in each task. Both behavioral and fNIRs data will further evidence predictions made by COVIS as it pertains to dual-system learning, but the addition of subjective assessment provides a unique opportunity to determine the consciousness of participants’ learning and explain the implicit acquisition of categorization knowledge. In evaluating COVIS as a theoretical model, we can further the neurobiological understanding of categorization while developing a framework by which to explain the acquisition of category knowledge and improve of category learning.
What Is Implicit About Implicit Learning?
Categorization is a fundamental cognitive process that underlies our ability to understand items as distinct and related. Present evidence suggests that we make categorical decisions by means of rules or boundaries between categories. One prominent theory, COVIS (Competition between Verbal and Explicit Systems), expands upon decision bound theory by assuming separate, competitive rule-based (RB) and information-integration (II) category learning systems, mediated mainly by the prefrontal cortex (PFC) and striatum, respectively. The RB system relies on easily verbalized one-dimensional rules while the II system requires integration of more than one stimulus dimension. Present fMRI neuroimaging and behavioral data suggest that this dual-system learning occurs, but it remains unclear as to how these decision boundaries develop over time (i.e., whether participants are aware that they are learning) and the types of rules participants apply to both RB and II tasks (i.e., whether participants are applying the “correct” rules). Participants in this study will categorize lines varying in length and orientation as belonging to one of two categories while corrective feedback shapes learning. Participants in the RB condition need only attend to line length but integrate both length and orientation in the II condition. The manipulation introduced here is the addition of a scale by which participants will attribute their categorical decisions to guess, intuition, rules, or memory, with responses serving as representations of participants’ consciousness of their learning when compared to accuracy across trials. Additionally, we will measure participants’ neural responses in PFC using a 20-optode fNIRs cap designed by NIRx (the NIRsport). fNIRs measures the relative changes in concentrations of oxy- and deoxy-hemoglobin (the same biomarker as fMRI) using LED light. Ideally, we expect an increase in accuracy across trials for both tasks. PFC activation should decline over trials in the RB condition in which participants learn and apply the easily verbalizable rule but remain stable while participants attempt to learn the more difficult, multi-dimensional rule in the II condition. In the RB task, we anticipate participants’ confidence to be higher as it relates to the easily acquired rule, their consciousness of which will be reflected in improved accuracy over trials. In the II task, however, it is likely that participants will believe they are guessing; with improved accuracy, this would suggest that learning is occurring but participants are unaware of it. Changes in participants’ relative hemodynamic response over trials should reflect whether they learned the appropriate rule in each task. Both behavioral and fNIRs data will further evidence predictions made by COVIS as it pertains to dual-system learning, but the addition of subjective assessment provides a unique opportunity to determine the consciousness of participants’ learning and explain the implicit acquisition of categorization knowledge. In evaluating COVIS as a theoretical model, we can further the neurobiological understanding of categorization while developing a framework by which to explain the acquisition of category knowledge and improve of category learning.