Further exploration of cortical activity during explicit and implicit category rule learning

Presentation Type

Paper

Abstract

Hill, Patel, Murray, and Bohil

We tested the COVIS-based prediction that activation in dorsolateral prefrontal cortex (DLPFC) should remain higher during information-integration category learning than unidimensional rule-based category learning. This effect was expected to be stronger when learners use a unidimensional rule with information-integration categories. We replicated conditions from prior neuroimaging studies (Cincotta & Seger, 2007; Filoteo, et al, 2005) but measured BOLD response in DLPFC using functional near-infrared spectroscopy (fNIRS). The categorization task was made up of two conditions, rule based and information integration, that presented 2D line stimuli with immediate feedback following response. The rule based task requires that participants utilize only 1 dimension while the information integration necessitates both dimensions be utilized to obtain maximum accuracy. Most participants performed above chance in the learning tasks, and we found that oxy-hemoglobin levels were indeed often higher across blocks in the information-integration condition. Decision-bound modeling analyses indicated that in many cases – during information-integration category learning – participants responded according to a unidimensional rule rather than a more appropriate rule combining information from the two stimulus dimensions, supporting our hypothesis.

This research has practical implications in the areas of training and education in both a professional and academic setting. Understanding whether an individual is employing optimal decision strategies is crucial towards maximizing learning and overall performance. Future research will look at the effects of aging on these neural responses associated with implicit and explicit learning.

Comments

Authors: Hill, Audrey, Patel, Pooja, Murray, Matthew, and Bohil, Corey

Share

COinS
 

Further exploration of cortical activity during explicit and implicit category rule learning

Hill, Patel, Murray, and Bohil

We tested the COVIS-based prediction that activation in dorsolateral prefrontal cortex (DLPFC) should remain higher during information-integration category learning than unidimensional rule-based category learning. This effect was expected to be stronger when learners use a unidimensional rule with information-integration categories. We replicated conditions from prior neuroimaging studies (Cincotta & Seger, 2007; Filoteo, et al, 2005) but measured BOLD response in DLPFC using functional near-infrared spectroscopy (fNIRS). The categorization task was made up of two conditions, rule based and information integration, that presented 2D line stimuli with immediate feedback following response. The rule based task requires that participants utilize only 1 dimension while the information integration necessitates both dimensions be utilized to obtain maximum accuracy. Most participants performed above chance in the learning tasks, and we found that oxy-hemoglobin levels were indeed often higher across blocks in the information-integration condition. Decision-bound modeling analyses indicated that in many cases – during information-integration category learning – participants responded according to a unidimensional rule rather than a more appropriate rule combining information from the two stimulus dimensions, supporting our hypothesis.

This research has practical implications in the areas of training and education in both a professional and academic setting. Understanding whether an individual is employing optimal decision strategies is crucial towards maximizing learning and overall performance. Future research will look at the effects of aging on these neural responses associated with implicit and explicit learning.