Fractionating adaptive learning: A meta-analysis of the reversal learning paradigm. (July 2019)

• Record Type: Journal Article
• Authors: Yaple, Zachary A.; Yu, Rongjun
• Is Part Of: Neuroscience and biobehavioral reviews. Volume 102(2019)
• Page Start: 85
• Page End: 94
• Publication Date: 2019-07
• Subjects: Adaptation -- Flexibility -- FMRI -- Prediction errors -- Reversal errors -- Reversal learning; Psychophysiology -- Periodicals; Human behavior -- Periodicals; Animal behavior -- Periodicals; Neurology -- Periodicals; Behavior -- Periodicals; Ethology -- Periodicals; Neurology -- Periodicals; Psychophysiologie -- Périodiques; Comportement humain -- Périodiques; Animaux -- Mœurs et comportement -- Périodiques; Neurologie -- Périodiques; Animal behavior; Human behavior; Neurology; Psychophysiology; Periodicals; Electronic journals; 573.8
• DOI: 10.1016/j.neubiorev.2019.04.006
• Languages: English
• ISSNs: 0149-7634

Abstract

Highlights: Prediction errors yielded more concordant activity within the striatum and amygdala. Reversal errors yielded more concordant bilateral frontal parietal activity. Reversal switching yielded more concordant inferior frontal cortical. All contrasts yielded anterior cingulate and bilateral insulae. Abstract: In constantly changing environments, individuals need to overcome old habitual behaviors in order to learn new associations. Neuroimaging studies have focused on prediction errors, reversal errors and reversal switching in the reversal learning paradigm. Due to the inconsistencies of brain functioning across studies, we attempt to shed light on the concordant activity by performing meta-analyses on different components of reversal learning. While all contrasts yielded anterior cingulate and bilateral insulae, specifically prediction errors yielded more concordant activity within the striatum and amygdala, reversal errors yielded more concordant bilateral frontal-parietal activity, and more concordant inferior frontal cortical occurred from reversal switching. These findings suggest that reversal learning is supported by a core saliency network in all aspects of reversal learning as well as other reward and control related regions in distinct stages of this cognitively complex task. Our meta-analyses results provide stereotaxic maps that can be used for further neuroimaging work on adaptive learning.