Article Review - Learning and altering behaviours by reinforcement: Neurocognitive differences between children and adults

Shepard, E., Jackson. G. M., Groom, M. J. (2014). Learning and altering behaviours by reinforcement: Neurocognitive differences between children and adults. Developmental Cognitive Neuroscience, 7, 94-105.

Background: Impaired reinforcement learning has been implicated in the pathology of several neurodevelopmental disorders including ADHD and Tourette syndrome. Previous studies have reported performance differences between children and adults in reinforcement learning. The specifics of these differences remain unknown and the precise deficits associated with neurodevelopment disorders remain unknown.

Purpose: To establish whether children differ from adults in behavioral and brain correlates of learning before they undergo the significant maturational changes that take place during adolescence.

Hypothesis: Children would show smaller learning-related changes in performance and ERP amplitudes, reflecting poorer learning ability at this age. Children would also show greater disruptions to performance and greater reliance on feedback.

Definitions: Reinforcement learning – the ability to learn and modify behaviors based on the positive and negative outcomes of our actions

Selection: 14 children (9-11 years old, 12 male, mean age 10.2) and 15 adults (5 male, mean age 25.5) recruited from local primary schools and the University of Nottingham, UK. Participants were typically developing with no known neurological or psychiatric problems, which may have impacted brain function, right-handed, and had normal or corrected-to-normal vision.

Methods: Participants completed 5 blocks of a reinforcement-learning task with stimulus-response associations to associated visual stimuli through trial-and-error. Each block had 48 trials and at the fourth block, the stimulus-response associations reversed unexpectedly. Circular yellow happy-face images and blue sad-face images were used as positive and negative feedback. The words “too slow” were displayed in green for late responses. EEG data was recorded continuously throughout task performance. 

Results: Children and adults did not differ in learning-related performance in the initial phase. Performance was, however, significantly more disrupted in children than adults in the reversal phase and when reinforcements changed. There was also a developmental difference in neural correlates of consolidation and feedback processing.

Conclusion: Study findings suggest that children, like adults, use feedback to drive goal-directed learning behavior. Children use such feedback even more than adults as indicated by ERP amplitudes, although they seem to have difficulty integrating feedback information. When the processing demands become too great for children, they may experience a reduced capacity for learning and consolidating information to alter behavior to produce correct responses.

Implications: Given the nature of performance disruptions when children were unable to consolidate feedback, this has potential implications on the provision of developmentally appropriate tasks. Children seem to rely more heavily on experience-based learning and utilize eternal reinforcements at greater frequency than adults. Children also, perform the best when reinforcement feedback is clear and consistent. 

Future Research: Further research is needed to investigate the possible correlation between differences in reinforcement learning and differing styles of learning.