Motivation 

This article explores how the brain processes feedback during learning, offering insights that complement more common educational theories about learning from feedback from cognitive, behaviorist, sociocultural, and constructivist approaches.

Feedback-Based Learning in Education and Neuroscience

Whether for academic achievement, professional development, or lifelong learning, learning from feedback is essential. As any educator (or parent or mentor or friend or …) knows, providing feedback does not ensure that someone will learn from it. For this reason, much research has studied the efficacy of different factors and types of feedback and whether that efficacy differs based on learners’ characteristics (e.g., this book chapter summary discusses factors related to education technology). These variables are studied from many different perspectives, including cognitivism, behaviorism, socioculturalism, and constructivism. However, the brain mechanisms and neural correlates of feedback processing have rarely been part of our understanding of how people learn from feedback.

This systematic literature review collects papers that bridge the educational and neuroscience research on learning from feedback. Only papers with both functional neuroimaging data and experimental behavioral data were included, connecting traditional educational research methods with those that show the brain mechanisms at play. The results show how brain development affects the processing of feedback and how situational, social, and emotional factors affect learning from feedback.

Methods

Inclusion Criteria: In addition to 1) including both functional neuroimaging data and experimental behavioral data, papers also had to 2) present empirical evidence on factors and conditions of learning from feedback, 3) be published in a peer-reviewed journal between 2005-2015, and 4) include healthy, human participants.

Corpus: From an initial pool of 345 articles, 30 met all inclusion criteria. These were separated into two categories

• 12 studies about developmental factors (i.e., compared different age groups)
• 18 studies about situational, social, and emotional factors.

Developmental Factors

These studies showed that children (ages 3-11) learn better from positive feedback compared to negative feedback. The ability to learn from positive feedback remained stable across ages, and it was comparable between children and adults. In contrast, the ability to learn from negative feedback improved as learners matured. This finding was linked to the maturation of prefrontal brain regions responsible for cognitive control, performance monitoring, and managing negative affect. This effect was particularly significant for adolescent learners (ages 11-17) whose brains are in different stages of maturation and whose performance monitoring improves with the development of these areas.

Because performance monitoring is worse in children than in adults, the article argues that feedback is not processed as efficiently by children. As a result, feedback must be more external (i.e., not from internal monitoring) and more explicit for younger learners. Children’s neural immaturity especially impairs learning when feedback is ambiguous or when they need to evaluate its validity. The article summarizes that for preschool and primary school children, feedback should be “kept as clear, direct, and explicit as possible” (p. 253), with a stronger emphasis on positive outcomes than negative ones.

Situational, Social, and Emotional Factors

While ambiguous and irrelevant feedback is particularly ineffective for children, this section showed that it is also ineffective for adult learners. The results found that learning was impaired when feedback was unreliable, even when learners were aware in advance of its invalidity. When feedback is unreliable, the learner must use cognitive control processes to respond appropriately to that feedback, “which goes against the main reinforcement-learning model” (p. 253).

Why this is important

These findings about how evaluating feedback’s validity impairs learning make me think about AI-generated feedback, especially feedback directly to learners. As I argue in my AI literacy training for teachers, evaluating the output of genAI is a critical step in using genAI tools. For this reason, I’ve been distrustful of genAI providing feedback directly to learners, especially young learners, who may not have the knowledge or skills to evaluate the feedback that they are receiving. This article justifies those concerns. It also takes it one step further by arguing that unreliable feedback is detrimental to learning from feedback, even when the learner has the necessary knowledge and skills.

While I am excited about the prospect of affordable, personalized feedback, and genAI seems like a significant step in that direction, this article shows that the quality and reliability of feedback are paramount. Reliability is especially important given that learners will often ignore feedback that does not align with their current understanding and beliefs, especially when it is not given by an authority figure (see the Rich 2016 article summary). Together, these findings show that knowledgeable educators remain essential in the process of providing effective feedback to learners, especially for young children.

Dion, J-S. & Restrepo, G. (2017). A Systematic Review of the Literature Linking Neural Correlates of Feedback Processing to Learning. Open Science in Psychology, 224(4), 247-256.

For more information about the article summary series or more article summary posts, visit the article summary series introduction.