About our Paper - Test-enhanced learning with feedback

Discover the story behind our paper, "Neural activations associated with feedback and retrieval success", which was published in npj Science of Learning

Like Comment

Our article, "Neural activations associated with feedback and retrieval success", was published in the journal npj Science of Learning recently.

I discuss the work behind the paper below.

Substantial research in cognitive psychology has challenged the conventional view that tests merely assess knowledge, and are a neutral event for the process of learning. Evidence from laboratory and classroom studies have showed that taking tests, compared to a diversity of other pedagogical methods, during initial learning is superior as it improves retention of the to-be-learned material – which is the main goal in education. This empirical phenomenon is known as the “testing-effect”.

What are the critical factors involved? Two such factors are retrieval success and feedback. Given the basic idea with test-enhanced learning – “test yourself during initial learning” – simply means that the initial knowledge level is generally low, so even if testing produces superior retention in the absence of feedback, correct answer feedback is a critical factor to support improvement in learning. Feedback compensates for unsuccessful retrievals, and retrieval changes the memory. But what is the neuro-cognitive processes that produces the testing-effect?

Here, we used fMRI to examine functional brain activity related to test-enhanced learning with or without feedback. Akin to a school situation, subjects learned foreign vocabulary across three consecutive cued recall tests with or without correct-answer feedback. Functional brain activity responses were analyzed in relation to both the test event (i.e. retrieval event) and the feedback event, respectively and across repetitions.

Higher hippocampal activity was found for feedback compared to no feedback, and learning from feedback was related to increased activity predominantly in the insula. Hippocampus is a key region for learning and memory, and insula involved in successful encoding. For retrieval success, up-regulated activity in fronto-striatal regions were evident at the first successful retrieval with a marked decrease across consecutive tests. These results indicate that while both feedback and retrieval success are key aspects that fosters the testing-effect; they operate at different functional levels in the brain. Of particular interest was the finding that test-enhanced learning taxes executive processes early in the learning-phase, but becomes less executively demanding as a function of improved learning. This raises an important question: is test-enhanced learning equally effective for all students, independently of individual variations in cognitive proficiency (e.g. working memory)? Such knowledge is of both scientific and educational significance and is a focus for future studies.

To date, we have limited knowledge about how individual variations in cognitive proficiency influence the effectiveness of the testing effect. Some studies do suggest, in line with the results from the current study, that test-enhanced learning with feedback might be a specifically powerful way to improve learning for students with lower cognitive proficiency.

In sum, the power of testing memory during learning does not only change your memory – it makes it also more probable that you will remember the to-be-learned material in the future!

Read the article in full here


Carola Wiklund-Hörnqvist

Post doc, Department of psychology


Go to the profile of Andrew Lyman-Buttler
about 4 years ago

Thanks for publishing this in an open journal. I teach high school chemistry and I've spent the last few years developing what I hope will be a broadly replicable model for blended learning, with traditional face-to-face instruction plus a Moodle site for progress monitoring. I've mostly eliminated the distinction between formative/summative assessments and practice assignments because computer-based item generation and grading means I can allow unlimited "quiz retakes" which really function as retrieval practice. I'm always looking for ways to make the system more effective so I appreciate your insight.

Andrew Lyman-Buttler
Science Teacher
Thomas Edison High School
Minneapolis, MN, USA