Chunking feedback in instructor-learner interaction facilities long-term learning transfer: behavioral and fNIRS hyperscanning studies

Abstract: Feedback is a crucial driving factor for attitude and behavior change, as well as the acquisition of knowledge and skills. Previous research has shown that providing and receiving feedback with elaborated content in teacher-student interactions can deepen the comprehension of conceptual knowledge and promote knowledge transfer to new contexts. However, the impact of altering the presentation of feedback during interaction on long-term learning and its underlying neurocognitive processes remains unclear.
To address this question, the current study employed a naturalistic teacher-student dyadic question-answer feedback task. It conducted both a behavioral experiment and a fNIRS hyperscanning experiment to investigate the facilitative role of chunked feedback in long-term knowledge transfer, its cognitive processes, and the underlying neural basis during interpersonal interactions. In Experiment 1, students learned psychology concepts and received chunked (i.e., a set of two related concepts) or separate (i.e., one concept) feedback (i.e., correct answer and example) from the teacher. The prior knowledge levels were manipulated through a learning introduction phase. The between-subject design included feedback presentation (chunking vs. separate) × prior knowledge (high vs. low). Learning performance was measured after the task regarding knowledge recognition and transfer, with a second test conducted after 7 days to explore the long-term effect. Experiment 2 adopted fNIRS hyerscanning to simultaneously record teachers’ and students’ brain activity during interactions. Additionally, a pseudo-chunk (i.e., a set of two less-related concepts) feedback group was included to clarify potential confounding between feedback timing and format.
The results of Experiment 1 showed that in teacher-student interaction, presenting feedback in a chunked manner, compared to separate feedback, was more beneficial for the long-term transfer performance of students with lower prior knowledge. With the introduction of the pseudo-chunk feedback group as a control, Experiment 2 replicated this finding and revealed no significant difference in long-term transfer performance between the pseudo-chunk and separate feedback groups. This indicates that merely changing the timing of feedback does not lead to long-term transfer gains for students with lower prior knowledge. Furthermore, it was found that chunked error correction partially mediated the relationship between feedback presentation format and long-term transfer performance. In Experiment 2, students receiving chunked feedback reported greater cognitive effort compared to those receiving non-chunked feedback, but no relationship was found between cognitive effort and learning. These results provide support for the possibility that the long-term transfer effect of chunked feedback for low-prior-knowledge students may be due to more organized error correction rather than cognitive effort. Additionally, Experiment 2 revealed that during the process of providing and receiving chunked feedback, greater brain-to-brain synchrony was observed in the frontal and parietal areas between teachers and students, with frontal brain-to-brain synchrony predicting long-term transfer performance and chunked error correction, suggesting it as the interpersonal neural basis of chunked feedback promoting effective error correction and facilitating long-term deep learning such as transfer.
This study provides practical insights for improving the effectiveness and efficiency of feedback in real classroom settings. Moreover, it suggests that interpersonal frontal brain synchronization may play a crucial role in organized information representation, effective knowledge correction, and long-term transfer during real instructional interactions, thereby contributing to a better understanding of the cognitive and neural basis of instruction and learning activities.