Effects of different AI-driven Chatbot feedback on learning outcomes and brain activity

Area of Science:

• Neuroscience
• Educational Technology
• Artificial Intelligence

Background:

• AI chatbots offer instant feedback for learning.
• The effects of different AI feedback types on learning and brain activity are not well understood.

Purpose of the Study:

• To investigate the impact of metacognitive, affective, and neutral feedback from educational chatbots on learning outcomes and brain activity.
• To identify neurophysiological signatures associated with different feedback types.

Main Methods:

• Utilized functional near-infrared spectroscopy (fNIRS) to measure brain activity.
• Compared learning outcomes (transfer scores, retention scores, metacognitive sensitivity) across three feedback conditions: metacognitive, affective, and neutral.
• Employed a machine learning model to identify brain regions predicting transfer scores.

Main Results:

• Metacognitive feedback led to higher transfer scores, metacognitive sensitivity, and increased brain activation in the frontopolar area and middle temporal gyrus.
• Affective feedback improved retention scores and showed higher activation in the supramarginal gyrus compared to neutral feedback.
• Neutral feedback resulted in higher activation in the dorsolateral prefrontal cortex.

Conclusions:

• Different types of AI chatbot feedback have distinct effects on learning outcomes and brain activation patterns.
• Metacognitive feedback shows particular promise for enhancing learning transfer and cognitive control.
• Findings provide neurophysiological evidence for the effectiveness of tailored feedback in AI-driven education.