Using N400 Event-Related Potential to Detect Differences in Design-Mode and Belief-Mode Scaffold Use

Background/Objectives: Scaffolding plays a vital role in sustaining collaborative discourse and shifting attention. However, current research lacks a detailed understanding of how scaffold use affects participants' discussions at the neural level. This paper investigates whether epistemic scaffold types (Design-mode and Belief-mode) influence participants' collaborative discourse and subsequently modulate N400 event-related potential amplitude during sentence processing. Methods: Participants in two experimental conditions engaged in an online discussion using scaffolds either representing Design-mode (My theory) or Belief-mode (I agree/I disagree). Participants then individually completed a stimulus-based decision-making task involving sentences representing the two modes. Pre- and post-surveys assessed changes in participants' attitudes across the study. Machine learning models were used to examine participants' discourse patterns while event-related potential (ERP) analyses of the N400 component assessed neural responses during the decision-making task. Results: Machine learning analyses indicated differences between the two scaffold modes, while ERP analyses revealed a modest N400 amplitude difference between the two modes, during the 380-430 ms time window. Conclusions: Findings suggest that epistemic scaffolding can influence collaborative discourse and neural processing, offering implications for the design of scaffolded learning for researchers and practitioners.