Machine Learning Classification of Student Learning Outcomes Using During Online Learning Using Functional Neuroimaging

• In Event: Equity in Science Assessment and Measurement

Abstract

Automated interactive learning management systems and the associated online video-based learning environments are thought to increase student learning outcomes related to science content and practices. The purpose of this study is to examine how hemodynamic response data may be used to develop student level answer predictions via machine learning algorithms as students engage with an online learning management system in a science classroom. Forty participants (n=40), 21 females, and 19 males were recruited from a charter school. Students watched a recorded video consisting of a 20-minute lesson and explanation of the process of DNA replication. Results suggest that hemodynamic responses observed during content presentations to the students is predictive of student correct and incorrect responses.

Authors

• Richard Lamb, East Carolina University

• Knut Neumann, Leibniz Institute for Science and Mathematics Education