Understanding Physiological Responses of Students Over Different Courses

TL;DR

This study investigates how students’ physiological responses reflect engagement across different courses by analyzing intra-student synchrony using data from 23 high-school students wearing Empatica E4 devices over four weeks and ten courses. It combines FastDTW-based synchrony measurements on phasic EDA with comprehensive EDA and HRV feature extraction to relate autonomic patterns to course type. ANOVA reveals significant cross-course differences, particularly involving PE, with Assembly showing lower synchrony than PE, and detailed analyses reveal course-specific arousal and autonomic activation dynamics. The work suggests that unobtrusive physiological sensing can inform personalized instructional strategies, though it calls for larger, more balanced datasets and deeper investigation into course attributes to generalize findings.

Abstract

Student engagement plays a vital role in academic success with high engagement often linked to positive educational outcomes. Traditionally, student engagement is measured through self-reports, which are both labour-intensive and not real-time. An emerging alternative is monitoring physiological signals such as Electrodermal Activity (EDA) and Inter-Beat Interval (IBI), which reflect students' emotional and cognitive states. In this research, we analyzed these signals from 23 students wearing Empatica E4 devices in real-world scenarios. Diverging from previous studies focused on lab settings or specific subjects, we examined physiological synchrony at the intra-student level across various courses. We also assessed how different courses influence physiological responses and identified consistent temporal patterns. Our findings show unique physiological response patterns among students, enhancing our understanding of student engagement dynamics. This opens up possibilities for tailoring educational strategies based on unobtrusive sensing data to optimize learning outcomes.

Figures (5)

• Figure 1: Physiological Synchrony of Students
• Figure 2: Error bars for all the Students
• Figure 3: Student-wise PS scores
• Figure 4: Changepoint and Autocorrelation Analysis
• Figure 5: Time domain features