To optimise learning in education remains an ongoing challenge. This study aims to identify an indicator for learning from physiological data, developing a quantitative video data analysis to link sensory input and movement to learning through sympathetic activation. Such measures are believed to help improve teaching quality by enabling assessment of different instructional activities, even within one lesson, due to the high temporal resolution.
More specifically, this is done by quantifying brightness, brightness change, loudness, and AI-estimated movement from video data and validating their predictive power for electrodermal activity (EDA). EDA is a well-known indicator of sympathetic activation, i.e., the system behind the ‘fight or flight’ response.
To this end, pre/post-tests, video, and EDA data were collected from 12 students (17-18 y) taking part in a physics lesson where instructional activities were varied and in a laboratory experiment with 100 students (12-21 y) watching two different versions of one instructional video, which only differed in strength of sensory input.
Findings indicate a causal relationship between sensory input and sympathetic activation. Moreover, an EDA variable correlating with learning was identified and found to vary according to different instructional activities and could be predicted by sensory input strength but not movement. These findings imply that video data can be used to assess sensory input and movement, whilst EDA measures can inform research on student learning. Both measures work on short timescales, making them appropriate for comparing different instructional activities within school lessons.
