Use of Electrodermal Wristbands to Measure Students' Cognitive Engagement in the Classroom

Abstract

A pilot project was conducted to study the feasibility of using electrodermal activity sensors embedded in a watch-like device to measure skin conductivity in real time. In the field of education, it may be interesting to use this technology to assess the students' cognitive engagement in the classroom. A few volunteer students as well as the professor were wearing an Empatica E4 wristband during some class periods where different activities were organized such as lectures, workshops and exams. Monitoring several individuals simultaneously makes possible to compare the collected data among students and between the students and the professor. Also, since the activities were weekly repeated, it was possible to assess to which extent the observed patterns were similar from one group to the other. In brief, the collected data is very difficult to interpret, since some external factors seem to have a significant effect on the measurements. Indeed, discrepancies are observed in the data curves representing the students’ electrodermal activity. Also, the data generated by the professor is quite different from one group to the other, even if he repeated the exact same activities at two different times of the week. It is suggested to improve the understanding of all the phenomena that could affect the electrodermal activity measurements before trying to draw conclusions related to the students’ cognitive engagement in the classroom.

Sensor-Based Prediction of Mental Effort during Learning from Physiological Data: A Longitudinal Case Study

“…Wearables such as the Empatica E4 have been used to measure electrodermal activity, skin temperature, and heart rate [11]. Some have linked wearable data to engagement [17][18][19], while others have used the data as indexes or measures of mental effort, mental workload, or cognitive load (for the purposes of this paper, we use these terms synonymously, referring to the load placed on the working memory during a task). Based on existing research, we believe that mental effort or mental workload is an appropriate correlate [1] to physiological activity, because mental workload demands can cause an unconscious, automatic response by the autonomic nervous system [20].…”

“…The second prominent challenge researchers face is how to best analyze the data from wearables in educational contexts. Some researchers have analyzed physiological variables individually, such as heart rate, electrodermal activity, or skin temperature [18,24,25]. However, a recent systematic review revealed that there was no single physiological measure that captured all aspects of mental workload measurement [21].…”

“…Meanwhile, [11] used a multimodal approach and found that using a random forest classifier resulted in strong classification accuracy when classifying students as being engaged in activities that required different levels of mental workload (area under the receiver operator characteristic (ROC) curve (AUC) = 0.93-0.99, F1 = 0.85-0.94). While unimodal or multimodal data analysis is a concern, the analysis techniques that are best suited for this type of work have also been explored, ranging from descriptive [18] through machine-learning methods [11,27]. At the present time there is not enough evidence to say which machine-learning algorithm works best for predicting mental workload.…”