Use of Heart Rate data obtained from wearable devices to monitor and assess student cognitive engagement in university computer programming lectures

Author(s):

Aidan McGowan(presenting / submitting)Philip Hanna Des Greer John Busch Neil Anderson Matthew Collins

Conference:

ECER 2017

Network:

11. Educational Improvement and Quality Assurance

Format:

Paper

Session Information

11 SES 10 B, Improving Schools by Using School Knowledge

Paper Session

Time:

2017-08-24

15:30-17:00

Room:

W2.05

Chair:

Heidi Flavian

Contribution

There have been substantial amounts of study over a lengthy period of time into the variability of heart rate (HR) in relation to cognitive activity. HR is the peripheral measure most used to assay affect and cognition (Guerra, 2015). As far back as the 1970s a series of experiments by Lacey (1970) and Lacey (1974) demonstrated that tasks requiring increased cognitive processing are associated with HR acceleration. Numerous clinical experiments have been conducted measuring HR and cognitive tasks, including a reported increase in HR for computer gamers performing complex gaming tasks and by subjects performing difficult mental arithmetic (Turner & Carroll, 1985). Abundant real world occupation studies have also reported similar HR increases due to increased cognitive task load, such as for air traffic controllers (Wilson & Eggemeier, 1991), fighter pilots (Wilson, 1993) and university lecturers (Filaire, 2010). Increased memory load (number of items) was shown to be accompanied by accelerated HR (Backs & Selijos, 1994 : Pearson & Freeman 1991). Indeed Cranford’s study (2014) directly linked HR to varying degrees of cognitive load in problem solving and concluded that HR monitoring has further significant potential use in measuring cognitive load during the learning process. Scholey (1999) reasoned that these observed increases in HR during cognitive processing are the body’s facilitation of the delivery of metabolic substrates to the brain that are then utilised by neural mechanisms underpinning cognitive performance.

However there have been a very limited number of studies in the use of HR as a measure of student cognitive engagement in university lectures. Bligh carried out a series of classroom lecture studies showing that student HR decreased over the course of a 50-minute lecture (Bligh 1972, 1998, 2000). The decline in HR was interpreted as a measure of decreasing arousal, which Bligh considered as one component of cognitive engagement. In addition, Bligh reported a single event where a question from a student resulted in an elevation of HR in other students. Darnell and King (2014) expanded on this work, and concurred with Bligh that there appears to be a decrease in average HR across a 50 minute lecture class and a temporary increase in HR in response to student questions. In addition, they concluded that pair-share sessions resulted in elevated average HR.

The devices used in most HR cognitive studies were generally expensive and obtrusive. Consequently most of the studies suffered from small sample sizes and limited sampling points. The prominent nature of the HR measuring device likely also affected the results with the students acutely aware throughout the experiment that their HR was being sampled. Indeed Anttonen and Surakka (2005) pointed out the need for new methods for inconspicuous heart measurement.

The recent proliferation of accurate, cheap and unobtrusive wearable devices with biometric sensors presents a new opportunity to perform a relatively inexpensive, natural, large scale study on the biometric effects on students during a series of lectures.

Objectives - stages

The initial objective is to use wearable devices (Microsoft Band 2) to measure and record heart rate of a large representative number of students during a number of lectures and analyse the data to seek to identify any general patterns in HR during the length of the lecture.
To analyse the data to seek to identify and potentially link repeatable HR patterns to the cognitive activities of students during the lecture.
To relate the analysis of the data to the relative effectiveness of the various interactive and non-interactive teaching methods employed during the lectures to potentially increase future student cognitive engagement and lecturer performance with the aim to increase overall student attainment.

Method

The study was conducted with a cohort of post graduate students taking a 12 week compulsory module in Java programming in semester two of a one year MSc course in Software Development. The research purpose and the methodology to be employed were both explained to all the students before the course and volunteers were requested. A large number expressed an interest and willingness to participate and subsequently ten students were chosen at random. Similar to other studies in the area any students that had diagnosed cardiovascular defects and smokers were not included. Each student was given a Microsoft Band 2 wearable device and encouraged to wear it regularly. This was to ensure that a baseline resting heart rate could be established for each student and also to lessen the potential influence on the results that measuring the HR during the lecture may have had. After each lecture each student in the study uploaded their HR data to a secure central server which was accessible to the researchers. A total of 20 individual 50 minute lectures were included in the study. Each lecture was fully video recorded and the lecturer’s HR was also measured throughout the lecture. The HR for students and lecturer was sampled per second. The initial analysis involved a linear analysis of each individual student and then group HR per lecture and subsequently for all lectures. A comparison was made with the known resting HR for each individual. In order to establish if indeed themed cognitive activities have an influence on HR each lecture video was independently transcribed and sectioned by time into common teaching formats and lecture events. These included teacher lead activities such as teaching theory, code demonstrations, live code development, questioning students and student lead activities such as individual or paired exercises, group discussions, peer demonstrations and student questions to the lecturer.

Expected Outcomes

The initial results have shown that similar to the findings of Bligh (2000) and Darnell and King (2014) that there was in most cases an overall decline in HR of students over the duration of a 50 minute lecture. The same is true for the lecturer. Preliminary results have also illustrated that there are identifiable patterns of increased HR that may be associated with cognitive activity. As such it may well be possible to better understand various teaching formats and their effect on cognitive performance. On that basis it also may well be possible to fine tune teaching formats and delivery mechanisms to optimise student cognitive performance during lectures. There appears to be some obvious lecture events that cause a student’s heart rate to increase, including peer demonstrations, lecturer questions and individual or group activities. It is also notable while most HR’s tend to decrease throughout the lecture that there are many significant spikes and drop-offs in the rates. It is of interest to further investigate these variances and potentially identify if the cause is due to cognitive or lack of cognitive engagement. While the focus of this study is on the lecture delivery of a computing programming module the researchers feel that there are many areas of commonality that will be of interest and comparable to other university subjects and their delivery.

References

Anttonen, J., Surakka, V. (2005). Emotions and heart rate while sitting on a chair. CHI '05 Proceedings of the SIGCHI Conference on Human Factors in Computing Systems Pages 491-499. Portland, Oregon, USA — April 02 - 07, 2005 ACM New York, NY, USA 2005 Backs, R. Selijos, K. (1994). Metabolic and cardiorespiratory measures of mental effort: The effects of level of difficulty in a working memory task. Int. J. Psychophysiol., 16 (1994), pp. 57–68 Bligh DA. What’s the Use of Lectures? Jossey-Bass Publishers, SF (2000). Or Intellect Books (1998). Originally published in 1972. Boucsein, W.: Electrodermal activity. Plenum Press, New York (1992) Cranford, K., Tiettmeyer, J., Chuprinko, B., Jordan, S., Grove, N. (2014). Measuring Load on Working Memory: The Use of Heart Rate as a Means of Measuring Chemistry Students’ Cognitive Load. J. Chem. Educ., 2014, 91 (5), pp 641–647 DOI: 10.1021/ed400576n Darnell, D., Krieg, P. (2014) Use of heart rate monitors to assess student engagement in lecture. The FASEB Journal vol. 28 no. 1 Supplement 721.25. Filaire, E., Portier, H., Massart, A. Ramat, L., Teixeira, A. (2010) Effect of lecturing to 200 students on heart rate variability and alpha-amylase activity. European Journal of Applied Physiology March 2010, Volume 108, Issue 5, pp 1035–1043 Lacey, J., Obrist, B., Black, P, Brener, A., DiCara, L. (1974). Studies of heartrate and other bodily processes in sensorimotor behaviour. Cardiovascular psychophysiology, Aldine, Chicago Lacey, J, Lacey, B., Black, P. (1970). Some autonomic-central nervous system interrelationships Physiological correlates of emotion, Academic Press, New York Lieberman, H., Farina, E., Caldwell, J., Williams, K., Thompson, L., Niro, P., Grohmann, K. (2016) Cognitive function, stress hormones, heart rate and nutritional status during simulated captivity in military survival training, Physiology & Behavior, Volume 165, 15 October 2016, Pages 86-97, ISSN 0031-9384, http://dx.doi.org/10.1016/j.physbeh.2016.06.037. Guerra, P., Sánchez-Adam A., Miccoli L., Polich J., Vila J. (2015) Heart rate and P300: Integrating peripheral and central indices of cognitive processing. Int J Psychophysiol. 2016 Feb;100:1-11. doi: 10.1016/j.ijpsycho.2015.12.008. Epub 2015 Dec 23. Pearson, G., Freeman, F (1991) Effects of extraversion and mental arithmetic on heart-rate reactivity. Perceptual and Motor Skills, 1991, 72, 1239-1248. Perceptual and Motor Skills 1991. DOI:10.2466/pms.1991.72.3c.1239 Shi, Y., Ruiz, N., Taib, R., Choi, E., & Chen, Galvanic skin response (GSR) as an index of cognitive load. In CHI '07 extended abstracts on Human factors in computing systems (2007). ACM, New York, NY, USA, 2651-2656.