by Jack White-Foy, Dulwich College
Neuroscience encompasses every aspect of brain and behaviour, from tiny molecular interactions to exchanges between groups of whole organisms. Educational neuroscience is the relatively new field of applying our understanding of development, cognition and behaviour to teaching and learning. This project set out to introduce students to cutting-edge Brain-Computer-Interface (BCI) technology through wearable, wireless EEG headsets. Students were encouraged to design and run their own EEG research projects through a school society. They were then asked to reflect on using such technology in schools and their views on the usefulness of neuroscience for students and teachers. The results show that students as young as 12 years old can successfully learn how to set up and carry out small-scale research projects. Whilst they struggled to take full advantage of the technology, they have grasped a rudimentary understanding of the relationship between recorded electrical activity and associated brain function. Additional findings related to the practicality and benefit of setting up such projects both to teachers and students themselves. Limited value was attached to the findings of EEG activity, which they ranked low to both students and teachers. Teachers’ enthusiasm and subject knowledge were ranked highest.
In an unfortunately over-cited article from 1997, John Bruer discussed the aggressive progress in neuroscience and its appeal to educators. Whilst full of praise for the technological achievements and advancement of knowledge, Bruer’s article is best known for stating that educational neuroscience was a “bridge too far” and that neuroscience was of little value to a classroom teacher (Bruer, 1997). Bruer went on to state that this was due to a lack of knowledge and that with more information, perhaps a bridge could be built. More than 20 years later and educational neuroscience has shown no signs of slowing down. We know far more about the development of brain structures and individual differences in cognitive abilities and many traditional ideas around the brain have been proven to be myths. Despite this progress, there is still not much to show for it in terms of direct impact on teaching practice (Thomas et al, 2019). More troubling has been the prevalence of neuro-myths that are still lingering today, which are often taken seriously because they are loosely based on genuine research findings (Grospietsch & Mayer, 2019). The repercussions of teachers adopting strategies based on neuro-myths has been well documented and includes wasted resources, time and even impaired student attainment (Dekker et al, 2012). More research clearly needs to be done in neuroscience and education before the two can be bridged effectively. The field is enormous and the organ at the centre of it is the most complex system in the universe (Kaku, 2014). It is of little wonder then that it might take some more time to unlock the brain and translate how it works into a language that is understood by educators.
A Google Scholar search for ‘neuroscience+research+education’ returned nearly 2.5 million results (June 2019). However, a search for ‘neuroscience+research+classroom-based’ returned just 6,590 results. Of particular interest to me was conducting neuroimaging-based research in real classroom settings (1,380 results on Google Scholar). With limitations notwithstanding, this could provide a unique setting away from a laboratory experiment and would hopefully generate ecologically valid results, which would be more relevant to practising teachers. I also wanted to carry out this research with students acting as assistants. The hope was that they would learn simple concepts in neuroimaging as a platform for developing research skills. With training in neuroscience to help them access its principles, I was also interested in the value they thought the discipline could offer to teaching and learning.
Zhang et al (2019) investigated the extent to which school-aged children could learn how to use EEG technology. Their aim was to identify factors that affect the efficacy of Brain-Computer-Interface use with severely disabled children. Zhang et al argued that the bulk of research into such EEG technology has involved adults, despite the technology being of most benefit to children. The study found that children were able to effectively learn how to use the BCI technology to manipulate virtual environments. The study did not explore whether their participants were familiar with setting up the devices or with the neuroscience principles behind the technology. The educational neuroscience programme described in this report sought to investigate this. The programme was created with the following aims:
- Short-term: set up a new school society, inviting all students aged 11 to 18 to learn how to use the EEG devices. Then assess whether students were interested in and able to learn how to research neuroscience in education. Finally, ask the students how relevant they saw neuroscience to teaching and learning based on their first-hand experience.
- Long-term: build a network of students and teachers across disciplines and partnership schools to research neuroscience in and out of classrooms.
The current report reviews progress made against the short-term goal.
THE NEURO LAB
Neuroimaging technology has revolutionised the field of neuroscience (Cacioppo, 2008). From its humblest beginnings of recording static structural images, the latest scans can show beautiful functional videos of the active brain. The very latest developments in ultra-high speed microscopy, called mesoscale selective plane illumination microscopy (mesoSPIMS), has even been able to image individual 3D neural networks (Voigt et al, 2019). The two main technologies have been functional magnetic resonance imaging (fMRI) and electroencephalography (EEG). Whilst EEG has superior temporal resolution (≈.05s with EEG compared with ≈1.0s on MRI), EEG is limited to measuring event-related potentials at the cortex and as such cannot provide data on subcortical (deeper brain) activation (Aue et al, 2009).
One analogy commonly used to compare fMRI and EEG is that of a sports stadium packed with spectators. With fMRI, the spatial resolution is such that you could record a single conversation between two individuals but with a delay. With EEG, the recording would be equivalent to lowering a single microphone into the middle of the stadium; there would be virtually no delay. The sound would be a combination of thousands of voices all at once with no way of distinguishing one individual. With such limitations, EEG is perhaps better suited to recording general brain wave patterns, rather than trying to identify neural activity in discrete brain regions (see Figure 1). One such example is gamma waves, which are high frequency and are associated with high-level cognitive tasks such as problem-solving (Roohi-Azizi et al, 2017). Delta waves are low frequency and are associated with coordinating large regions of neural activity moving front to back (prefrontal cortex to occipital lobe). This frequency of wave is typically observed during periods of deep sleep (see Figure 1) and is associated with consolidation of memory. Ideally, fMRI would be the imaging tool of choice but with costs in the hundreds of thousands of pounds, it is accessible only to the most advanced of university research laboratories. Even EEG, the cheaper alternative can exceed tens of thousands of pounds. It is no wonder then that neuroscience research has had trouble entering classrooms. This is about to change.
In 2015, the company Emotiv released its first version of the Insight, a wireless EEG headset with 5 channels and costs just £200. The advanced model, the EPOC+, costs £600 and has 14 channels (see Figure 2). A recent study at the University of Macquarie in Australia compared the EPOC+ with university research-grade EEG hardware. It was determined that the EPOC+ was equal in performance (Kotowski et al, 2018).
Whilst considerably cheaper than a university EEG, the devices produced by Emotiv are still not insignificant for school budgets. Fortunately, in June 2018 my own department agreed to purchase a single Insight device. By September 2018, the headset arrived and in October 2018, The Neuro Lab society was launched with 20 members from Year 7 to 13 (11 to 18 years old). At this time, I was still learning how to set up the headset and use the software. The advantage of this was that I was not much further ahead of the society’s new members. This gave us the genuine situation where we were all learning together. My input was limited to an initial idea for a project, which was based on a video I had seen by IBM to demonstrate their new cloud service. A programmer had connected an Emotiv Insight to a Raspberry Pi (a credit card sized computer) which translated the brain pattern from the Insight into commands to control a remote-controlled Star Wars toy (the BB-8). A diagram of the setup is in Figure 3.
This project at first appeared to be a challenging and interactive task for the new society to complete. Instructions were available online and the hardware was simple to obtain. The Neuro Lab students developed the idea to include a challenge, so that cognitive performance could be measured. The task was to control the route of the BB-8 device through a maze. The time to complete the maze and the number of errors (e.g. hitting walls, moving in the wrong direction) would be recorded and used to track improvement with practice. The design of the project is shown in Figure 4 below.
In practice, the integration of the different software packages and learning the different programming languages meant this project had to be put on hold. An alternative approach to The Neuro Lab was needed to adjust the difficulty level to be more appropriate to the members. The students chose to focus on developing confidence with the following software, developed by Emotiv:
1) EmotivBCI: this is the main software suite that displays live electrode activity translated as mood indicators (see Figure 5). This uses the award-winning algorithm designed by Tan Lee, the founder of the company.
2) The software also contains a Brain-Computer Interface environment, whereby the programme is trained to recognise distinct brain patterns and translate them into specific commands. These commands are carried out by a virtual cube (see Figure 6). Commands include moving the cube up and down as well as growing and shrinking in size.
3) 3D Brain Visualizer: This software produces a live 3D model of the brain and maps regional activation. The colours correspond to wave frequencies (see Figure 7 below and Figure 1 above).
The Neuro Lab sessions were entirely student-led, with each learning at their own pace and developing their own project ideas. Due to the enthusiasm of the members, I increased the sessions to lunchtimes every Monday, Wednesday and Friday. This gave the members the opportunity to maintain a steadier rate of progress. Figure 8 shows photographs taken during real sessions in The Neuro Lab.
EXAMPLE PROJECT: DULWICH CREATIVE
During a special curriculum enrichment week at the College in Summer 2019, The Neuro Lab wanted to get involved and celebrate the theme: creativity. Members chose to observe neural activity (using the visualiser software) with the creative task of unguided drawing; no parameters or goals were set as to what or how to draw. Figure 9 shows the setup and a screenshot from the video that was produced.
EVALUATING NEUROSCIENCE IN EDUCATION
The potential benefits of increased understanding in neuroscience have been debated for decades and will likely continue to be for years to come. What is arguably true is that greater knowledge is always welcome. The aim of this project was to see how students would respond to opportunities to study neuroscience in school and whether it was too advanced. I was also interested in what they thought of the potential applications for them as learners. Near the end of the 2018/2019 academic year, I asked members of The Neuro Lab to complete a short online questionnaire. They completed it either in a Neuro Lab session, or in their own time.
The survey recorded respondents’ names to ensure no duplicates were submitted and to allow me to know who to send reminder requests to. Students were assured that their names would be removed and replaced with a numerical ID, that only I would have access to. Students were informed that taking part was entirely voluntary and that they could remove their data at any time.
The questions were organised into four sections:
- Familiarity with the neuroimaging headsets
- Knowledge and understanding of neuroscience
- Views on applications of neuroscience in education
- Views on teaching and learning priorities for teachers and for students.
For section one, the first question asked respondents if they felt confident carrying out various tasks. They could choose as many as they wished. This was followed by open qualifier questions that challenged their responses to the previous check-box question, by asking for technical responses to scenarios (e.g. “How could you improve signal quality from an electrode on the EPOC+?”). For all sections, opinions were sought using Likert-style questions to give respondents greater choice and to help differentiate between opinions. Eight students (aged 11-13) answered the questionnaire. All responses are in Appendix 1 with highlights described below.
Figure 10 summarises responses to the question on confidence using the technology. All respondents indicated they were very confident with setting up and using the devices (options 1-6). Fewer responses stated they were confident using the software and interpreting the results (option 7). In the follow-up questions, all students were able to give correct solutions to technical problems. Their knowledge of which brain structures could be measured with the headsets varied. Half of students had misconceptions, thinking that EEG can measure subcortical structures, which it cannot.
Question 9 in the survey asked students to consider how challenging certain aspects of The Neuro Lab were (see Figure 11). There was considerable variation in response, which reveals the differences in students’ strengths. Some members were full of ideas for projects, whilst others found this challenging. Interestingly, the question that drew the greatest range of results was regarding setting up the headsets, which will be discussed later.
Questions 10, 12 and 13 interrogated students’ views on the value on neuroscience in schools. Specifically, they were asked to rate suggested applications of the field and then rank teaching and learning priorities for students and for teachers (see Figures 12, 13a and 13b).
The majority of responses were positive when rating the applications of neuroscience in schools. This contrasted with where they ranked neuroscience findings, specifically brain region activation with associated tasks, in priorities for students (5th out of 9) and teachers (5th out of 6). The most highly ranked factor was a teacher’s subject knowledge and enthusiasm. The factor ranked last for both student and teacher priorities was knowledge about what makes students forget information.
The students involved with The Neuro Lab have learned a great deal of technical knowledge and skill. They are confident using the equipment and have made considerable progress in terms of their understanding of the applications of the technology as a Brain-Computer-Interface and a way of observing brain activity. They are still developing their skills in creativity for research ideas and there is a strong desire to learn more about cognitive neuroscience. As a subject, it is clear that neuroscience is not necessarily too advanced for school-age students and instead has been a surprising motivator. It has provided a challenging and interesting off-curriculum avenue for all students, from the strongest science fanatic to the weakest non-academic. It is possible that the freedom to design, test and abandon research ideas in a low-stake environment has given some students greater confidence where they might otherwise be lacking and the room to stretch for those already feeling comfortable with the pace of their formal studies.
The survey revealed that students felt confident describing how the technology could be used and they had a strong understanding and skill in this area. Interestingly, they also found this to be the most challenging aspect of The Neuro Lab. This was also my experience. Whilst the EEG technology and underlying algorithms are state-of-the-art, the wireless aspect still relied upon Bluetooth technology, which was the weakest link in the setup. By June 2019, we had established a reliable sequence of setting up and connecting the devices. In hindsight, the instructions provided by Emotiv were for individual units. The Neuro Lab was working with seven simultaneously. This presented considerable risk of interference from the different Bluetooth connections, which was arguably the main reason for frustration felt by members and the most significant drag factor in affecting progress made on projects. At times, it would take up to 40 minutes just to set up the devices, which left 20 minutes to use them. Since the society restarted in September 2019, we have consistently set up all seven devices in 5 to 10 minutes by following the sequence established by members.
A surprising result from the survey was the factor ranked lowest as a priority for teachers and students: ‘understanding forgetting’. Knowing how we learn and how we forget are arguably of comparable significance, yet they were ranked quite differently by the students. This highlights the limitations of students’ grasp of pedagogy and cognition. They are not experts in teaching and learning. Teachers are experts, as a result of years of training and experience. To expect a student to have a similar understanding of and appreciation for the usefulness of a field of research, is unrealistic. Whilst valid, caution should be taken when evaluating any model, intervention or policy using student opinion. In hindsight, I would have provided students with relatable examples of each factor in the survey.
The teenagers in this project did learn how to use neuroimaging technology and are interested in neuroscience. They have yet to explore neuroscience principles in-depth, which goes some way to explain why they placed neuroscience so low in their priorities for teachers and students whilst at the same time were positive about the potential applications of neuroscience in schools. To some degree, this reflects the current state of play of neuroscience in education. There has been a growth in enthusiasm for learning more about the teenage brain and the implications of the research. Whilst this hunger has grown, the real-world applications of neuroscience for teachers and students has lagged. In the past five years, I have attended half a dozen specialist conferences, talks and workshops on neuroscience in education. At each of these events, they have ended with the same questions: what are the practical lessons for teachers to take from neuroscience and how can we bridge the divide between researchers and teachers? These questions were followed by a ten-minute brain-storming session where attendees were asked to come up with ideas. Suggestions were vague, unrealistic and narrowly focused, lacking in experience on both sides. The teachers had insufficient experience of the research and most of the researchers had never set foot in a classroom.
There is a real risk that in the rush to satisfy educators’ hunger for neuroscience, there will be a return of the neuromyths that resulted in wasteful and even damaging policies in schools (e.g. VAK learning styles, Brain Gym, hemispheric learners and critical periods). I am a strong advocate for educational neuroscience. There are only advantages to be gained from greater knowledge of how the brain develops throughout childhood and adolescence and the relationship with academic performance. Obtaining this knowledge takes time and careful research design because accessing the brain in valid settings is challenging. Wireless EEG offers one method we can use to achieve this, but we must remember the limitations of the technology and the considerable differences between children. Above all, we need to be patient.
In the current academic year 2019/2020, the students are now ready to use the technology with confidence. It is our intention to begin more detailed learning of the theories of neuroscience to give more context to the data we collect and to feed the students’ creativity. To start us off, I have been in contact with a consultant in pain management who has experience in using EEGs to monitor cortical activity in patients during surgery. This activity is used to measure the level of consciousness in the patient during surgery and allows the anaesthesiologist to adjust the anaesthetic accordingly. The consultant has offered to visit The Neuro Lab to give the members a course in the neuroscience of pain and make links to the EEG technology with which they are familiar. I suspect that a greater level of theoretical knowledge will not only challenge the students, it will help to inspire them all to design their own projects, an area in which a large number struggled.
When discussing projects for 2019/2020, The Neuro Lab members decided they would still like to work on the BB-8 project. I contacted the Head of Developer Marketing Content Strategy at IBM, the original designer of the BB-8 Brain-Computer Interface. He has since been in touch and has agreed to help The Neuro Lab work on the project. With the fundamentals of the technology firmly under their belts, the society’s members have already been making good progress. They have been opening up the Python and Scratch coding languages of the devices and are discussing ways in which the findings could be used: providing alternative communication methods for those living with paralysis and tools for improving cognitive abilities in children with special educational needs. There is a long way to go but we are on the right road.
Aue, T., Lavelle, L. A., & Cacioppo, J. T. (2009). Great expectations: What can fMRI research tell us about psychological phenomena? International Journal of Psychophysiology, 73(1) 10-16.
Bruer, J. T. (1997). Education and the brain: a bridge too far. Educational Researcher, 26(8), 1–13.
Cacioppo, J. T., Berntson, G. G., & Nusbaum, H. C. (2008). Neuroimaging as a New Tool in the Toolbox of Psychological Science. Current Directions in Psychological Science, 17(2),62-67.
Dekker, S., Lee, N. C., Howard-Jones, P., & Jolles, J. (2012). Neuromyths in Education: Prevalence and Predictors of Misconceptions among Teachers. Frontiers in psychology, (3), 429.
Grospietsch, F., & Mayer, J. (2019). Pre-service Science Teachers’ Neuroscience Literacy: Neuromyths and a Professional Understanding of Learning and Memory. Frontiers in human neuroscience, 13, 20.
Kaku, M. (2014). Future of the Mind. DoubleDay: New York.
Zhang, J., Jadavji, Z., Zewdie, E., & Kirton, A. (2019). Evaluating if children can use simple brain computer interfaces. Frontiers in Human Neuroscience, 13, 1-7.
Kotowski, K., Stapor, K., Leski, J. & Kotas, M. (2018). Validation of Emotiv EPOC+ for extracting ERP correlates of emotional face processing. Biocybernetics and Biomedical Engineering 286, 1-9.
Roohi-Azizi, M., Azimi, L., Heysieattalab, S., & Aamidfar, M. (2017). Changes of the brain’s bioelectrical activity in cognition, consciousness, and some mental disorders. Medical journal of the Islamic Republic of Iran, 31, 53.
Thomas, M., Ansari, D., & Knowland, V. (2019). Annual Research Review: Educational neuroscience: progress and prospects. Journal of child psychology and psychiatry, and allied disciplines, 60(4), 477–492.
Voigt, F.F et al (2019). The mesoSPIM initiative: open-source light-sheet microscopes for imaging cleared tissue. Nature Methods 2019.