Opinion

How Eye Tracking Technology Contributes to the Development of Adaptive E-Learning Systems

Zhanni Luo*

School of Educational Studies and Leadership, University of Canterbury, New Zealand

Submission: December 19, 2019;Published: January 07, 2020

*Corresponding author: Zhanni Luo, School of Educational Studies and Leadership, University of Canterbury, New Zealand

How to cite this article:Zhanni Luo. How Eye Tracking Technology Contributes to the Development of Adaptive E-Learning Systems. JOJ Ophthalmol. 2020; 8(1): 555730. DOI: 10.19080/JOJO.2019.08.555730

Keywords:Eye-tracking; E-Learning; Education; Computer science; Adaptive e-learning; Digitalization; Environments; Hook up; Categorize; Code; Match; Electroencephalograms

Opinion

Studies in Ophthalmology mainly focus on the treatment of disorders and diseases of the eye. Researchers also study eye movements collected by eye trackers to develop a means to prevent, diagnose and treat abnormalities or ocular disease in clinical situations. I propose that eye-tracking technologies not only serve clinical treatment but also contribute to the development of adaptive e-learning systems in the field of education and computer science.

Adaptive e-learning system refers to an educational system that provides customized learning materials to different learners to address each individual’s unique needs or preferences. In the era of digitalization, online learning enables students to study at any place at any time, as well as selecting any material they prefer. As learners are normally exposed to self-paced environments in which face-to-face tutoring is normally absent, the connection between learners and learning has been significantly reduced.

Accordingly, learners are easy to be distracted or unengaged in educational activities. Therefore, there is a call for adaptive e-learning systems, which “hook up” learners’ attention by providing adaptive learning materials that are preferred by the learners.

To establish an effective adaptive e-learning system, designers need to complete works of three steps. First, observe users’ behaviour patterns and preferences, summarize the commonality and therefore categorize users into groups (Step 1: Categorize). Normally this step is established based on literature review. Second, prepare adaptive materials and pair them with behaviour patterns (Step 2: Code). At this step, designers need to write codes in servers and arrange materials in databases. Third, provide adaptive experience in e-learning systems. The e-learning system collects users’ behaviour patterns, match them with the materials in the databases, and then provide adaptive learning materials as prepared (Step 3: Match) (see Figure 1).

According to Figure 1, it is obvious that the establishment of adaptive e-learning system relies on the categorization of users’ behaviour patterns and preferences, which mainly based on literature (Step 1). However, since previous studies in the last two decades focused on technologies, researchers tend to build and test systems in Step 2 and 3 with the Step 1 untested. As to the users’ behaviour patterns and preferences, researchers either make casual and unjustified categorization or adapt from existing learning style theories (which categorize students into different groups based on differences on receiving and proceeding information). Studies of the latter type are based on the assumption that the learning style theory is rational in the way of categorizing people into a limited number of groups.

Unfortunately, the reliability and validity of learning style theories have been criticised for decades. One major weakness of learning style theory is it categorized millions of people into a limited number of groups. Normally authors make the categorization based on their personal observations and the interpretation of theories from their own perspectives, so the established learning-style frameworks are doubtful with regards to rationality, reliability and validity.

Then, why we involve learning style theories in the establishment of adaptive e-learning systems? The answer is simple: the learning style theories provide abundant information on people’s behavioural traits and preferences, which is rarely covered in other theories.

I propose to solve the problems by recording users’ biometric data and analyses the data with the use of machine learning. I suggest choosing eye movement since it is easier to collect compared with other forms of biometric data such as brain electricity collected by electroencephalograms (EEGs). The procedures could be simplified as follows: first, enlist hypotheses about users’ behaviour patterns and preferences based on the learning style theories. Since there are various learning style models, the number of hypotheses could be unlimited. Second, build a material pool in which the materials are potential to reveal eye-movement differences in users’ behaviour patterns and preferences. Then, collect eye-movement data, followed by testing hypotheses with the data. At this step, a number of learning style theories would be challenged while some others tested. After that, categorize users into groups with the help of machine learning. Machine learning is a field of computer science that studies algorithms and techniques to complete tasks without explicit instructions. Instead, it makes predictions or improvements based on patterns of the data and inference.

The use of eye-movement data and machine learning approach is with the following benefits. First, it avoids interpreting theories from a specific angle, as it includes hypotheses of various learning style frameworks. Then, it categorizes students’ into an unlimited number of groups based on authentic data (real-time eye movements and preferences), which significantly improves the reliability of the categorization. Meanwhile, it also provides adaptive learning materials without testing students’ strengths or weaknesses, which avoids negative impacts of stereotypes on students. In short, the proposed method makes good use of learning style theories while totally breaking the restrains.

Conclusion

To sum up, there is a lack of studies on the validated categorization of learners, which is the basis of adaptive e-learning systems (see Step 1 in Figure 1). While most previous studies build adaptive e-learning systems based on the long-criticized learning style theories, I propose to employ eye-tracking technology and machine learning. It is able to validate the learning style theories, redefine what individual differences are, and eventually provide a truly adaptive and customized learning experience. In this way, we could both make good use of and eliminate the restrain of learning style theories. Or, we can even redefine learner differences and adaptive e-learning systems.

It worth noticing that the eye-movement data often contain “a great deal of noise” caused by blinking, mind-wandering, inaccurate calibration, etc. What’s more, there is a long way to go in this inter-discipline direction, so researchers are suggested to treat it with caution.

Researchers with eye-tracking equipment could cooperate with scholars in the field of psychology, computer science and education. Future work could explore the accuracy of eyetracking technology in identifying different types of users in learning systems, the feasibility of machine learning in analysing eye movement data, and the establishment of adaptive e-learning systems based on eye movement data.