The Future of Learning: Utilization of Thermal Imaging to Measure Cognitive Load

Learning vs. Cognition

Learning is the process of acquiring knowledge or ability. Typically learning is done through experience, instruction, or study and practice. Although it is natural to think about the brain when considering learning, ironically not much consideration is given to cognition when instruction and study are constructed. Cognition is central to learning, as it is the way that the brain collects and organizes information and ultimately determines how much an individual can acquire and retain.

Assessing Cognitive Load in Educational Research

Before we can use our understanding of cognition to redesign teaching practices, we must first measure what’s known as cognitive load: the mental resources required to complete a task. Current research depends on self-reported scales, a subjective measure of cognitive load. At Lab 58, we are investigating the use of thermal imaging cameras to measure cognitive load based on skin temperature.  Thermal imaging should be at the forefront of educational research. It gives us clear advantage with regard to measuring, estimating, and even anticipating cognitive load.

Cognitive Load Theory: Free Up Space to Learn More

The concept of cognitive load surrounds the appraised level of complexity or level of information processing resources that are mandatory for engagement in a particular learning experience. In other words, cognitive load is all about how difficult a task is and how many of our brains’ resources are required to complete the task. Cognitive load is connected to working memory, essentially the brain’s short-term memory bank. Working memory comes equipped with a limited capacity that varies from individual to individual. Once information enters our working memory, there is a short time before it is either discarded or begins its journey into long-term memory storage.

Cognitive load theory states that because our brain, particularly our working memory, is limited in capacity, we must be intentional about the way that we structure learning tasks and experiences so they are designed to reduce working memory load. Reducing working memory load encourages the production of new schema that will free up pathways to store new information.

Measuring Cognitive Load

Determining the most effective ways to measure cognitive load is significant. When cognitive load can be measured in a scalable way across varying tasks and people, knowledge and instruction can be shaped to encourage more constructive learning and comprehension. Measuring cognitive load has been a continuing challenge in experimental and educational research, partly because there are numerous ways to measure cognitive load, each with their own set of limitations. The two primary types of measures include self-report dual-task, and physiological parameters.

The Likert Scale

There are multiple self-report measures used in experimental research; however, the most widely used scale for measuring cognitive load is the rating scale developed by Paas et al. (1992). This Likert scale was designed for the purpose of measuring the difficulty of a particular learning task. Although this scale is widely used in research due to its low economic cost, it has its own set of limitations. Aside from the difficulty in distinguishing between true variance and measurement error, many researchers are deterred by the inconsistent wording, labels, and scale ranges. Additionally, like many self-reported scales, measures that require introspection from participants are typically unreliable, as the responses are subjective in nature. Often, participants may alter their responses based on situational attributes; as a result, their responses may not provide an adequate measure.

Dual-Task Measures

Dual-task measures are objective measures that require an individual to participate in two tasks simultaneously. The guiding assumption of this type of measure is that performance on a secondary task diminishes as the initial task becomes more and more complex. This assumption is acceptable in theory; however in practice, there are numerous disadvantages to this type of measure. The most salient limitation is that everyone has a unique working memory capacity. As a result, the complexity of a task and the utilization of cognitive resources among learners will differ as well.

Another disadvantage of dual-task measures is that it is difficult to determine whether the measured load is intrinsic, extraneous, or germane. Intrinsic load refers to the general difficulty of the learning task being performed. It is typically static, and unable to be adjusted by the instructional designer. Extraneous load has little to do with the learning task at hand. This type of measurement is also intrusive and therefore disrupts the natural learning process, adding to the amount of cognitive load experienced by the participant. Germane load refers to the mental resources allocated to programming schema into the brain’s long-term memory stores.

Measuring Cognitive Load via Physiological Parameters

Cognitive load can also be evaluated using measures of physiological parameters, as a wide range of them have been used as indicators of cognitive load. Some of these measures include heart rate, electroencephalography, hormone levels, fMRI, and pupil dilation. Pupillometry, or pupil dilation, is the most frequently evaluated physiological indicator. Individual differences in pupil diameter and blink rate represent noninvasive physiological indicators of attentional engagement. Typically, increased blink rate and dilation are correlated with greater cognitive load. Physiological assessments provide an objective and effective method of measurement, as they do not rely on the introspection of the learner.

Using Thermal Imaging to Measure Cognitive Load

The utilization of thermal imaging to measure cognitive load is a functional and efficacious method. Research, such as the study conducted by Abdelrahman et al. (2017), has proven thermal imaging to be a reliable, low-cost, unobtrusive, easy-to-use, objective cognitive load measurement tool that can change the face of instructional design and learning as we know it. Advances in thermal imaging technology have resulted in the availability of high-fidelity cameras that are priced below $500.00. Thermal cameras work by detecting light waves that are untraceable by the human eye and using those light waves to construct heat maps, or temperature profiles, of the object being recorded by the camera. In Figure 1, a computer vision algorithm has placed temperature markers—yellow dots—at distinctive points that are used to calculate cognitive load output measurements.

The process performed by the camera is so beneficial to cognitive load measurement because research has associated changes in skin temperature with changes in cognitive load. The thermal imaging process begins with the user engaging in a learning experience, such as a math activity or a logic-related task, while being recorded by the thermographic camera. The recording process is unobtrusive as the thermographic camera is either completely invisible, hidden within the computer, or so small that it is nearly unnoticeable to the subject. After completing and saving the video recording, the recorder then analyzes the video frame by frame to create a thermal heat map.  

Thermal Imaging in Educational Research

As previously stated, thermal imaging should be at the forefront of educational research due to the clear advantages it provides with regard to measuring, estimating, and even anticipating cognitive load. Not only that, but this advantage extends to working memory capacity, which is a fundamental aspect of learning. This method has been tested and proven numerous times through experimental research. In a 2015 study, Pinti and associates used thermal infrared imaging in conjunction with functional Near Infrared-Spectroscopy (fNIRS) to reveal autonomic correlates of prefrontal cortex activity. Essentially, the researchers measured changes in the prefrontal cortex’s hemoglobin levels using fNIRS, and they utilized thermal infrared imaging to measure changes in cognitive load during a learning task.

During this study, researchers attained significant evidence that thermal infrared imaging is an effective evaluation of the autonomic nervous system (ANS) response to the allocated learning experience. Due to the previously discussed link between skin temperature and the ANS, as a whole, researchers found thermal imaging to be an adequate evaluation of peripheral systemic behaviors, such as heart rate and blood pressure, which tend to fluctuate during participation in a cognitive task. As a task grows more complex, many experience increased blood pressure, accelerated heart rate, sweating, and increased cortisol levels. Measuring these behaviors with thermal imaging allows us to gain perspective on the load being placed upon the very limited working memory.  

In the 2017 study, to improve the process of creating novel experiences and increase cognition associated with digital systems, Abdelrahman and associates used a commercial thermal camera to observe the research participants’ nose and forehead changes in temperature to evaluate cognitive load. These researchers found that participation in complex tasks causes changes to facial temperature that can produce an approximate cognitive load measure when using a thermal camera. Additionally, another successful study on patients with psoriasis utilized thermal imaging to detect and evaluate the activation and deactivation of the sympathetic nervous system.

RTI's Role in Advancing the Use of Thermal Imaging to Measure Cognitive Load

More research and attention should be given to utilizing thermal imaging to measure cognitive load. RTI has joined in on innovation and development of this measurement tool by helping reduce the once high cost of thermal imaging technology. In 2018, RTI launched a company that commercialized and assisted in mass-producing the thermal camera. With the help of companies like RTI, the thermal camera will continue to evolve into a more feasible solution for measuring task difficulty. As the cameras continue to decrease in cost and become smaller and less detectable, they will yield even more accurate results and allow us to structure instructional design to work with the brain and working memory instead of against them.