"Spontaneous" blink frequency serves as indicator of learning satisfaction

A research team including Graduate Student Mari Miya, Junior Associate Professor Yu Akaishi, and Professor Masanaga Yamawaki of the Department of Medical Education Research and Development, Graduate School of Medicine & Dentistry, Institute of Science Tokyo, has discovered that there is a correlation between the frequency of a learner's blinks during online clinical medicine lectures and their learning satisfaction. They quantified subtle facial movements through AI-based facial expression analysis. They found that when spontaneous blinking increases, learning satisfaction tends to rise.

This discovery is expected to be utilized for understanding learners' states and supporting learning. The findings were published in the journal BMC Medical Education on April 15. The knowledge that medical students need to learn is increasing exponentially, making the improvement of educational methods and quality a critical challenge.

It has been reported that instructors estimate the level of understanding of a lecture from the facial expressions of learners during the class, and learners also respond to non-verbal information, including the instructor's facial expressions. Non-verbal information has the potential to serve as a source of information for understanding a learner's psychological state and engagement (which consists of behavioral, emotional, and cognitive aspects), representing the learner's involvement in learning activities. On the other hand, it has not been clarified which specific non-verbal information should be focused on.

Therefore, this time, the research team focused on individual parts of facial expressions and verified the relationship between their subtle movements and lecture satisfaction. In 1978, Paul Ekman and his colleagues developed the Facial Action Coding System, which anatomically describes and classifies facial movements in units of muscles. They defined the minimum units of movement called AUs (Action Units), laying the foundation for facial expression analysis.

In this study, using software that can quantify AUs (OpenFace 2.0), they analyzed and quantified the facial expressions of participants during online clinical medicine lectures. As a result, they found that among the three types of blinking (spontaneous, reflexive, and voluntary), learning satisfaction increases when spontaneous blinking increases.

A significant correlation was recognized between the satisfaction level of the overall lecture and the measurement time of blinking (AU45). Regarding the movements of facial parts other than blinking, no significant correlation with satisfaction was recognized. It is known that spontaneous blinking increases or decreases during scenes of concentration or knowledge integration.

A learner's blinking can be observed in real time and can be utilized as a means for instructors to estimate the learner's engagement. Non-verbal information has the potential to serve as a valuable source of information for understanding the psychological state and engagement of learners. Akaishi said, "We advanced our research based on the hypothesis that 'true learning satisfaction might be hidden in non-verbal information during lectures,' and found that 'blinking' in particular becomes a useful indicator showing the psychological status of learners.

In the future, we hope to apply technologies that numericalize and quantify facial expressions, gaze, and blinking, which tended to be judged subjectively until now, to the field of medical education, and contribute to training high-quality medical professionals." Journal Information Publication: BMC Medical Education Title: Artificial intelligence analysis of facial movements and satisfaction in online medical lectures: a cross-sectional study DOI: 10.1186/s12909-026-09175-x Medicine This article has been translated by JST with permission from The Science News Ltd.

(https://sci-news.co.jp/).

Unauthorized reproduction of the article and photographs is prohibited. This article has been translated by JST with permission from The Science News Ltd. (https://sci-news.

co.jp/). Unauthorized reproduction of the article and photographs is prohibited.