Back to Blog

Can we use AI to understand what students are thinking during science lessons? Prof. Mei-Hung Chiu shares groundbreaking research on using facial expression analysis to detect students' cognitive and emotional states - and how this could transform science teaching.

A Research Journey

Prof. Chiu's research has evolved through multiple domains, each building on the last:

Student Science & Conceptual Change

How students construct knowledge and change their minds

Modeling-Based Instruction

Using models to support science understanding

Augmented & Virtual Reality

Immersive technologies for science education

Facial Expression Research

Reading students' cognitive states through their faces

AI in Science Education

Current focus: integrating AI to support learning

The Inspiration: "Lie to Me"

From TV to Classroom

The research was inspired by the American TV sitcom "Lie to Me." Prof. Chiu and colleagues wondered: if we can read deception through facial expressions, why not read students' understanding?

The Core Idea

If teachers could read students' minds through their facial expressions, they could provide more targeted guidance during learning. Facial expressions reveal cognitive and emotional states that words might not capture.

View Details

The research question: Can we use facial expression analysis to:

  • Detect when students are confused or surprised?
  • Identify moments of cognitive conflict (when reality contradicts expectations)?
  • Help teachers intervene at the right moments?
Original Quote

Mei-Hung Chiu: "So my colleagues and I watched the TV, and then we said, why don't we apply this facial expression in science education in our classroom? If the teachers could read the students' mind through their facial expressions, then maybe the teachers could provide some guidance to their learning. So that was the beginning of the story."

The Flask Experiment: Counterintuitive Science

The Demonstration

Heat water in a flask until it boils. Then flip the flask upside down and place ice on top. Ask students: What will happen next?

The Counterintuitive Result

Students expect the ice to melt and the boiling to stop. But something surprising happens: the water starts "boiling" again - even with ice on top. This creates cognitive conflict perfect for studying facial expressions.

View Details

What students expect:

  • Ice will melt
  • Water will cool down
  • Boiling will stop

What actually happens:

  • Bubbles appear inside the flask
  • Water appears to "boil" again
  • Ice is on top, but bubbles form below

The science: The ice cools the air in the flask, reducing pressure inside. This lowers the boiling point of water, causing it to boil at a lower temperature - a counterintuitive demonstration of vapor pressure principles.

This cognitive conflict - when reality contradicts expectation - creates visible facial expressions that can be analyzed.

Original Quote

Mei-Hung Chiu: "Because we put the ice cube on the top, and the water was boiling, so it's supposed to have the ice melting, right? But in fact, you see the bubbles inside the flask. Did you see that counterintuitive thinking? Because you have the ice on the top, but you also see the water boiling again. So that's our scenario for the students to make a prediction before they see the rest of the part of the video."

Research Design: Capturing Facial Expressions

The Method

Students watch the demonstration video, make predictions, then see the counterintuitive result. Their facial expressions are recorded and analyzed to detect moments of surprise, confusion, and realization.

View Details

The procedure:

  1. Show students the video of water boiling in a flask
  2. Stop the video before the ice is placed
  3. Ask students to predict: "What will happen if we put ice on top?"
  4. Resume the video showing the counterintuitive result
  5. Capture and analyze facial expressions throughout

What we're looking for:

  • Surprise expressions when predictions are violated
  • Confusion when trying to reconcile expectations with reality
  • Engagement patterns during conceptual conflict

Implications for Teaching

Facial expression analysis offers a window into student thinking that traditional assessments cannot provide:

By reading facial expressions, we can detect moments of cognitive conflict - when students' expectations clash with reality. These are the teachable moments where real learning happens.