Mental health: High-tech patch reads human emotions

Relying solely on facial expressions to understand emotions can be misleading. Many people put on a brave face, but aren't. To help medical professionals better distinguish between fake and real emotions, a team of researchers has developed a sticker. While the stretchable and rechargeable sticker is still in the research and development phase, it will eventually be able to use body signals to measure what's really going on inside a person.

Because bottling up emotions for too long risks health consequences. The technology is intended to help people who struggle with mental health issues but aren't willing to admit it to themselves or others. By tracking certain signals, the sticker is designed to detect problems like anxiety or depression earlier. This is the goal of the development led by lead author Huanyu Cheng, Associate Professor of Engineering and Mechanics at Pennsylvania State University.

The smart patch offers another advantage: It bridges cultural and social differences. Many people express their emotions toward healthcare professionals in a rather stoic manner, while others express themselves very expressively, as Cheng pointed out.

Specifically, the patch measures various physiological responses associated with diverse emotional states, such as skin temperature, humidity, heart rate, and blood oxygen levels. Each sensor operates independently, so the measurements cannot influence each other. The patch also uses flexible metals such as platinum and gold, which are processed in a wave-like pattern. The goal was to ensure that the sensitivity of the sensors remains even when the material is twisted or pulled.

The researchers also used layers of material that conduct electrical current differently depending on temperature. They also incorporated hollow tubes made of carbon atoms into the sticker, which absorb water and measure the moisture content. The sticker then transmits the recorded information wirelessly to mobile devices or a cloud.

For the study, the team of authors first trained an artificial intelligence (AI) model to read and understand signs of simulated and real human emotions. Real people then displayed six common facial expressions: joy, surprise, fear, sadness, anger, and disgust. The researchers also fed this information into the learning system. Later, the patch classified the facial expressions performed on additional test subjects with an accuracy of more than 96 percent.

When it came to detecting real emotions, Cheng and his colleagues tracked people's reactions while they watched a video clip designed to evoke specific feelings. The patch's accuracy rate for the displayed reaction was nearly 89 percent. The sensor measurements confirmed known connections between reactions and emotions, such as the increase in skin temperature and heart rate during surprise and anger, respectively.