How are you feeling today?

In a paper that will be presented next month at MobiCom, the Annual International Conference on Mobile Computing and Networking, researchers from MIT’s CSAIL announce that they have developed a device that can determine people’s emotions by analyzing reflections in wireless signals.

Unlike other emotion detection systems, EQ-Radio doesn’t rely on typical emotional cues, like facial expressions, which are not always reliable. Instead, it works like this: a wireless device that the team calls EQ-Radio sends out a wireless signal that bounces off your body. The reflection measures not only your breathing, but also your heart rate.

AI photo

Using that information, the device can predict whether the person being observed feels happy, sad, angry, or excited. The team reports in the paper that in a sample size of 30 people ranging in age from 19-77, EQ-Radio guesses the emotion correctly 87 percent of the time–a success rate that they say is higher than the competing emotion-detection solutions of rivals like Microsoft’s Emotion API.

“Our work shows that wireless signals can capture information about human behavior that is not always visible to the naked eye,” Dina Katabi, the leader of the project said. “We believe that our results could pave the way for future technologies that could help monitor and diagnose conditions like depression and anxiety.”

Because it can measure heart rate, it might also be a less invasive way for doctors to monitor patient’s heartbeats, potentially watching for conditions like arrhythmias without the need to be hooked up to monitoring devices.

There are also more fun applications for the device. TV and Movie studios could use the tech to get an accurate read on what parts of their movie worked in focus groups, and smart homes could read your emotions and move the mood music and lighting to match, well, your mood.