MIT Invents A Device That Detects Your Sadness–With Radio Waves

The technology opens the door for devices that recognize your emotions, even when you can’t.

The next generation of conversational UIs will do more than just understand our words. They’ll listen our tone, getting to the heart of what emotions we express as we speak to them. Yet that’s nothing compared to what a team at MIT is working on: computers that use radio waves to detect what you’re feeling instantaneously.


So imagine this. You come home after a hard day of work at the office, disheartened. No one’s home, but as soon as you walk inside, your internet of things-controlled stereo cranks your favorite music, and the lights readjust themselves, turning your living room into a warm, relaxing space. Or maybe you’re watching a horror movie, and it’s a bit too scary for you: the lights around you might get a bit brighter, while the television lowers the volume and bass of the film itself. Heck, imagine playing a video game on your console that gets more challenging or exciting when it begins to bore you.

Those are just a few of the user experiences that an experimental technology called EQ-Radio could enable. Developed by a team of researchers led by Dina Katabi, a Professor of Electrical Engineering and Computer Science at MIT’s Computer Science and Artificial Intelligence Laboratory, the EQ-Radio uses standard wireless signals–like the ones used by RFID transmitters–to detect how hard a person’s heart is beating.

It works something like a cross between an electrocardiogram and a dolphin’s echolocation skill, allowing devices elsewhere in the room to measure the intensity of your pulse, then predict what emotions you’re feeling based on what pattern the results take: the waveform for arousal, for example, looks very different than the waveform of depression.

How well does this work? According to Katabi, the technology is surprisingly accurate. The EQ-Radio measures heartbeats just as accurately as an ECG monitor, and can correctly predict a person’s emotional state 87% of the time, given repeated exposure, or up to 70% if the EQ-Radio has never “heard” their heartbeat before. And yes, the EQ-Radio can even distinguish between multiple people at once, thanks to their unique heart patterns.

That’s good enough that Katabi envisions the EQ-Radio being a great supplemental sensor in many smart homes, helping control ambient settings like music and lighting. “By looking at your mood over a longer length of time, it could even make suggestions for things it thinks you might enjoy,” she says. Imagine an EQ-Radio hooked up to Amazon Echo. That could be dangerous on the wallet–mass orders of chocolate that arrive like clockwork every time you’re blue–but Katabi has some more benign ideas in mind. For example, if you’ve been depressed for a few days, your EQ-Radio might just encourage you to go out for a walk.

That also touches on some of the EQ-Radio’s medical applications. It could be used to help manage everything from heart arrhythmia to depression. An EQ-Radio could detect when you have a heart attack, and alert the hospital. Caregivers could use an EQ-Radio to make sure their charges were in good spirits. And forget wearables–because it can detect how fast your heart is beating, the EQ-Radio could also help you track your fitness. Imagine an aerobics app that automatically adjusts your workout according to how hard your heart is beating.


Sounds like the kind of sensor that should come standard with all smart gadgets, doesn’t it? Unfortunately, EQ-Radio isn’t a consumer-facing product yet, and as a researcher, Katabi doesn’t know when it might come to market. But from a technology perspective, computers that can read our hearts and make intelligent guesses as to what we’re feeling are a solved problem.

But should we be so quick to invite these mind-reading devices into our lives? Such technology also raises significant privacy concerns, as Fast Company‘s Steven Melendez points out today. Although EQ-Radio currently requires users to opt-in by making specific gestures to indicate their consent to be tracked, there’s nothing stopping others from deploying this same technology in a more invasive way, he writes. For example, employers could use something like the EQ-Radio to call your bluff during salary negotiations.

Balancing such concerns with the benefits of truly intuitive computers will be a challenge to anyone bringing such a product to market. So depending on where you’re sitting, EQ-Radio is either the next evolution in user experience, or a dystopian privacy nightmare waiting to happen. Which is it for you?

[Photos: Jason Dorfman/MIT CSAIL]