In science it's all about quantifiable numbers and results while emotions are more subjective. The two concepts are showing up in some new research and much is being learned. Using high tech devices to read emotions might seem like an impossible task, but as it turns out it’s possible. Researchers from MIT, in the Computer Science and Artificial Intelligence Laboratory (CSAIL), have developed a device they call “EQ-Radio” that is a wifi way to read a person’s emotional state.
It’s working pretty well too. It measures changes in a person’s breathing and heart rhythms by bouncing radio waves off the body and back to the device for interpretation. The team has reported an 87% accuracy rate at detecting a person’s feelings of sadness, anger or excitement. Current methods of measuring emotional responses involve either facial recognition software, which can be easily misinterpreted or on-body sensors that record heart rate and breathing. The problem with wearable sensors is that they are affected by movement or other factors such as temperature and position and are not that reliable for specific information on emotion.
MIT professor and project lead Dina Katabi said in a press release,“Our work shows that wireless signals can capture information about human behavior that is not always visible to the naked eye. We believe that our results could pave the way for future technologies that could help monitor and diagnose conditions like depression and anxiety.” MIT PhD students Mingmin Zhao and Fadel Adib co-wrote a paper on the research with Katabi.
Katabi is no stranger to using wireless technology for monitoring and even predicting human behavior. Aside from teaching at MIT, she owns a company, Emerald, which makes a device that is aimed at detecting and predicting falls among the elderly. She hopes to include the EQ radio technology into the company’s current products so that in addition to preventing falls and injury, perhaps the devices can also alert family and caregivers to emotional issues like anxiety and depression which are common in the elderly.
The wireless signals reflected of people’s bodies were found to measure heartbeats as accurately as a traditional EKG machine. It wouldn’t be an MIT project if some serious number-crunching wasn’t involved of course. The device breaks the heart rhythm into single heartbeats and then extraction algorithms are used to analyze the small variations in heartbeat intervals to determine their levels of arousal and positive affect. While each person is different, it’s about correlation. Signals that correlate to low arousal and negative affect would indicate sadness while a correlation to high arousal and positive affect would likely be tagged as excited.
The learning curve isn’t even that steep. The study reported that even when the device had not been previously used on a particular test subject, it still read emotions with a 70% accuracy rate. Compared to the Microsoft product, Emotion API, which is based on facial expressions, the team found the EQ-Radio system to be far more accurate at correctly identifying emotions. The team will be presenting their paper and the device at the International Conference on Mobile Computing and Networking in New York City in October 2016. The video below explains more about the project and what the team hopes to accomplish with it.
Sources: MIT MobiCom