Sensors tucked inside a bra can detect emotional states that lead to overeating in time to head off binges, according to studies assisted by a team from Microsoft Research.
In combination with a smartphone app that receives the emotional data in real time, the technology could help overeaters stop themselves before they give in to the impulse, according to "Food and Mood: Just-in-Time Support for Emotional Eating", a paper written by a team from Microsoft, the University of Rochester in New York and the University of Southampton in the U.K.
MORE STUDIES:6 of the very coolest new Microsoft lab projects
The researchers still have to come up with a way for the electrocardiogram and skin-conductivity data gathered by the bra sensors to be sent real-time to the mobile app, but once that hurdle is cleared the intervention system looks promising, the researchers say.
The bra was the garment of choice because it is worn near the heart to accommodate the EKG and if placed on the side the skin-conductivity sensor sits over an area moist enough to give good readings, the paper says.
In their experiments, the researchers tested the data-gathering potential of the sensor-equipped bra, but the data was simply logged. The technology was able to detect with greater than 70% accuracy the two key factors used as a standard measure of emotion in psychological testing.
To test the mobile application, called EmoTree, subjects self-reported their emotions regularly and recorded their emotional state they remember experiencing just before they ate anything. Users reported their current mood and how engaged they were with the task they were performing. If their self-reported current mood indicated they were about to eat based on emotion, the app told them to go to an intervention screen that instructed them to take 10 slow breaths.
Test subjects reported that they would prefer a personalized intervention screen that offered other options than slow breathing, and the researchers say they will work on it.
Meanwhile the sensor bra's usefulness may be short-lived because preliminary work with wrist sensors looks promising, the authors say, which would be much less intrusive.
Read more about software in Network World's Software section.