What if your smartphone could detect your mood simply by analyzing changes in your usual phone-usage patterns? A new app from, a spinoff from the MIT Media Lab, can do that, and it is almost ready to be deployed to diabetes patients, who are prone to sudden mood shifts that signal dangerous blood-sugar levels. collected 320,000 hours of data from research participants’ cellphones and used it to determine what kinds of usage patterns signal the beginning of issues like the flu or anxiety. The app platform runs in the background, analyzing call frequency, location, and text-messaging habits to figure out when users aren’t feeling like their usual selves. The app only looks at statistics, not phone-call content, so it doesn’t know what you’re saying, only that the number of calls you’re making has changed or that you’re not leaving the house. When you seem to be displaying abnormal behavior, the app sends out an alert to your network of caregivers. "The idea is for psychosocial support. This helps [users] go a long way to stay on their treatment regiment," a rep told "It’s a 'check engine’ light." Next year, the startup will run a study with diabetic patients and clinicians to see if the app makes a difference. Researchers also say they see applications for assisting people with mental-health issues.