Podcast: Play in new window
BOB HIRSHON (host):
Mood prediction by phone. I’m Bob Hirshon and this is Science Update.
Once patients with bipolar disorder are in deep manic or depressive states, they may not be capable of seeking help, or even taking meds. University of Michigan researchers have developed a phone app that might one day predict an oncoming episode by analyzing phone conversations. Computer scientist Zahi Karam explains that the app doesn’t listen to what patients say—it analyzes the pace, volume and pitch of their voice to gauge their mood. Preliminary tests shows a high degree of accuracy.
KARAM (University of Michigan):
We hope to improve our model so we do a better job of detecting mood in these personal phone calls, and ultimately we hope to show that we can anticipate upcoming mood events.
HIRSHON:
Bipolar disorders affect tens of millions of people. Alerting patients or caregivers to an approaching event could help them better manage their condition. I’m Bob Hirshon, for AAAS, the science society.