People hospitalized for epilepsy testing helped an artificial intelligence system make fewer mistakes simply by telling it when its seizure alerts were wrong. In a small NIH-funded study at the University of Pennsylvania, patient feedback cut the system's false seizure alarms by about 80% in a computer simulation, according to results published in Nature Medicine. The study included only 13 patients and was a proof of concept, not a test of a new treatment.
The research points to a future in which brain-monitoring devices can communicate with patients in real time rather than storing data for doctors to review days later. For families living with epilepsy, frequent false alarms can mean needless worry, disrupted sleep, and lost trust in a device. About one-third of the world's roughly 70 million people with epilepsy have seizures that medication does not control, the researchers wrote.
