A brain implant may be able to help predict seizures for epilepsy patients who don't respond well to drugs, a new study finds.

The small pilot study, published in Lancet Neurology, tested the device in 15 patients. The medical device company NeuroVista funded the study and several scientists who work for the company were also responsible for performing the research. NeuroVista stands to profit if the device is a market success.

"We just wanted to see if this is feasible, and this study shows that it is," said lead researcher Dr. Mark Cook, of the University of Melbourne and St. Vincent's Hospital in Australia.

If people with epilepsy know that a seizure is coming, they can avoid driving or swimming that day, added Cook. They might also be able to advantageously adjust their medication use.

The surgical procedure entails implanting electrodes between the skull and brain, and connecting them with another electrode implanted in the chest.Data from the implants is transmitted to a handheld device that patients can use to monitor the likelihood of experiencing a seizure. High risk is indicated by a red light, medium risk by a white light, and low risk by a blue light.

The patients in the pilot study were aged 20 to 62, and suffered between two and 12 seizures per month.

The devices correctly predicted seizures with a "high warning" sensitivity greater than 65 percent, and gave better than random predictions in 11 of 15 cases.

Some patients experienced negative side effects, including serious complications like infection and chest pain caused by the chest implant moving. Two of the 15 patients decided to have the device removed.

Complications occurred at a rate typical of implantable deep-brain devices, including devices related to Parkinson's disease, researchers said.

"Knowing when a seizure might happen could dramatically improve the quality of life of people with epilepsy by giving them back some independence in their lives," said Cook. "A lot of patients with epilepsy will tell you it's not the seizures themselves, but the fact they don't know when they will happen, that is the worst part of their condition."

There are plenty of questions left, said Dr. Ashesh Mehta, director of epilepsy surgery at the North Shore-LIJ Comprehensive Epilepsy Care Center in Great Neck, New York.

"This study is an important first step," said Mehta, who was not involved in the research. "The next step would be to implant these in a larger sample of patients. And you need to see which groups of patients might be good candidates for this."

The authors also recommend more research be done, and said that their work could be used as the basis for future studies.