Using software to pinpoint seizure activity

Sridevi Sarma was trained as an electrical engineer, but during her doctorate studies at the Massachusetts Institute of Technology, she became more interested in neuroscience through her minor in brain and cognitive studies. Now an assistant professor in the Johns Hopkins Department of Biomedical Engineering, Sarma is using that engineering expertise to help doctors understand the electrical activity inside the brain during epileptic seizures.

Today, doctors trying to stem the tide of seizures in epileptic patients rely on discovering the epileptogenic zone, a region or regions in the brain where the seizures originate. Some patients end up requiring surgery so electrodes can be implanted into the brain. Patients might be required to sit in a hospital room for days with electrodes implanted into their brains, waiting to have seizures as doctors monitor EEG data to see how the brain is misfiring. Once doctors determine where seizures are presumably starting, the epileptogenic zone is removed through surgery. The average success rate, Sarma says, is about 40 percent.

“Sometimes patients have very large regions of their brains removed, which is irreversible, but then their seizures come back,” Sarma says.

As an alternative to interpreting EEG data, Sarma is working with colleagues from the School of Medicine to develop EZTrack. The software product takes in the same recordings as those collected while epileptic patients sit inside hospitals, but it uses graph theory — network-based data analysis — that allows doctors to interpret recordings in real-time as a graph of electrical signals for more accurate pinpointing of the epileptogenic zone.

“We’re trying to find a region of interest — a set of electrodes that doctors believe they should pay attention to in terms of where the epileptogenic zone is,” says Sarma of her work with William S. Anderson and Nathan Crone of the Johns Hopkins Department of Neurology and Neurosurgery. In testing with 19 patients so far, EZTrack has been accurate 100 percent of the time in predicting surgical outcome. “In preliminary trials, [EZTrack shows] that the surgery was a failure before the surgery even happens,” she says. In other words, the software could help doctors avoid performing surgeries that won’t help patients.

With $100,000 in grant funding from the Johns Hopkins-Coulter Translational Partnership, established in the Department of BME, Sarma plans to build a prototype of EZTrack and expand clinical testing. The Coulter Foundation awards are designated to advance promising biomedical engineering innovations toward commercialization and patient use. The program runs from 2012 to 2017.

— Andrew Zaleski