A flurry of coordinated activity in a brain-spanning network of neurons may sound like the formation of a brilliant new idea, but it is actually the description of a seizure. Understanding why and how this synchronization spreads would be a critical tool in treating severe epilepsy.

As a network scientist, Danielle Bassett, the Eduardo D. Glandt Faculty Fellow and associate professor in the Department of Bioengineering in the School of Engineering and Applied Science, studies how the interconnections between members of a group influence the behavior of the whole. Looking at epilepsy through that lens, she is collaborating with Perelman School of Medicine researchers, including Brian Litt, a professor of neurology and neurosurgery at Penn Medicine and director of the Penn Epilepsy Center, and Ankit Khambhati, Bassett’s postdoctoral fellow and a recent graduate of the Litt Lab.  

In an earlier study, Bassett, Litt, and Khambhati developed a computer model of seizure networks based on brain recordings from Penn’s epilepsy patients.

Now, the researchers have shown that a second network acts on the one directly involved in the seizure, influencing whether the pathological synchronization remains confined to a local area or spreads across the brain.   

The researchers added a new dimension to their simulation, developing a technique known as “virtual cortical resection.” By simulating the surgical removal of different sections of the brain—a last-ditch option for treating severe epilepsy—the researchers could test how the two networks interact.  

“Using this method,” Bassett says, “we were able to show that there are some regions of the regulatory network that can push the seizure network into a less active state, or pull it out of that state.”

This “push-pull” mechanism appears to work in a manner similar to other biological processes that maintain homeostasis, such as the regulation of heart rate or body temperature, but the researchers are the first to show this kind of regulatory network for epilepsy.    

Identifying which regions are which in a patient’s regulatory network could guide new treatment options, such as implantable stimulation devices that would bolster the nodes that help quiet seizure activity, or laser surgery to eliminate the nodes that promote it.

“Our team’s method offers an exciting way to simulate the effect of different therapeutic interventions on patients and predict outcome and side effects,” Litt says.