Neural Networks

Pinpointing epilepsy  

Gwendolyn Guay began staring into space and mumbling to herself when she was 4. Over time, her episodes became more frequent and stranger—she would become unresponsive for up to a minute, pressing her fingers together while making soft clucking sounds. A specialist in Maine diagnosed Gwen with epilepsy.

Gwen’s epilepsy proved to be intractable. After five years trying different medications, the Guays met with Blaise Bourgeois, MD, director of Children’s Hospital Boston’s Epilepsy Program, who leads a multidisciplinary evaluation by a team of neurologists, neuropsychologists, radiologists, imaging specialists and neurosurgeons.

Diagnostic evaluation

The team’s goal was to pinpoint the “trigger point” causing the seizures and remove it. “It’s like trying to defuse a bomb,” says Bourgeois. “The hope is that you can go in and remove the trigger.” While this kind of surgery is the most effective treatment—and epilepsy’s only real cure—only about 5 percent of patients are surgery candidates.

Using the latest brain imaging devices, the epilepsy team searched for the source of Gwen’s seizures as well as areas housing her crucial functions, like movement, language or memory.

Neurophysiologists conducted a week-long session of 24-hour video electroencephalogram (VEEG) monitoring, recording the electrical activity of Gwen’s brain through 26 electrodes. Results suggested a left brain focal point, but weren’t conclusive. Sophisticated brain imaging tests—a Positron Emission Tomography (PET) scan that tracked chemical activity, Single Photon Emission Computerized Tomography (SPECT) scan which tracks blood flow, and a Magnetic Resonance Imaging (MRI) which painted a three-dimensional picture—provided additional clues. Together, the results suggested the left temporal lobe.

Understanding what impact a resection would have and its proximity to healthy tissue were also important. Special attention was paid to language and memory—skills traditionally housed in the left temporal lobe, close to Gwen’s epileptic area. To clarify which parts of her brain controlled these, doctors performed a Wada test. They numbed one whole side of Gwen’s brain, causing the opposite half of her body to go as limp as if she’d had a stroke. While each side was temporarily paralyzed, the doctors asked Gwen to perform tasks. She couldn’t speak or remember when her brain’s left side was asleep.

Neurosurgeons needed a detailed map of her brain’s functions—millimeter by millimeter—which required implanting subdural strips of silicon, studded with 108 electrodes. “It’s the most accurate data you can get,” says Bourgeois. A few days later, with scores of fine wires coming neatly out of Gwen’s head, doctors performed a cortical stimulation test; they asked her questions while applying small shocks to each electrode in turn, temporarily rendering that part of her brain useless. If she failed a task, doctors knew that the part of her brain receiving the stimulus was being used for that thought process. Results were good: They zeroed in on the epilepsy, and the areas right around it didn’t seem to be used for language. But whether Gwen’s memory would be safe was still unclear.

Team discussion

With the “blueprint” of Gwen’s brain in hand, the team began an hours-long debate about whether to proceed with surgery. Neurologists, neurosurgeons, radiologists, nurses and psychologists examined the monitoring and imaging results and debated the likelihood of success. “We had lots of concerns about Gwen’s memory,” says Bourgeois. “If we took out functional tissue, she wouldn’t be able to recall or retain words.” But that possibility had to be balanced against the risks of living with such debilitating seizures. “Most of my patients tell me that having even one seizure a week ruins their lives,” says Bourgeois. “Living with her seizures would be a constant worry and self-esteem issue for her as she got older.”

Meanwhile, the team made a significant discovery. Just days before her scheduled surgery, a specialized high-powered MRI revealed a cortical abnormality in Gwen’s left temporal lobe, near the hippocampus but not involving it. “It was a big reassurance and we recommended her for the surgery with a very high certainty that it would be a success,” says Bourgeois. And the procedure did go smoothly: Gwen’s neurosurgeon removed the electrodes that mapped her brain, and easily located the epileptic tissue.

“Grids and strips can be risky, but Gwen clearly benefited from it,” says her neurosurgeon Joseph Madsen, MD. “The collective data from all of Gwen’s tests correlated with the abnormal cortex seen on her MRI, outlining a very clear cut area for me. My job is to remove the bad tissue, but I couldn’t do that without the expertise of our team.”

Within the week, the Guays left for home. Gwen will return for regular follow-up visits, but all signs point toward a seizure-free future.