More than 3 million Americans have epilepsy, a neurological disorder characterized by unpredictable seizures. For almost 100 years, doctors have diagnosed epilepsy with a scalp electroencephalography (EEG) head scan, looking for abnormalities in the scanned brain waves between seizures, when the brain is considered at rest. The abnormalities don’t always present themselves, however, resulting in a nearly 30% misdiagnosis rate.
Biomedical engineering professor Sri Sarma and neurologist Khalil Husari have developed an EEG analytics algorithm that uses at-rest data to build a “heat map” of a patient’s brain activity that a doctor can then quickly and definitively interpret.
“Our tool looks for pathological connections in the brain that persist during resting state and thus can identify when a patient is more likely to have epilepsy even when the EEG looks normal to clinicians,” Sarma says. “[The tool] has the potential to reduce misdiagnosis rates to under 10%,” she adds.
Sarma and Husari plan to use a grant from the Louis B. Thalheimer Fund for Translational Research to perform a study of 200 patients to validate the algorithm’s accuracy and customize another algorithm that removes excess brain activity from the heat map. Johns Hopkins Technology Ventures is seeking a patent for the technology.