Researchers at Carnegie Mellon University have developed a new method that could improve how doctors identify the brain regions responsible for drug-resistant epilepsy. The technique, called spatial-temporal-spectral imaging (STSI), uses machine learning to analyze all major types of epileptic brain signals within one computational framework.
Currently, many epilepsy centers use invasive intracranial electroencephalography (EEG) recordings to find where seizures start in the brain. This process can take days or weeks and is physically demanding for patients. While noninvasive scalp EEG is safer, it has been unclear which biomarkers—such as spikes, high-frequency oscillations (HFOs), or seizures—are most reliable for locating seizure-generating tissue. Traditionally, each biomarker required its own analysis method.
The STSI framework addresses this by jointly analyzing where, when, and at what frequencies brain activity occurs. It can image both transient events like spikes and oscillatory events like seizures and HFOs.
In a multi-year study involving 42 patients with drug-resistant epilepsy and 2,081 individual EEG events, the team found that pathological HFOs—those overlapping with spikes—were the most accurate interictal biomarker for identifying epileptogenic regions. These pathological HFOs localized the epileptogenic zone within about nine millimeters of invasive seizure mapping results, which is close to the seven-millimeter accuracy achieved using actual seizures.
"You can record pathological HFOs in under an hour, instead of waiting days for a seizure. The accuracy is only two to three millimeters different," said Bin He, professor of biomedical engineering at Carnegie Mellon University. General HFOs performed poorly as biomarkers compared to pathological HFOs.
This research provides a faster and noninvasive method to support presurgical planning for epilepsy surgery. According to the researchers, STSI can also be used to analyze any EEG or magnetoencephalography (MEG) signal—including those related to memory, attention, pain, psychiatric disorders, and normal brain function.
Looking forward, He aims to secure additional funding to validate STSI in larger patient groups and prepare it for clinical use.
"The whole point is to help others," He said. "If we can provide a noninvasive, precise alternative that spares patients from days of invasive monitoring, that would have a major impact. We're commited to improving the patient experience through our expertise."
