Unified EEG imaging method improves accuracy in identifying epileptogenic zones

A new advance from Carnegie Mellon University researchers could reshape how clinicians identify the brain regions responsible for drug-resistant epilepsy. Surgery can be a life-changing option for millions of epilepsy patients worldwide, but only if physicians can accurately locate the epileptogenic zone, the area where seizures originate.

Bin He, professor of biomedical engineering, and his team have developed a unified, machine learning-based approach called spatial-temporal-spectral imaging (STSI) to assist. It is the first technology capable of analyzing every major type of epileptic brain signal within a single computational framework. Their work, published in PNAS, presents a technical breakthrough and promising new direction for noninvasive presurgical planning.

Today, most epilepsy centers rely on invasive intracranial electroencephalography (EEG) recordings to determine seizure onset. Patients may be monitored for days or weeks until a seizure naturally occurs. Although the method is accurate, it is time-consuming, costly, and physically taxing. Noninvasive scalp EEG holds great potential as a safer alternative, but clinicians have lacked clarity on which biomarkers such as spikes, high-frequency oscillations (HFOs), or seizures are most reliable for pinpointing seizure-generating tissue. Each biomarker has traditionally required its own analysis pipeline, which has left the field without a unified way to compare them.

He's STSI framework changes this. By jointly analyzing where, when, and at what frequencies brain activity occurs, it can image transient events like spikes and oscillatory events like seizures and HFOs. 

For the first time, one algorithm can handle all epileptic biomarkers. This unified computational approach has never been done before."

Bin He, professor of biomedical engineering, Carnegie Mellon University

Using STSI, the team conducted a multi-year study of 2,081 individual EEG events from 42 drug-resistant epilepsy patients. This is the first rigorous quantitative comparison of all major epileptic biomarkers for source localization, and is the product of a long-standing collaboration with clinicians at the Mayo Clinic, who collected all patient data.

He's group found that pathological HFOs, which occur only when HFOs overlap with spikes, are the most accurate interictal biomarker for identifying epileptogenic brain regions. These pathological HFOs localized the epileptogenic zone within about nine millimeters of invasive seizure mapping, approaching the seven millimeter accuracy achieved using actual seizures.

He explained, "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." In contrast, general HFOs, once considered a promising biomarker, performed poorly. This finding helps clarify years of inconsistent results across clinical studies.

The work also represents a major conceptual shift in electrophysiological source imaging. It provides a noninvasive and faster method to support presurgical planning. Its implications extend beyond epilepsy, because STSI can analyze any EEG or magnetoencephalography (MEG) signal, whether transient or oscillatory. This capability opens doors for studying memory, attention, pain, psychiatric disorders, and normal brain function.

Looking ahead, He hopes to secure new funding to validate the technique in larger patient cohorts and prepare it for clinical adoption. 

"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."