SEEG Guides Surgery in Bilateral or Unclear Scalp EEG Onset

Stereo-electroencephalography (SEEG) can localize the epileptogenic zone in patients whose scalp EEG shows bilateral independent or unclear seizure onset, according to a study by Soni and colleagues published in Epilepsia (2026). The paper addresses one of the most difficult referral scenarios in epilepsy surgery programs: the drug-resistant patient whose noninvasive Phase 1 evaluation fails to point to a single hemisphere.

The clinical problem is familiar to every monitoring unit. Prolonged video-scalp EEG with standard montages sometimes captures seizures that begin independently on both sides, or ictal patterns that offer no clear lateralization. Historically, many of these patients were declined for surgery on the assumption that bilateral onset implied a diffuse or multifocal epilepsy. The Soni study examines whether SEEG — the stereotactic placement of depth electrodes for intracranial recording — can salvage surgical candidates from this diagnostic gray zone.

SEEG differs from subdural grid ECoG in both technique and logic. Electrodes are implanted stereotactically through small burr holes on the basis of a preformulated hypothesis about the anatomic network of seizure onset and propagation. The method allows sampling of deep structures — mesial temporal, insular, cingulate, opercular — that surface grids cannot reach, and it does so bilaterally in a single implantation when the hypothesis requires it. That bilateral capability is precisely what makes SEEG attractive when scalp EEG cannot exclude a second focus.

The study population consists of patients with bilateral independent or unclear scalp electroencephalographic seizure onset who proceeded to SEEG implantation. The authors report on how often the intracranial recording resolved the laterality question, whether the implanted hypothesis led to a defined surgical target, and what the operative outcomes were. For reading-room neurophysiologists, the relevant question is concrete: when scalp ictal patterns are ambiguous or discordant, what proportion of SEEG cases still yield a focal, resectable onset zone?

This is investigational clinical research in the sense that matters here — not device development, but a single-center evaluation of an established invasive diagnostic workflow. SEEG hardware and implantation technique are standard practice in comprehensive epilepsy centers in Europe, Canada, and increasingly in the United States. The study's contribution is evidence about patient selection and yield in a specific, hard-to-stratify group, not a new FDA-cleared technology.

The practical implications for monitoring-unit practice are worth stating plainly. First, bilateral or non-lateralizing scalp onset is not, by itself, a reason to end the presurgical evaluation. Second, SEEG implantation in this population demands an explicit anatomic hypothesis — typically formulated from MRI, PET, ictal SPECT where available, and semiology — because electrode coverage must be planned before laterality is confirmed. Third, intracranial review in these cases carries particular weight: the interpreting neurophysiologist must distinguish true independent bilateral onsets from rapid secondary bilateral synchrony, a distinction scalp EEG often cannot make but depth recordings can.

For clinical neurophysiology teams, the study also underscores the reporting burden. In SEEG cases with bilateral coverage, ictal onset should be described contact by contact, with attention to the earliest electrographic change and to whether consecutive seizures arise from the same or different electrodes. That level of granularity determines whether the surgical conference sees one resectable focus, two independent foci, or a network unsuitable for focal resection.

The authors connect their findings to the broader question of who reaches surgery at all. Patients with ambiguous scalp onset have historically carried poor odds of progression from Phase 1 to any intervention, and the paper's data speak to whether SEEG changes that trajectory — in localization yield, in resection or neuromodulation decisions, and in seizure outcome at follow-up.

Limits remain. SEEG samples only the territories covered by the implantation plan, so a negative or multifocal result partly reflects the hypothesis implanted, not the whole brain. Complication rates for depth electrode placement, though low in experienced centers, are nonzero, and patient selection criteria still vary widely across programs.

The Soni paper gives epilepsy programs a quantified basis for offering SEEG to patients whose scalp EEG once closed the surgical door, and future multicenter work will need to test whether those localization rates and postoperative outcomes hold beyond the single-center experience reported here.