Preictal scalp EEG connectivity lateralizes EZ in 91% of cases

Connectivity analysis of scalp EEG during the 10 seconds before seizure onset correctly lateralized the epileptogenic zone in 91% of cases, versus 46% when computed over the first 5 seconds of the ictal period. The retrospective study, published open access in Clinical Neurophysiology (Elsevier, 2026), included 13 adults with drug-resistant focal epilepsy monitored on 40- or 64-channel scalp EEG.

The researchers combined Electric Source Imaging (ESI) with Granger causality and hubness centrality metrics. They defined ictal onset as the first EEG change identified by experienced epileptologists, then computed connectivity over the preictal interval and the early ictal segment. The reference standard for epileptogenic zone localization came from resective surgical outcome or stereo-EEG findings — not from the scalp recordings themselves.

What do the numbers show?

Localization performance followed the same pattern as lateralization: 73% concordance in the preictal phase against 27% in the ictal phase. In short, the window before the first EEG change carried more lateralizing and localizing information than the seizure itself.

Lobe of onset shaped the results.

• Temporal lobe epilepsy: 7/7 for lateralization, 6/7 for localization
• Frontal lobe epilepsy: 2/2 for both measures
• Parieto-occipital epilepsy: 1/2 for lateralization, 0/2 for localization

The cohort comprised 7 females and 6 males with a median age of 35 years. Each patient had at least one recorded seizure during prolonged EEG monitoring.

When does discordance matter?

In two cases — one with poor surgical outcome, one with no surgical indication — the connectivity findings disagreed with clinical estimates of the epileptogenic zone. The authors frame this discordance as clinically relevant: it may flag uncertainty in non-invasive presurgical evaluation rather than simply constituting method failure.

That interpretation carries practical weight for the reading room. A preictal connectivity map that contradicts concordant clinical data could prompt closer scrutiny before a patient proceeds to intracranial recording or resection.

What are the limits for clinical practice?

This is investigational methodology applied to a small retrospective sample of 13 patients, not an FDA-cleared diagnostic tool. The parieto-occipital subgroup performed worst, at 0/2 for localization, and the study draws no conclusions about pediatric populations or lower-density montages.

The workflow, however, relies on standard long-term monitoring infrastructure: 40- or 64-channel scalp EEG already in routine use in epilepsy monitoring units. The analysis window — 10 seconds preictal plus 5 seconds ictal — requires only that epileptologists mark seizure onset as they already do.

The authors conclude that connectivity analysis across the preictal-to-ictal transition provides relevant lateralizing and localizing information, particularly in temporal and frontal lobe epilepsy. Their stated significance is that integrating high-density EEG, ESI, and functional connectivity in the phase preceding the first EEG change may support non-invasive presurgical evaluation — a claim this 13-patient series demonstrates but does not yet establish at scale.