Scalp fast oscillations predict seizure recurrence following a first unprovoked seizure

First unprovoked seizures (FUS) pose a prognostic challenge because recurrence risk varies substantially and conventional EEG has limited predictive sensitivity. Scalp fast oscillations (SFOs) have emerged as potential biomarkers of epileptogenesis, but their predictive value in adults remains uncertain. We investigated whether SFOs detected on routine EEG predict seizure recurrence after FUS in adults.

In this retrospective cohort study, adults (≥18 years) who underwent EEG evaluation for FUS between 2010 and 2020 were included. Time-frequency analysis was performed to identify SFOs, classified as isolated (ISFOs) or superimposed on spikes/sharp waves (SFOSs). Seizure recurrence was monitored through 2025. Cox proportional hazards models assessed associations between SFOs and recurrence, adjusting for interictal epileptiform discharges (IEDs), slow waves, EEG system, and antiseizure medication (ASM) initiation.

Among 86 patients, 24 (27.9%) experienced seizure recurrence. Both ISFOs and SFOSs were significantly more prevalent in the recurrence group (p < 0.001). The presence of SFOs ≥40 Hz was independently associated with recurrence (adjusted HR: 10.80; 95% CI: 3.00-38.87; p < 0.001), outperforming IEDs and slow waves. Receiver operating characteristic analysis showed that SFOs ≥40 Hz had the highest AUC (0.804), compared to IEDs (0.663) and slow waves (0.609).

Gamma-range SFOs (≥40 Hz) detected using standard EEG protocols are independently associated with seizure recurrence after FUS and outperform conventional EEG markers. These findings support SFOs as a clinically accessible biomarker for early epilepsy risk stratification, warranting prospective validation in larger and more diverse cohorts.