Developmental profile of physiological high-frequency oscillations in the human brain

High-frequency oscillations (HFOs), transient burst of ≥ 80 Hz activity, are increasingly recognized as promising EEG biomarkers of the epileptogenic zone for guiding epilepsy surgery. However, the presence of physiological HFOs in healthy cortex complicates their interpretation. The developmental profile of physiological HFOs in children remains poorly characterized, despite profound maturational changes in pediatric EEG. Using a large multi-institutional cohort of 185 pediatric patients undergoing intracranial EEG, we constructed a normative developmental atlas of non-spike HFOs derived from non-epileptic channels, excluding seizure-onset and spiking regions, representing putative physiological HFOs. HFOs were identified using automated HFO detectors followed by deep learning-based artifact rejection and spike-HFO classification. HFO characteristics (rate, power, peak frequency, and duration) were compared across age groups (young: 0-6 years; middle: 7-12 years; old: ≥13 years), and correlations with age were assessed. Topological analysis revealed pronounced developmental shifts in HFO distribution. In younger children, HFO rates exhibited fronto-parieto-temporal predominance that progressively decreased with age, whereas occipital HFO activity increased across development. Power showed a modest age-related increase, particularly in older participants, whereas HFO duration was longer in younger children. Peak HFO frequency remained stable across age groups. These findings delineate the developmental profile of physiological HFOs in the pediatric brain. Establishing normative age-dependent patterns of physiological HFOs may improve interpretation of high-frequency activity as spatial EEG biomarkers and support more precise delineation of the epileptogenic zone in pediatric epilepsy.