Authors :
Presenting Author: Srijita Das, MS – University of Nebraska Medical Center
Matthew McCumber, MS – University of Nebraska Medical Center
William Stacey, MD, PhD – University of Michigan
Stephen Gliske, PhD – University of Nebraska Medical Center
Rationale:
High-frequency oscillations (HFOs) are promising biomarkers of epileptic tissue. While circadian and ultradian patterns of epileptic activity have been reported, the temporal dynamics of HFOs across the 24-hour cycle remain underexplored. Understanding whether HFO rates exhibit consistent circadian fluctuations could enhance our mechanistic understanding of epilepsy and support future chronobiological strategies in seizure forecasting. This study aimed to quantify circadian rhythmicity in HFO rates across a large cohort of epilepsy patients undergoing long-term intracranial EEG (iEEG) monitoring as pre-surgical evaluation.
Methods:
We retrospectively analyzed continuous, multi-day, sleep-scored iEEG recordings from 68 consecutive patients with drug-resistant epilepsy at the University of Michigan between 2016 and 2022. Poor-quality data were excluded, resulting in 61 patients. HFOs were detected using a validated automated algorithm (qHFO detector)1, including artifact rejection, which was applied to non-overlapping 10-minute windows. For each segment, HFO rates were computed and assigned a corresponding circadian time (0–24 hours). To assess rhythmicity, we computed the first circular statistic (first Fourier moment) within multiple vigilance states (NREM 1-3, REM, Awake), correcting for non-uniform sampling and incomplete data using the unfolding method2. We focused on quantifying temporal dispersion of HFO rates across the 24-hour cycle. Patients with insufficient data or technical issues were flagged for separate investigation.
Results:
A majority of patients exhibited prominent circadian modulation in HFO rates. We observed significant 24-hour cyclicity in HFO rates during NREM sleep in all subjects except one (p < 0.05). The first circular moment displayed statistically significant 24-hour periodicity in over 75% of cases (p < 0.05), with individual peak times varying across patients. Circular histograms showed non-uniform distributions of HFO activity over time, and the analysis supported structured, non-random fluctuation in rates across the circadian cycle. Three patients did not show clear rhythmicity, and further analysis is ongoing to determine whether this is due to biological variability or technical artifacts.