Abstracts

Rationale: High Frequency Oscillation (HFOs) are a powerful biomarker of epileptic tissue and a potential specific marker of epileptogenicity. But their utility is hampered by requiring long-term spontaneous recordings and difficulties detecting them with non-invasive means. With probing stimulation, we could analyze averaged stimulation evoked HFOs, which increases signal to noise ratio (SNR), and perform tests for limited time in specific brain states. Understanding the characteristics of evoked scalp HFOs is the first step towards its use as both a localizer and predictor of epileptogenesis.

Methods: Six patients with depth electrodes implanted to localize their epileptic focus underwent pseudo-random multi-region stimulation during sleep. Single pulse electrical stimulation (SPES) was delivered to two adjacent contacts at 7 mA, biphasic pulse duration 0.253 ms or 1 ms, every 3-5s +/-0.25s (random jitter). 10-20 pulses were delivered to each location.

The following pipeline ensured that we were not filtering the stimulation artifact and that edge effects were minimized. 1) The scalp EEG was re-referenced to bipolar montage; 2) For each bipolar channel and trial, 15 ms around stimulation were removed (with a Tukey-windowed median filter); 3) the inverted signal was used as padding to reduce edge effects; and 4) then band-filtered in the Ripple band (FIR sharp filter 70-110 Hz, to avoid line harmonics). 5) The extra padding was then removed, and 6) the filtered EEG was normalized by z-scoring per trial. We defined early (15-100ms after SPES) and late ( >100ms) responses.

Results: In analysis of sleep datasets from six patients, stimulation in neocortical regions evoked both intracranial and scalp HFOs in half of the patients. In these patients, scalp HFOs were induced in at least half the trials when stimulating in the seizure onset zone (SOZ) but not when stimulation was delivered to a contralateral channel. SOZ was in the frontal (N=2) or temporal (N=1) lobes. Early and late HFO responses were observed. Figure 1 shows an example of evoked scalp HFOs and intracranial HFOs following individual SPES delivered to the SOZ. In this representative example, stimulation evoked intracranial HFOs are visible at the single trial on contacts in the SOZ, near the stimulation channel. Corresponding early scalp HFOs are also observed at the single trial level. Early responses are detectable with time averaging locked to the SPES.

Conclusions: This initial study confirmed the presence of stimulation-evoked HFOs on the scalp and in the intracranial SOZ following SPES in neocortical areas. It also points to the challenges in detecting these small events and the advantage of probing the system and tailoring the analysis window. Stimulation evoked scalp HFOs could help identify and potentially predict patients with epilepsy.

Funding: Please list any funding that was received in support of this abstract.: This work was supported in part by the Tiny Blue Dot foundation, NIH grant K24-NS088568, and NIH grant R01-NS062092. RZ was supported in part by the MGH ECOR Fund for Medical Discovery Postdoctoral Fellowship Award.