Abstracts

Rationale: High frequency oscillations (HFO, ripples: 80–250 Hz, fast ripples: 250–500 Hz) are novel EEG markers of epileptic tissue. Additionally, there is evidence for physiological ripples that occur mainly in mesiotemporal structures. Epileptologists are challenged to distinguish epileptic and physiological HFO, if they aim to use epileptic HFO for clinical diagnostics during intracranial EEG, especially in temporal epilepsies. Differentiation is complicated by the fact that both types of HFO are most likely generated in the same brain structures. No simple frequency measure for distinction could be identified yet. Recent research suggested that ripples associated with sleep spindles are physiological events and likely play a role in memory consolidation. Ripples in the seizure onset zone (SOZ) and co-occurring with epileptic spikes are most likely epileptic in nature. This study aimed to identify parameters in the amplitude (AMP) and frequency (FRQ) domain that serve as distinguishing features of physiological and epileptic ripples.Methods: We included patients who underwent chronic intracranial EEG with mesiotemporal implantation and simultaneous scalp EEG at the Freiburg Epilepsy Center. EEG sampling rate was 2 kHz. The following EEG events were visually marked for each patient during slow wave sleep: 50 sleep spindles in frontal and parietal scalp contacts; all ripples in mesiotemporal contacts during these spindles (spindle ripples); all ripples and spikes in SOZ contacts and the irritative zone during 1 min. Ripples co-occurring with spikes (spike ripples) and ripples inside the SOZ (SOZ ripples) but outside of spindles were considered epileptic and ripples co-occurring with spindles physiological. All ripple events were analyzed for duration, 5 parameters describing waveform morphology (of these 3 described AMP-features) and 4 parameters characterizing FRQ-content (see table 1). Groups of physiological and epileptic ripples were compared using a Kruskal-Wallis test (α<0.05).Results: We analyzed 437 electrode contacts of 19 patients. 2858 spindle ripples and 3836 epileptic ripples (including 1290 spike ripples) were identified. We found significant differences between HFO groups with respect to all parameters. Epileptic HFO had higher values in all 3 AMP-parameters, a longer duration and more peaks per sample (details in table 1). Physiological spindle ripples had higher values in 3 of 4 FRQ-content parameters. Significant AMP-parameter differences were seen not only for spike ripples but also for SOZ ripples compared to spindle ripples (see example in figure 1).Conclusions: Physiological spindle ripples differ in waveform morphology (especially AMP-features) as well as in FRQ-content from epileptic ripples. Analysis of spindle ripples may allow identifying a subgroup of ripples different from epileptic ones. The predictive value of the depicted features to distinguish between physiological and epileptic ripples remains unclear. The presented results suggest that an improved separation of epileptic ripples may facilitate the identification of the epileptic focus in the future.