Abstracts

Rationale: It is well known that pathologic high-frequency oscillations (pHFOs) underlying the genesis and evolution of seizures. The pHFOs have been categorized into fast ripples and ripples based on the differences of frequency spectrums and the sources of corresponding neurons. Although the pHFOs are recognized as important biomarker for localization of an epileptic region, the interaction of fast ripples and ripples during a seizure are not understood well. In this research, we extracted the features of burst repeating and power changing of fast ripples and ripples from intracranial seizure signals. Then we measured the change of cross-correlation coefficient between the fast ripples and the ripples to exam the reciprocal influence from a seizure onset to termination. Methods: Eleven intracranial recordings from 5 patients who underwent epilepsy surgery were analyzed. Signals derived from subdural grid and depth electrodes were sampled by 4K Hz. We firstly acquired fast ripples and ripples through 2 bandwidth filters: 80-250 Hz and 250-500 Hz. We enveloped these 2 filtered ECoG waveforms to measure the bursting features of fast ripples and ripples serially by autocorrelation function. Then we used cross-correlation to measure the similarity of these two series to observe the lag of one relative to the other. Finally, we compared the power change of the fast ripples and ripples to measure the proximity of alignment between these two given pHFOs Results: We analyzed the seizure dynamics of ripples and fast ripples from the seizure-onset zone (SOZ) and the vicinity. In the figure 1, enveloped waveforms of fast ripples and ripples presented with the fluctuation of bursting. The time windows at seizure initiation and phase transition for autocorrelation, the intervals between pHFO bursts slowly were slowly prolonging and turning to a relatively steady before ECoG discharging stopped. No matter the fast ripples or ripples, the time-series autocorrelograms of the SOZ channels and the seizure propagation channels clearly demonstrated rhythmic repetition in a seizure (Figure 2). Abrupt drop of cross-correlation coefficient (red line superimposed on raw ECoG) co-occurred with discontinuation of rhythmic fluctuation of fast ripples when visible ECoG discharging stopped. The termination of ECoG discharging was independent with the power change (black line superimposed on time-series autocorrelogram) of ripples. Conclusions: We analyzed the seizure dynamics of ripples and fast ripples from the seizure-onset zone (SOZ) and the vicinity. In the figure 1, enveloped waveforms of fast ripples and ripples presented with the fluctuation of bursting. The time windows at seizure initiation and phase transition for autocorrelation, the intervals between pHFO bursts slowly were slowly prolonging and turning to a relatively steady before ECoG discharging stopped. No matter the fast ripples or ripples, the time-series autocorrelograms of the SOZ channels and the seizure propagation channels clearly demonstrated rhythmic repetition in a seizure (Figure 2). Abrupt drop of cross-correlation coefficient (red line superimposed on raw ECoG) co-occurred with discontinuation of rhythmic fluctuation of fast ripples when visible ECoG discharging stopped. The termination of ECoG discharging was independent with the power change (black line superimposed on time-series autocorrelogram) of ripples. Funding: This work was supported in part by the Ministry of Science and Technology, Taiwan, under the project MOST 103-2220-E-040-001 and in part by the "Aim for the Top University Plan" of National Chiao Tung University and the Ministry of Education, Taiwan