Abstracts

Rationale: Single pulse electrical stimulation (SPES) on subdural grid electrodes (ECoG) evokes delayed cortical responses (DRs, >0.1 s after stimulation) containing high frequency information and these high frequency DRs are specific for the seizure onset zone (SOZ) [1]. Early responses (ERs, <0.1 s after stimulation) might as well be clinically relevant, but analysis, especially the frequency content, is complicated by their proximity to the stimulation artefact. The Hilbert Huang transformation (HHT) is a novel method for time-frequency analysis on a fine time scale [2]. We studied ER characteristics in the spike (0-80Hz), ripple (80-250Hz) and fast ripple (250-500Hz) band using HHT. We explored frequency characteristics of ERs in relation to the SOZ to increase our understanding of signal propagation in epilepsy and to improve the clinical use of SPES. Methods: We selected two out of 13 patients with refractory focal epilepsy that underwent presurgical ECoG monitoring including SPES at high sampling rate (2048Hz). SPES stimuli (10 monophasic pulses of 1 ms, 4-8 mA, 0.2 Hz) were given on adjacent electrode pairs and responses were recorded in 64 channels. HHT was performed for each epoch, [-1 s 1s] around the stimulus, for each stimulation pair and all response channels. Spectral changes were defined as differences from a bootstrapped pre-stimulus baseline [-1 s -0.2 s]. Colour coded images were constructed with a frequency and time resolution of 15 Hz and 7.3 ms. Averaged images of 10 epochs were visually evaluated for ERs within the spike, ripple and fast ripple band. ERs were defined as clusters of increased power at least partly separated from the stimulus artefact within the post-stimulus ER interval [<0.1 s]. We compared the occurrence for ERs between SOZ electrodes and non-SOZ electrodes. Results: The HHT is able to reveal ERs in all frequency bands (table 1 & fig 1A). For both patients, ratio of ERs in the spike band at SOZ channels are relatively higher than in non-SOZ channels (fig 1B). For patient 2, this relation is also observed for the ripple and fast ripple band (table 1).Conclusions: These results suggest that an increase of ERs in the spike band is related to SOZ. For patient 2 this also holds true for the ripple and fast ripple band. This correlation is similar to what is known for DRs. Traditionally, ERs are considered as physiological. However, our results suggest that ERs contain pathological information as well. High prevalence of ERs could reflect increased irritability to stimuli, resulting from electrophysiological alterations of underlying epileptogenicity. More patients will be analysed to confirm these results.