Abstracts

Rationale: HFOs are discrete electroencephalographic events described as ripples (80-250Hz) or fast ripples (250-500Hz) and suspected to play a role in epileptogenesis. They were first described in rats or epileptic patients with intracerebral microwires (contact surface 0.0013 mm2), but can also be recorded using clinical depth electrodes with macrocontacts. Current clinical depth electrode contacts range from 1 to 9 mm2. The effect of contact size on HFO detection has not been assessed. This is important because the generators of HFOs are believed to be very small: small electrodes may record them better if they are near a generator but large electrodes provide better spatial sampling. The purpose of this study was to assess detection rates and durations of HFOs from electrode contacts of different sizes. Methods: Hybrid intracerebral depth electrodes containing two pairs of adjacent large (1 mm2) and small (0.2 mm2) contacts at their distal end were used in four patients undergoing invasive recordings for the investigation of intractable focal epilepsy between October 2007 and May 2008. These deep electrode contact pairs were located in the amygdala (4 electrodes), the hippocampal head (4 electrodes), the posterior hippocampus (4 electrodes) and cortical areas (8 electrodes). EEG was sampled at 2000Hz after filtering at 500Hz, and 5 minutes of interictal slow wave sleep (no seizure within 6 hours) was selected for analysis. Ripples and fast ripples were marked independently, and the marker was blinded to the contact size. Rates and durations were compared between adjacent large and small contacts using Wilcoxon’s signed rank test. Results: Ripple rates were significantly higher in the large versus small contacts (49.8 vs 38.5 ripples/minute, W(39)=229, p=0.012; fig. 1), but fast ripple rates were not different (W(20)=69, p=0.09). Mean ripple duration was not different between large and small contacts (W(39)=313, p>0.09), but a significantly greater mean duration of fast ripples was noted in the large contacts (54.2 vs 48.7 msec, W (18)=16, p=0.0006; fig. 2). Conclusions: These results indicate that ripples are better detected using the larger of the two contact sizes studied here. As well, the larger contacts detect fast ripples of longer duration. The inclusion of additional patients should confirm these findings and perhaps show better detection rates for fast ripples (the current results show a non-significant trend). Further analysis should include assessment of contact size on interictal spike detection and contact efficacy (artefact), further characterization of detected HFO events (amplitude, co-detection rate), as well as a broader range of contact sizes. Although HFOs were originally observed in microelectrodes, larger contacts may provide a broader spatial sampling and therefore be more efficacious in detecting them.