Abstracts

Rationale: The control of refractory epileptic seizures remains an area of intense research that spans a number of disciplines. Amongst the novel therapeutic strategies being investigated is the delivery of electrical stimuli to neural targets thought to be responsible for the generation, propagation or termination of epileptic seizures in order to prevent or more rapidly terminate seizure events. Intrinsic to this approach is the critical problem of what is the nature of the stimulus that should be applied for therapeutic benefit to be conveyed? . Given the many parameters of stimulation, there exists a massive scope for experimentation outside of the typical values whilst still adhering to the prescribed safety limits for intracranial electrical stimulation. Despite this possibility, only a few publications exist that have attempted to further describe the effects of varying these parameters in a controlled manner. The present study shows that combinations of stimulus parameters outside of those normally used can also convey therapeutic benefit and should be further explored for use in implantable devices.Methods: Five adult, female Genetic Absence Epilepsy Rats from Strasbourg (GAERS) were used to explore the effect of a number of novel electrical stimulation paradigms. In particular, the effect of pulse widths (300 s, 500 s and 1000 s), stimulation rates (125 Hz, 500 Hz and 1000 Hz) and randomised variations in the temporal aspects of stimulation pulses (periodic/synchronous and aperiodic/asynchronous) were explored. Aperiodicity was achieved by creating a Poisson distribution in the inter-stimulus intervals (ISIs), whereas asynchronicity reflected that each pair of electrodes received a pulse-train that was different to the other with respect to ISIs. Electrical stimuli were delivered bilaterally to the somatosensory areas of the rat brains, via implanted screw electrodes. Stimuli were delivered in a closed-loop fashion on the basis of an online seizure detection algorithm using a line-length feature. Seizure duration was quantified by offline analysis of recorded electrocorticograms.Results: The stimulation parameter that was most effective at terminating seizures was the 500 Hz, 1000 s, aperiodic/asynchronous combination which had a significantly different median seizure duration of 2.7 s (Standard error=0.42 s, n=129), versus a non-stimulated median of 9.5 s (Standard error=0.61 s, n=137) (P<0.001, Mann-Whitney Rank Sum Test). The least effective combination was the 125 Hz, 300 s, periodic/synchronous paradigm which had a median seizure duration of 8.8 s (Standard error=0.41 s, n=134), versus a non-stimulated median of 8.3 s (Standard error=0.48 s, n=125). Overall, a multiple linear regression showed both pulse width and temporal type to be significant regressors of seizure duration.Conclusions: This study shows that novel stimulation parameter combinations exist that convey therapeutic benefit for the treatment of seizures and should be further explored for use in implantable devices.