Abstracts

Rationale: The presence and location of spikes on the interictal EEG are central to the diagnosis of epilepsy. Yet the relation between spike dynamics and epileptogenesis (e.g., frequency of seizures, pharmacologic tractability, etc.) in human recordings remains poorly understood. In previous work (Kalamangalam, GP. Epilepsia 48(s6): 2.198, 2007), a positive correlation was demonstrated between spike morphology variability and seizure propensity in recordings of patients exhibiting either benign focal epileptiform discharges (BFEDs) or periodic lateralized epileptiform discharges (PLEDs) on interictal EEG. This work examines the probability distribution function (PDF) of interspike intervals in the same two diagnostic classes, archetypal of epilepsy syndromes with abundant spikes but widely-differing short-term prognoses. Methods: Digital EEG from routine 20-minute recordings were analyzed from two groups of five patients each with these diagnoses. A single representative channel from each multichannel trace was processed. Spikes were detected and clustered into morphological subgroups with an adapted version of WAVE_CLUS (Quian Quiroga R et. al. Neur Comput 16:1661-87, 2004; software freely downloadable at www.vis.caltech.edu/~rodri). Interspike intervals were computed and histograms charted from spike timings without regard to morphology. The data were fit to parametric distributions (Statistics Toolbox, MATLAB, The MathWorks, Inc.). Results: Spike counts approximately ranged from 300-1000 per trace. As groups, PDFs of the interspike interval for BFEDs were significantly different from PLEDs. In particular, PDFs for BFEDs in general followed a monotonic decreasing profile, modelled adequately by a negative exponential. Under this model, spike generation for BFEDs recalled a Poisson process. PLED PDFs followed a biphasic profile, modelled by a gamma or generalized extreme value function. Conclusions: The qualitative behaviour of different epilepsy syndromes is reflected in the details of time-dependent spike behavior, controlled for overall spike burden. Clinically, BFEDs signify a benign epilepsy syndrome; correspondingly, the PDF of interspike intervals suggested a ‘memoryless’ Poisson process (Yates RD & Goodman DJ. Probability and Stochastic Processes, John Wiley & Sons, 1999) that may underly the infrequency of seizures in this condition. In contrast, a more aggressive epilepsy syndrome (i.e., associated with PLEDs) exhibited a non-Poisson interspike interval PDF, arguing for time-dependent correlation that may underly the propensity for repeated seizures in this condition. Study of interictal spike dynamics affords insight into epileptogenesis, and may help classify and monitor epilepsies of uncertain prognoses.